Articles, Reports, Texts, Videos, Talks

No results.

  • 2026
  • Senate Appropriations Committee, via AIP FYI — Senate Science Funding Proposals Revealed. AIP FYI. November 13, 2026. The Senate’s FY2027 draft bills (12 November) give NSF $8.83 billion, up 0.9%, against $7 billion in the House bill; they reject the proposed elimination of NSF’s social, behavioral and economic sciences directorate and require NSF to post funding opportunities within 90 days of enactment unless it explains a waiver to appropriators. The current stopgap runs to 11 December. See also the committee’s bill summary.
  • Science|Business — Votes in Parliament Research Committee Push Horizon Europe and ECF Forward. Science|Business. November 10, 2026. The European Parliament’s ITRE committee adopted its FP10 negotiating position (67 to 1, 18 abstentions) at €200 billion, against €175 billion proposed by the Commission and €158 billion by the Council presidency, with European Partnerships capped at 35 and an independent High-Level Council. A plenary vote is due in the 23–26 November session.
  • Mark Rubin — Gold Standard Science, the RoRI Metascience Report, and Moving Beyond Methodological Constraints. Critical Metascience. November 7, 2026. Quotes Berna Devezer against a centralised metascience unit and “gold standards”, then questions the claim in RoRI’s UK case study that the UK Metascience Unit funds “avowedly ‘critical metascience’”: going through the 23 projects in its first grants cohort, he finds work that takes a critical approach to science but none that questions metascience’s own assumptions, Columbusing, activism or progress. He also summarises his paper with Pekka Syrjänen in Review of General Psychology, which argues that credibility depends on how background assumptions, theory, methods and data support a conclusion together, so constraints on researchers’ methodological freedom are a poor guarantee of it.
  • Martin Kahanec, Naomi Ellemers, Adam Izdebski et al. — How Democracy Depends on Science, and Must Protect It. Nature. November 3, 2026. The European Commission’s Group of Chief Scientific Advisors set out three ways science supports democracy (shared factual baselines, trusted institutions and tools for adapting policy) and warn that cuts to public research budgets make scientists dependent on a narrower range of funders, while institutions often stay silent about selective political favouritism rather than risk their budgets.
  • Hedvig Ördén & Rachel Fishberg — Implementing Research Security in Denmark: The Emergence of Research Security Professionals. Policy Reviews in Higher Education. October 24, 2026. Drawing on media debates, guidelines issued after the Danish Security and Intelligence Service set up its Centre for Innovation and Knowledge Security, and interviews with research security staff, shows how the intelligence services’ involvement provoked public criticism and also sped up universities’ engagement, creating a new profession that mediates between science and the security state.
  • Kenneth M. Evans — New US Congress Should ‘Reimagine’ Science, Not Just Return to the Past. Nature (World View). November 17, 2026. Argues that the new Congress should use the window opened by two years of upheaval to rethink the postwar social contract for science, which privileged experts over the public, by asking who takes part in decisions about science and how policy institutions can include more public voices. Notes that NSF awarded nearly 50% fewer grants in fiscal year 2026 than in 2021–24 and carried more than $1bn into 2027 for the administration’s priorities.
  • David Matthews — How Much Does Science Spend on Improving Itself?. Science|Business. November 19, 2026. Report from RoRI’s Barcelona meeting and the launch of Seize the Meta-Moment!: 12 of 18 countries have no explicit metascience budget, the UK Metascience Unit’s £49m to 2030 is about 0.06% of UK public R&D, and Canada’s oversubscribed call was CAD 2.7m. Metascience is mostly a “side hustle” with few career paths, and evidence rarely reaches decisions (“The challenge in the Netherlands is not generating evidence. It is using it,” says NWO’s Puck Wildschut). Ireland is due to set up the first national metascience unit in the EU in December; NSF’s Jason Gallo considers an ethnographic study of the X-Labs, and Wellcome’s Kirsten Bound plans work on the organisational design of research institutions.
  • Danish Council for Research and Innovation Policy (DFiR) — Strengthening the Nordic Research and Innovation System: Prioritising Nordic Added Value in a Changing Geopolitical Landscape. DFiR policy brief. November 18, 2026. Commissioned by the Danish and Faroese presidency of the Nordic Council of Ministers. NordForsk, Nordic Innovation and Nordic Energy Research together spend about 0.4% of Nordic government R&D budgets, and their funding is small, dispersed and shaped by national co-funders. DFiR proposes four criteria for Nordic added value (shared geography and climate, comparable population data, shared security risks and trust, shared languages and societal models) and a four-step model: political priority areas funded through fewer, larger programmes co-funded with foundations; an independent evaluation agreed before the first calls; then a new mandate and more money only if the model proves itself. It flags the risks to bottom-up research and academic freedom. Based on a CFA report, The Nordics as an R&I Region (Bloch, Fuglsang, Graversen & Norn).
  • Virginia Gewin — A Grant to Hand Out More Grants, and Build a Research Area. Nature. November 18, 2026. The Experiment Foundation’s Field Building Accelerator will, in early 2027, give at least 50 field builders $70,000 each to design their own microgrant programmes of up to $10,000 per researcher around a niche problem, hosted on Experiment.com. Applicants name the field, define its questions and identify at least eight researchers or groups to fund; judges do not average scores, so a single champion can move an application forward. Topics reflect funders such as the Gordon and Betty Moore Foundation (oceans, climate, toxicity, conservation technology). Brian Nosek backs small grants for high-risk ideas.
  • Judit Varga & Wolfgang Kaltenbrunner — The RoRI NARRATIVES Project: Fostering Narrative, Qualitative Thinking on Two Fronts. Go Meta! (RoRI). November 16, 2026. Participant observation of 12 review panels at the Volkswagen Foundation, SNSF and the Dutch Research Council finds that narrative CVs change how excellence is judged only in supportive settings, where format, indicators and the instructions of funders and panel chairs are aligned with reform; CVs alone cannot transform assessment. A framework of 14 contextual factors on five dimensions underpins recommendations for funders, and the authors argue that qualitative evidence deserves the same standing in metascience and policy that narrative information seeks in assessment.
  • Astera Institute — Uncovering Hidden Science: Results from the Hidden Science Competition. Astera Institute. November 17, 2026. An experiment paying life-science PhD students $10,000 each to publish data, tools and results that never fitted a journal narrative: 12 applied and all 10 eligible proposals were funded. Preprint servers sometimes declined or delayed the work, Zenodo and GitHub proved dependable, and preparing work for open release worked as quality control, with one winner finding an instrument artefact that changed his conclusions.
  • Nature — Open Up Peer Review to Early-Career Researchers. Nature (editorial). November 17, 2026. The first large analysis of Nature Portfolio’s co-review programme finds early-career researchers make up 17% of reviewers at Nature and 11% at Nature Communications, that 40% of them are women against 24% of established reviewers, and that 88% of established researchers are more willing to review if they can co-review with one.
  • Global Research Council — Making a Case for Curiosity-driven Research. Global Research Council (discussion paper). November 2026. The discussion paper for the council’s May 2027 annual meeting in Cape Town, co-hosted by South Africa’s NRF and Germany’s DFG, asks how the world’s public research funders can make the case for curiosity-driven research, drawing on responsible research assessment and CoARA.
  • Holly Else — Preprint-Centric Publishing Model Gets Major Boost. Nature. October 28, 2026. Funders from 11 countries, including the DFG and the Swiss National Science Foundation, and research organisations such as CNRS join Open Research Europe’s “publish, review, curate” platform, giving more than 2.5 million researchers free editorial services; the €16m budget to 2030 is 60% funded by the European Commission.
  • Nature — How a New Wave of Science Philanthropists Can Have the Greatest Impact. Nature (editorial). October 21, 2026. New tech philanthropists, from the Arc Institute to the Xplorer Prize and Serrapilheira, fund high-risk, long-term work that public grants struggle with, but should publish their results and grant details, coordinate with public funders and remember that philanthropy still rests on a foundation of public funding.
  • Ministry of Research, Education and Digitalisation — Aftale om fordeling af Forskningsreserven 2027 (in Danish). November 17, 2026. The government and five other parties divide the DKK 2.1bn 2027 research reserve: DKK 1,213m to green research, 247m to critical technologies and defence (75m to AI), 227m to free research and talent (DFF’s free funds rise to DKK 1,144m, with one more year of Inge Lehmann and Sapere Aude), 168m to welfare and social research and 90m to health. The parties note the Danish National Research Foundation’s need for new capital without granting it, and fund a follow-up project to the February 2026 summit on academic freedom.
  • Charlie Coyte, Theresa Reichstadt, Ishita Puri & Susan Guthrie — Modelling Changes to UKRI’s Full Economic Costing Policy. RAND Europe (for the UK Metascience Unit). November 2026. UKRI has paid 80% of the full economic cost of the research it funds since 2002, but actual cost recovery averages about 69%, partly because applicants understate costs to win. An economic model and systems map suggest that raising the rate with a fixed budget would cut success rates and concentrate funding in fewer, research-intensive institutions, while lowering it could push smaller institutions out; either way the system’s capability is put at risk, and a higher rate “would not be a silver bullet” for university finances.
  • Peter Gluckman — Research Degrees: Radical Change Will Be Needed. Substack. November 2026. Argues that the phases of a PhD that build critical thinking (months of reading to find a gap, and the work of interpreting what results mean) are being replaced by “a click” on an AI model, so universities and the scientific community must decide what to protect if humans are still to learn scientific enquiry.
  • Gemma Milne — Taking Root: Focused Research Organisations in the UK. Innovation and Research Caucus. November 2026. From 30 interviews, a workshop and programme documents, traces how the FRO model reached the UK and how Bind Research (Research Ventures Catalyst and a Treasury endowment), Meridial and Echo Labs (ARIA through Convergent Research’s activation partnership) were created. Finds that new institutions need strong networks and sustained “wrangling”, that a different mode of research work motivates their staff, that there is no single template, that government already has the tools to fund institutional diversity, and that universities are valued but hard to engage; recommends more programmes for new institutional forms, shared documentation, opening UKRI eligibility, adaptive processes so new and existing institutions work together early, and a broader idea of the entrepreneurial researcher.
  • James Coe — “We Are a Vital Part of a Nutritious Breakfast”: Stian Westlake on What UKRI Is For. Research Agenda. November 12, 2026. The ESRC’s executive chair says research funding alone will not deliver growth, describes UKRI’s “buckets” (curiosity-driven research delegated to councils, collaborative applied R&D, and support for firms to start, scale and stay) and the first proper refresh of the ESRC’s applicant-led mode in decades, with Connect awards for big ideas, larger challenge awards and a rapid policy-relevant award. He describes experiments with the metascience unit on making review easier and on appropriate use of AI (“peer review is a real black box”), and asks whether strategic investments should selectively grow capability outside the south-east of England.
  • Archana Pyati — The White House Has Embraced Metascience. What Will It Mean for Research Universities?. AAU. November 10, 2026. The Association of American Universities welcomes federal metascience units (NSF’s office, its UKRI partnership and a new NIH metascience office) and traces the idea back to Marburger’s “science of science policy” and NSF’s SciSIP programme, but warns that it arrives alongside cuts, withheld funds and OMB’s grant rule. Former SciSIP directors fear peer review could be sidelined and note the proposed elimination of NSF’s social-science directorate, which built metascience expertise; reforms “should complement, not supplant, universities” (Jenn Gustetic, IFP).
  • Nature — Metascience Should Be a Source of Evidence, Not a Means of Control. Nature (editorial). November 10, 2026. Welcomes governments’ growing use of metascience, from lotteries to new units, but echoes Seize the Meta-Moment! in warning that it can be misused to serve government agendas: in the US, to push through cuts or changes without the scientific community, and in the UK, if funders tie themselves to a single contested view of research and growth. Policymakers should build measures of quality with researchers and learn from experiments.
  • Max Kozlov — California Votes to Give Record $8.4 Billion to Medical Research. Nature. November 10, 2026. Proposition 38, funded by state bonds, earmarks $8.4bn for immunology research, about $4bn of it in competitive grants to public universities and non-profit institutes in California; therapies it produces must be sold at a 20% discount in the state, with 10% of royalties returned. Supporters present it as stability against federal cuts and political control of grants; it has also drawn criticism.
  • Jenny M. Lewis & Kate Williams — National Research Systems Prize Impact, but Not in the Same Ways. LSE Impact Blog. November 10, 2026. Topic modelling of 96 policy documents finds that systems with standardised national impact assessment (Australia, the UK and Hong Kong) define impact more narrowly, around commercialisation, translation and accountability, than the US and the Netherlands, which leave more room for new knowledge, researcher autonomy and institutions’ own aims; the authors suggest combining some structured accountability with flexibility, as the Dutch Strategy Evaluation Protocol does. Based on their article in the Journal of Higher Education Policy and Management.
  • David Lang — The Art of Field Building. Experiment Foundation. November 2026. Argues that new fields of science are made by “field builders”, often outsiders, who connect people already working on an idea without knowing each other, and sets out their toolbox from workshops, small fast grants and journals to prizes, places, vision papers, shared language and new tools, with cases from artificial intelligence to synthetic biology. It launches a part-time Field Building Accelerator, described as “truly applied metascience”, building on more than 300 small grants of up to $10,000 made by empowered “science angels”.
  • Evidence for Democracy — Budget 2026: Targeted Investments in Science, but Big Questions Remain. Evidence for Democracy. November 6, 2026. Canada’s budget (5 November) puts commercialisation first: $930m over five years to widen the SR&ED tax credit to commercialisation costs, new IP and commercialisation policies at the three granting councils, a $1bn Strategic Response Fund from 2028–29 and $162m for 10,000 Mitacs AI placements, with no broad increase for investigator-led funding or indirect costs and no national science strategy.
  • Chris Parr — Research England Announces Overhaul of Targeted Funding. Research Professional News. November 4, 2026. Research England will close its Enhancing Research Culture, Participatory Research and Policy Support funds from 2027–28 (£390m over six years) and put up to £9m over three years into regional and national action on research culture; it also caps any single university’s share of research capital funding at 8%.
  • Andrew Maxwell — Mission-Driven Research (Part 2): What Must Change When We Change the Journey?. Substack. November 4, 2026. With Part 1 (3 November), argues that mission-driven funding needs staged commitments released as evidence comes in, progress judged by what has been learned and room to change teams midway, and that the funding system itself should be redesigned through experiments with the people who must adopt it.
  • Dan Garisto — US Social Sciences Face Existential Crisis After NSF Cuts. Nature. November 3, 2026. In fiscal year 2026 NSF funded fewer than 100 new social-science grants against a historical average of about 850 a year, with none in archaeology, economics or sociology; long-running surveys such as the General Social Survey are at risk.
  • Jonathan O’Callaghan — Is Research Funding Now a Lottery? Elite Funder ERC Swamped by Applications. Nature. October 28, 2026. The European Research Council received 7,710 Starting Grant proposals for 2027, up from 4,807, pushing the success rate below 5%; researchers point to AI making applications far easier to write.
  • Daniel P. Gross & Maria P. Roche — Coordinated R&D Programs and the Creation of New Industries. NBER Working Paper 35978. November 2, 2026. Using archives on the US wartime radar programme, finds that knowledge in papers and patents spread widely while trained people spread mostly locally, and that the programme’s split between component makers and system builders reappeared in the postwar industry.
  • Shohreh Nasri & Sepehr Ghazinoory — The Strategic Rhetorics of Research and Technology Organizations (RTOs). Science and Public Policy. October 31, 2026. A discourse analysis of 19 RTOs’ mission and strategy documents finds four ways they justify themselves (industrial competitiveness, social welfare and sustainability, technological sovereignty, frontier science) and a shift towards acting as network actors in innovation ecosystems.
  • E. Jason Baron, Shawn Kantor & Alexander Whalley — Can Government Support for Innovation Create the Next Silicon Valley?. Journal of Economic Perspectives 40(4). November 2026. Reviews causal evidence from universities, new research institutions, wartime R&D and the space race: public R&D raises local innovation and income, but unevenly and often modestly relative to cost, depending on the entrepreneurs, capital and skills around it.
  • Hemant K. Bhargava et al. — Fighting Fire with Fire: Infusing Artificial Intelligence into Peer Review to Sustain Quality Scholarship. Management Science. August 7, 2026. With submissions to the journal up 22% from 2024 to 2025 as AI-assisted writing spreads, proposes a governed, journal-specific workflow with AI as a frontline reviewer whose output authors can see and contest, and humans in the loop.
  • Christian Hendriksen (Copenhagen Business School) — Kunstig intelligens kalder på omkalfatring af vores forskningssystem (in Danish). Information. November 4, 2026. A kronik arguing that researcher assessment, grants, careers and journals were built for a time when analysis took months, so publication counts stop working once AI does much of the work. Citing an AI proof of the Navier–Stokes problem, AI screening of grant applications in a UK network and the Harvard comparison of AI and human reviews, he proposes judging researchers on what they made it possible to understand, using time freed by AI for better data and replication, and training PhD students to justify and check their use of AI. He warns of lost understanding, convergence of ideas and dependence on a few US model providers, and suggests Denmark is small enough for universities and foundations to coordinate experiments.
  • Catalyst Fund — A Catalyst Fund for Research Software in the Age of AI. Catalyst Fund. November 2026. A one-time fund, fiscally hosted by Hack Club Bank, for the norms, tools, methods and training that make AI-assisted research software trustworthy, arguing that AI coding can bypass the judgment and understanding researchers gain by writing code themselves, and that “when code and tests are both generated, validation risks becoming circular”. Exploratory grants of up to $50,000 for three to six months are open worldwide until 31 December 2026, to individuals and informal teams as well as institutions, reviewed by a panel of five to seven community members, with openly licensed outputs and a short note on what was learned; an impact track of up to $200,000 opens in 2027. Priorities draw on an Edinburgh workshop in March 2026.
  • Rachael Pells — Political Interest in Metascience ‘a Double-Edged Sword’. Research Professional News. November 3, 2026. Coverage of Seize the Meta-Moment!: metascience is shifting from a niche pursuit to a “high-stakes branch of statecraft”, and its authors warn that the surge of political attention risks tying the field to narrow national agendas unless transparency and diversity are protected.
  • James Wilsdon, Josie Coburn and Seunghyun Lee (RoRI) — Eighteen Ways of Looking at Metascience. Go Meta! (RoRI). November 3, 2026. The project coordinators’ launch summary of Seize the Meta-Moment!: most of the work is old (a Polish journal on the study of science in 1918, China’s science of science since 1982) and what is new is the travelling label, though the report itself notes that the word metascience also has a long history; the field is more than ten times larger than the communities that use the name (about 59,000 labelled publications against 476,154 records in the atlas); and there are five national models (directed, delegated, dedicated, decentralised, entrepreneurial) that are “configurations, not grades”. Twelve of the 18 case studies give no investment figures, seven countries lack an integrated research information system, and the authors argue for investing in what lasts beyond the moment: linked open data, routes from evidence to decisions, careers and institutional memory, and published tests of alternatives.
  • Research on Research Institute — Seize the Meta-Moment! How Metascience Is Evolving Worldwide. RoRI. November 3, 2026. The final report of RoRI’s META-MOMENT project, led by James Wilsdon and Ludo Waltman, maps the institutions, investments, infrastructures and capabilities for metascience across 18 countries: Austria, Australia, Brazil, Canada, China, Colombia, Germany, India, Ireland, Japan, Kenya, Korea, Namibia, the Netherlands, Norway, South Africa, the UK and the US. A microsite adds country evidence, a 131-topic atlas of the field, and ecosystem and collaboration analyses. Denmark is not among the countries covered.
  • Cole Donovan — What’s Really Behind the Science Slowdown the White House Wants to Fix?. Tech Policy Press. November 2, 2026. A former OSTP assistant director answers the Cowen–Kratsios interview: business already funds three of every four R&D dollars and a third of US basic research, and federal funding to universities (about $45bn in 2024, behind $70bn to business and $68bn to federal labs and FFRDCs) is roughly 5% of R&D performance, so a slowdown cannot be blamed on it alone. He points to the R&D tax credit, with more than 95% of credit dollars going to the largest 5% of firms and an estimated $190bn cost in 2025–2029 without peer or political review, as deserving at least as much metascientific scrutiny as grants. He also notes that the Genesis Mission has no appropriation of its own while research infrastructure deteriorates, and that Congress, GAO and programme officers already oversee federal grants.
  • Mahmud Farooque — Can (Meta) Science Solve It?. Substack. November 2, 2026. After AEI’s conversation with Michael Kratsios, Farooque (ASU Consortium for Science, Policy & Outcomes; ECAST) welcomes the willingness to experiment but finds a gap between Kratsios’s systemic diagnosis and reforms that concentrate on the productivity of discovery. Drawing on Don Kash, Dan Sarewitz and Motorola’s technology roadmaps of the 1970s (“a roadmap that all of us were to drive”), he asks how a collection of successful reforms would add up to an innovation system that delivers public purposes, and who decides those purposes. He calls for a science of science policy outcomes and an experimental system of innovation governance that is “distributed yet integrated”, drawing on foresight, participatory technology assessment and anticipatory governance: “If metascience is to go the distance, it needs a larger ambition than improving how we produce scientific knowledge.”
  • Kaare Riis Nielsen — Tænketank: Forskningen trænger til reformer – ikke Trumps politiske kontrol. Altinget. November 2, 2026. The founder of the Danish think tank POMUS argues that the US reform agenda addresses real problems (application burden, risk-averse peer review, publication-driven careers, short grants), and that Europe should learn from it without letting reform become a cover for political control: more than 7,800 US grants were cancelled or suspended in 2025, and political appointees now have more say over awards. New models such as NSF’s X-Labs, ARIA and SPRIND should be tested and evaluated over time without declaring existing institutions redundant in advance. “Metascience should be science’s organised self-criticism, not a scientific alibi for political control.”
  • Jen Gold — I Built the UK Government’s Metascience Unit: Here’s What Countries Need to Know. Nature. October 30, 2026. The co-founder of the UK Metascience Unit, run jointly by the science department and UKRI since 2024, draws lessons for the units now being set up in US agencies and in Ireland. Its experiments found that e-mail appeals to reviewers’ sense of obligation did not raise acceptance rates, and that distributed peer review speeds up assessment because the reviewer pool grows with the number of applicants; a trial of staff desk rejection is under way. Her first lesson is to stay small: a team of six could drop an unpromising line quickly. Units need freedom and money to experiment, and leaders willing to make change.
  • Mark Hahnel — What’s So Wrong with Everyone and Anyone Publishing “Research”?. OpenResearch.wtf. October 30, 2026. Digital Science’s VP of open research argues that more publishing is not harmful in itself: if output grows a hundredfold, the share of useful work can collapse while its amount still rises. The harm is that authors keep the benefit of making claims while passing the cost of checking them to others, and that repetition starts to look like evidence. arXiv submissions rose from 113,380 in 2016 to 284,486 in 2025, with growth up to 25% in 2026, and preprints rose from 4.0% to 10.2% of what Dimensions indexes. He proposes a separate, machine-labelled shelf for conjectures and model-generated work that is never cited as evidence, with explicit promotion when an idea is tested, and credit for the replication, correction and synthesis that make other work dependable.
  • Harry Lambert — The Office of the Frontier. Colossus. October 2026. A long profile of Michael Kratsios, who turned the White House Office of Science and Technology Policy from a source of mandated reports into the coordinator of US technology policy, with OSTP drafting the executive orders. It traces his science agenda to Peter Thiel and Patrick Collison and describes the October science summit: metascience units with “the authority to run experiments and compel program officers to participate”, money for people over projects, faster grants, alternatives to peer review and support for young scientists (NIH’s share of main grants to lead scientists aged 36 and under fell from 18% in 1983 to 3% in 2010). It also notes that NSF staffing is down 35% since Trump’s return, that Jim O’Neill is nominated to lead NSF, and that firms pledged $2.4bn in compute to the Genesis Mission.
  • Stian Westlake — Rigour, Pluralism, Impact. Magic Smoke. October 30, 2026. The head of the UK’s Economic and Social Research Council proposes three criteria for a healthy social science ecosystem: rigour (following methods well and adopting powerful new ones), pluralism (a real diversity of interpretive frameworks and priors) and orientation to impact. Each missing criterion has its failure mode: airport-bookshop research and replication crises, ivory-tower work, and groupthink. He admits ESRC collects little scientometric information on past grants, has no systematic review of outputs, and no way to assess pluralism in the fields it funds, though a new analytic capability and the UK Metascience Unit it hosts are changing that.
  • Berna Devezer — Holding Science to a Gold Standard Will Hold It Back. The Scientist. October 29, 2026. A metascientist argues that a centralised Office of Metascience is itself misguided: science is too varied to optimise with one lever, agency-led experiments judged on short-term proxies will become de facto standards, and the push to “double the scientific productivity and impact generated by each federal research dollar by 2036” invites Goodhart’s law. There are no gold standards in science, she writes; even replicability and RCTs are context-bound. She asks for independent ethics review of experiments on funding, which have real consequences for people, and notes that NSF withheld $1.4bn in appropriated funds, hitting the Science of Science programme that funds independent metascience.
  • Noah Smith — A Golden Age of Science Is Coming. Noahpinion. October 29, 2026. Against the “ideas are getting harder to find” literature, Smith argues AI removes three constraints: the number of “effective researchers”, the theory bottleneck and the burden of reading the literature. He cites OpenAI’s thousands of agents analysing a whole-brain CRISPR atlas and Anthropic’s early drug-development results, most of them still preliminary, and speculates about “cloud laws” that AI can exploit but humans cannot understand, which would need vast new sensor data. He allows that humans may remain the bottleneck for open-ended conceptual work.
  • Michael Salmon — A Map of the AI for Science Landscape. Research Agenda. October 29, 2026. Reports on think tank Science Works’ open mapping of UK organisations built around AI-enabled science: more than twice as many founded since 2020 as in the five years before, many of them fellowships, training programmes and networks. Government dominates funding of organisations that fund research, run facilities or train people; venture capital leads on models and tools, with Wellcome the main philanthropic funder. 69% of 276 UK-headquartered organisations and 84% of 116 companies are in London, Oxford and Cambridge.
  • Matt Clancy — Tentative Agreement on How Science Should Be Supported?. The Abundance and Growth Blog. October 28, 2026. Drawing on a New Atlantis interview in which Ari Schulman and M. Anthony Mills describe the US administration’s science policy as a contradictory mix of factions (populists, social conservatives, MAHA, the tech right, national conservatives and science reformers), Clancy finds a possible consensus. The OSTP–OMB budget memo and Matt Faherty’s interviews with people close to NSF, most of whom defend the agency, point the same way: more agency for funders (field maps, empowered programme officers), more variety (X-Labs, ARPA-style programmes, golden tickets, prizes, longer grants) and more evidence (metascience capability, testing, better data). Faherty found scientists far more open to metascience-informed reform than in his 2022 NIH interviews. Schulman doubts that top-down reform can change scientific culture.
  • Mitul Luhar, Eri Saikawa, Julie Risien et al. — Toward Holistic Reform of Researcher Assessment and Incentives at US Academic Institutions. PNAS. October 28, 2026. An editorial from the National Academies’ October 2025 workshop on researcher assessment finds a growing gap between what US universities say they value (community engagement, mentoring, team science, open data, integrity, technology transfer) and what promotion and tenure reward. It warns against the “five-legged sheep”: broader criteria must allow different roles and strengths, without asking everyone to excel at everything. Integrity and open science may be “basic hygiene” rather than rewarded extras (Elizabeth Gadd), and isolated policy changes will not work while the wider incentive system stays the same. It cites the Golden Age report’s call to reform tenure, and proposes a National Academies consensus study and coordination through DORA, HELIOS, PTIE and APLU, with CoARA as a model to adapt.
  • Roger Pielke Jr. — China and the Endless Frontier. The Honest Broker. October 26, 2026. Reading the China Fundamental Research Report against US policy, Pielke argues that calls for more US funding to keep pace with China will fail: China’s government R&D already matches the US at research-specific purchasing power; the report credits China’s rise to institutional continuity while the US shifts from mission-oriented science set by Congress to science directed by the White House; and the US social contract for science has weakened. He asks Congress to reclaim priority-setting through regular reauthorisation and multi-year budgets, to write peer review into statute, and for scientists to ground their case in benefits households can see.
  • Adam Mastroianni — Fake It ‘til You Make It (and Then Keep Faking It). Experimental History. October 28, 2026. A links roundup with two metascience items. On new Data Colada allegations against Dan Ariely, he argues that fabricated findings succeed by confirming what we already suspect, that peer review can tell researchers which results they need, and that the accused survive by admitting nothing. On AI review, he pairs Litvak’s planted-error benchmark with a new Harvard study in which both AI and human judges preferred AI reviews to paid expert human reviews on every dimension, and argues that AI can do the quality control reviewers rarely did, without fixing pre-publication review itself. He also notes that the Golden Age report cites his post “Ideas aren’t getting harder to find”.
  • Conor Griffin — Tools>Theories: A Review of The Engine of Scientific Discovery, by Alexander Krauss. AI Policy Perspectives. October 28, 2026. Krauss’s dataset of 761 major discoveries finds almost all were enabled by a recently developed tool or method, that only 1% were later discarded, and that nearly a third broke the textbook sequence of the scientific method; he proposes grants and fellowships for toolmakers, citation practices that track tools, and “Methods Labs”. Griffin adds four caveats: survivorship bias, the neglect of curated data (no database developer has won a Nobel), the role of imagination, and the back-and-forth between theory and tools, and asks how AI could upgrade, curate and create scientific tools.
  • Duncan McDonnell — Unravelling the University. The Research Whisperer. October 27, 2026. Leaving Griffith University after twelve years, the political scientist describes “the ties that no longer bind”: a widening gap between management and academics, with legalistic HR emails, consultants in senior posts and managers who “fail upwards” (62% of staff in a 2024 survey did not believe the executive listens); students absent from online-first teaching; and collegiality hollowed out, with academics working like independent contractors. He calls for action on university governance and for academics to keep turning up for each other.
  • NSF — Office Head for the Office of Metascience. NSF. October 26, 2026. NSF opens a nationwide search for the inaugural head of its Office of Metascience: a rotator appointment of up to three years, reporting to the NSF Director, with an independent budget and dedicated staff. Topics include high-risk research, proposal review and selection, the cost and burden of funding processes, award size and duration, early-career support, impact measures, AI in research workflows, and replication. The office’s page (23 October) lists pilots of distributed and phased review and golden tickets, and promises to share “experiments and findings clearly and publicly”.
  • Martine Mengers (CBS) — Grant: How Does Research Funding Affect Researchers’ Willingness to Take Risks?. Copenhagen Business School. October 22, 2026. The Villum Foundation gives Professor Hans Christian Kongsted (CBS) DKK 4,999,998 for “The Villum Experiment: Risk, Novelty, and Scientific Trajectories”. Using years of applications to the Villum Experiment programme, which assesses proposals anonymously, the team will measure how novel and risky proposals are with bibliometric and text methods, test whether the programme encourages and selects riskier ideas, follow outcomes in publications, patents and citations, and examine which applicants, early-career or established, are willing to propose risky projects given the career risks. Kongsted tells Science Report (in Danish, November 2026) that, paradoxically, the researchers with the most original ideas often have the least room to take the risk.
  • Tyler Cowen & Michael Kratsios — ‘A Golden Age of Science’ Brought to You by the White House. The New York Times (Interesting Times). October 22, 2026. Cowen presses the White House science adviser on the Golden Age plan. Kratsios says the post-war linear model no longer fits, that the government ran grants with “no metrics whatsoever”, and that metascience units are needed to find what works; he calls the focus on budget totals “very misguided”. He argues political appointees should oversee whether grant review is “meritorious” but not decide individual grants, defends a four-year PhD with a year in industry, and cites NIH grants to over-65s outnumbering those to under-35s. Cowen raises universities as a “holding tank” for dormant ideas, demoralised scientists, falling high-skilled immigration, concentration of power in the executive branch and of AI rents in a few firms.
  • Prachee Avasthi — A Policy Can’t Make True Believers. Prachee Avasthi (Substack). October 21, 2026. Avasthi helped write the Astera Institute’s open science policy: no Astera money or time for journal articles, publication fees or waiting on reviewers, and code, data, protocols and narratives shared openly as soon as they are reusable. Much of her open-science time now goes to adjudicating loopholes, so the mental load has “moved from the journal to the policy”. A policy mainly filters and signals values; the work is to fund people who already work in the open and make openness easy, with tools for citable self-publishing and paid data stewards.
  • Mark Rubin — Self-Correction in Science, Preregistration as Cheap Talk, and a Critical Analysis of the Case for Preregistration. Critical Metascience. October 21, 2026. Three items. Douglas Allchin (Studies in History and Philosophy of Science, 2026) argues science self-corrects through active research and organised scepticism, and that replication can simply repeat an original error, since “artifacts are very, very reproducible”. A SocArXiv preprint by Zion Little, Jacob Montgomery and Keith Schnakenberg models preregistration and finds that selective reporting and vague plans, combined with the credibility a plan confers, can leave the record worse than no plan; Rubin adds that more prescriptive plans mean more deviations. Luke Conway’s long critique concludes that preregistration has “a small-but-important function” but should not be central to psychology.
  • Tommy Shih, Ana Sánchez Cobaleda & Veronica Vella — Emerging Research Security Approaches: Organisational Arrangements at Three European Universities. Policy Reviews in Higher Education. October 23, 2026. Practitioner-researchers compare how KTH (Sweden), the University of Barcelona and the University of Liège turn European research security policy into practice. None has a comprehensive research security policy; each builds incrementally on existing structures, with export control compliance as the usual entry point because it carries clear legal duties. Framing differs: responsible internationalisation at KTH, compliance and foreign interference at Barcelona, prevention of misuse at Liège. The authors ask policymakers to set shared principles without prescribing one organisational model, to invest in universities’ capacity to interpret and coordinate, and to embed research security in wider research governance rather than build parallel systems.
  • Emil Drevsfeldt Nielsen (Dansk Metal) & Mikkel Haarder (DI) — Det er nu, Folketinget skal investere, hvis vi ikke vil være afhængige af teknologi fra andre kontinenter. Science Report. October 23, 2026. As parliament starts negotiating the 2027 research reserve, DKK 2.1bn this year, Danish Industry and the metalworkers’ union ask that at least a third go to critical technologies, rising to at least half by 2030, up from about a quarter last year, backed by an active industrial policy. Part of a Science Report series asking research voices for one piece of advice on the 2027 allocation.
  • Tom Kalil & Eric Gilliam — Kalil’s Corner: The Hidden Value of Field Strategists. SCI PHI (Renaissance Philanthropy). October 23, 2026. A conversation on “field strategists”, people who can see the leverage point that would move a whole field, from J.C.R. Licklider onwards. Kalil argues the US has “prematurely standardized” on a few ways to fund and organise research, mainly committees giving small grants to professors, and that funders should apply the experimental method to how research is funded. Gilliam makes the case for rebuilding BBN-style contract research organisations alongside focused research organisations, and Kalil sets out how to make a moonshot fundable: an ambitious goal, why now, why philanthropy, and social proof.
  • Michael W. Nestor — The Economy Has Changed. We Need a National Frontier Science Fund. Science Politics. October 21, 2026. With federal interest costs projected to double to $2.1 trillion by 2036 and discretionary spending shrinking as a share of GDP, Nestor proposes a dedicated, multi-year fund for basic research, shared infrastructure and high-risk work, financed by matched corporate philanthropy and small assessments on sports betting and digital-asset trading. A core pool would follow an endowment-style spending rule and a smaller pool would use ARPA-style programme managers; payers get no say over awards, and a supplement-not-supplant rule would stop Congress cutting appropriations to match.
  • Susanne Ditlevsen & Stine Falsig Pedersen (Videnskabernes Selskab) — Hvis Danmark skal beholde en af sine største forskningssucceser, kræver det statslig kapital nu. Altinget Forskning. October 21, 2026. The president and secretary-general of the Royal Danish Academy warn that the Danish National Research Foundation lacks capital: without a one-off state injection, the round of new Centres of Excellence now under way will be the last, and the model will wind down after 2027. They cite the 2025 external evaluation, which found the centres outperform Danish research in general and other Nordic centres of excellence, and ask parliament to secure the fund’s capital. Also on the Academy’s site.
  • Tanya Karoli Christensen, Adam Diderichsen & Anders Engberg-Pedersen — Vores snævre syn på innovation overser humanioras potentiale. Altinget Forskning. October 21, 2026. Since the January political agreement to write innovation into the purpose clause of the Danish University Act, with DKK 1.8bn in free innovation funds to 2029, universities are building their own innovation policies. The authors argue the agreement’s view of innovation, as companies, technology and growth, is too narrow. Humanities and social sciences already account for 58% of the roughly 290 companies started each year by University of Copenhagen students and graduates, and much humanities innovation creates value directly through new concepts and practices, such as the analysis that led Danish police to drop the term “husspektakler” for incidents that can involve partner violence.
  • Federation of American Scientists — Metascience Fellowship. FAS. October 2026. FAS will place three fellows in federal research agencies such as NSF and DOE for one year, after a pre-service residency at FAS, to generate evidence on grantmaking, portfolio design, peer review and open science and to design funding experiments with built-in evaluation. The call notes that NSF and NIH have both recently created metascience units. US citizens only; applications close 9 November 2026 and fellows start by 4 January 2027.
  • Anna Butters, Melanie Benson Marshall, Stephen Pinfield, Tom Stafford, Alexander Bondarenko, Robert Nuske, Pierre Schwidlinski & Hanna Denecke — Distributed Peer Review at the Volkswagen Foundation. RoRI Working Paper No. 23. October 2026. Evaluates two rounds (2024 and 2025) of the Volkswagen Foundation’s “Open Up” call, where applicants reviewed each other’s proposals in parallel with the usual panel. The two routes chose very different winners: of 18 funded proposals each year, only three (2024) and one (2025) were picked by both. Simulations show that only about 56% of the proposals selected by distributed review would be selected again if the process were repeated. Applicants valued the extra feedback and perceived fairness but worried about expertise, workload and gaming; most would take part again, and reviewer discipline and experience made little difference to scores.
  • Mohana Basu — China Picks First Researchers for 15-Year Mega Grants. Nature. October 20, 2026. China’s research council NSFC has chosen its first 44 researchers under 35 (women under 38), nine of them from outside China, for grants of up to 15 years in a “five-plus-five-plus-five” arrangement: 4 million yuan for the first five years, doubling to 8 million and then 16 million for those selected to continue. NSFC president Dou Xiankang promised no short-term target assessments: “We will assess only one thing: whether you are there in the laboratory.” Names and projects have not been released. Nature sets the scheme beside other long grants: the ERC’s new ERC Plus Grants (up to €7m over seven years, first call in 2026) and HHMI’s Freeman Hrabowski Scholars (five years, renewable once). Zheng Liu (Guilin Medical University) calls it “a good start”, since young researchers’ grants often end “before the work matures”.
  • James Wilsdon — Does Adding Mean Increase? Measurement, AI and the Remaking of Research. Frontiers Policy Labs. October 20, 2026. Across twelve funders, grant application volumes rose by an average of 57% between 2022 and 2025; Marie Skłodowska-Curie applications rose 64.6% from 2024 to 2025, low-scoring proposals became rare and a fifth scored above 95%. With AI removing effort and polishing text, criteria of “excellence” no longer do the work they did, so adding risks dilution. Wilsdon argues for changing what we count: contribution over rank (as in the China Fundamental Research Report), validated versus unvalidated over basic versus applied, and “pull” funding models. A RoRI summit of assessment-framework architects from eleven countries found that governance matters more than technology. He asks the ISC to set a norm that no funder, journal or university uses AI in assessment without disclosing it, measuring its effects and testing whether it spots what humans miss.
  • James Wilsdon (RoRI) — The Frontier Is Fundamental: Why the China Fundamental Research Report Matters. Go Meta! (Research on Research Institute). October 20, 2026. Wilsdon, who co-chaired the international review committee, on the China Fundamental Research Report (2026), commissioned by NSFC for its fortieth anniversary and launched in Beijing. It is the first open, system-level account of how China funds, organises and evaluates fundamental research: about 250 pages in English, built on open data (OpenAlex, CORDIS) with reproducible figures and some NSFC internal funding data. China now produces about a third of the world’s fundamental research papers, up from under 1% in 1985, but spent about 0.17% of GDP on fundamental research in 2023, against about 0.50% in the US and 0.74% in South Korea, and some 4.6 million Chinese-language articles sit outside global datasets. The report refuses to rank and asks what a system contributes. Wilsdon poses three questions for the governance of science: whose categories, who is visible, and who decides, as AI models trained on past winners enter assessment. RoRI publishes its 18-country study Seize the Meta-Moment! on 3 November.
  • UKRI & DSIT — Data for Metascience Sandpit 2027. UKRI. October 20, 2026. A £2m funding opportunity from all UKRI research councils and the Department for Science, Innovation and Technology: a five-day sandpit on 1–3 and 9–10 March 2027 to build projects on multi-faceted metascience questions about R&D funding that can inform policy. Expressions of interest open 27 October and close 24 November 2026.
  • Vinton Cerf & France Córdova — On the Future of Science and Its Wise Pursuit. Physics Today. October 20, 2026. The internet pioneer and the former NSF director (now president of the Science Philanthropy Alliance) argue that AI tools do not guarantee progress, which still depends on free researchers, institutions capable of long inquiry, shared infrastructure and patient funders. With US public research institutions under strain, they call for sustaining them alongside new partnerships, and propose an annual forum to assess whether the architecture linking government, universities, industry and philanthropy keeps pace with science, since infrequent field-by-field assessments cannot.
  • Ben Reinhardt — Rounds, SAFEs, and Glory. Spectech. October 20, 2026. What philanthropy can borrow from venture capital to fund risky work by new organisations. Philanthropy lacks funding rounds, ties money to projects so grantees fundraise constantly, clusters checks below $1m or above $10m, and gives funders no credit for backing successes early. Reinhardt proposes a “philanthropic SAFE” built on matching, shared “diligence objects”, prizes and impact portfolios for funders who back things early, anchor funders and syndicates, and “exits” such as long-term government funding for high-performing research nonprofits. An appendix sketches funding stages from idea generation ($100–500k) to scaling ($50–200m+).
  • Katharine Sanderson — ‘Landmark’ Report on Chinese Research ‘a Wake-up Call for West’. Research Professional News. October 19, 2026. News report on the China Fundamental Research Report. James Wilsdon calls it “the most comprehensive deep dive by China into its own system that we’ve ever seen” and a reminder that “weakening our science base as a short-term response to security challenges leaves us ultimately in a much worse position”. China is “putting a lot more money into early career and early mid-career scientists than we do in the West, and they’re giving a lot more smaller grants to a lot more people”. Lead author Lin Zhang says international discussion is dominated by a few indicators that “cannot adequately explain how research systems actually develop and function”.
  • The Economist — America’s New Era of Private Science. The Economist. October 19, 2026. Metascience advocates are putting their ideas into practice through private institutions, many founded in the past five years, while the government’s Golden Age report says universities should not be “the only home” of innovative research and NSF plans $1.5bn for its first X-labs. Cites a 2025 study of 100,000 NIH applications (Azoulay & Greenblatt) finding out-of-consensus projects less likely to be renewed, and a survey putting researchers’ administrative time at about 44%. Profiles Convergent Research’s FROs ($20–50m over three to seven years) and Episteme, backed by Sam Altman, which gives lead researchers near-total freedom within commercial scope; Arena BioWorks closed after under two years. Scientists worry that less money spread across more institutions will hurt university labs, as the FY2027 request would halve NSF.
  • Stefano Baruffaldi, Felicia Jorzick & Cornelia Lawson — Research Grant Effects Across Funding Modes and Industry Collaboration. Research Policy (December 2026 issue, open access). A regression discontinuity study of 2,123 EPSRC applications from 2007, comparing researchers just above and just below the funding line. A grant raises publications by about 13% over five years and brings about 0.8 more grants, but no gain in novelty or disruption. Open, investigator-initiated grants work best without industry partners; directional, funder-initiated grants work best with them. The authors read this as a case for matching a funding mode to the way knowledge is produced, while stressing that the subgroup patterns are not causal.
  • Andrew Gerard & Jenn Gustetic — Eight Questions About Metascience Units. Macroscience. October 19, 2026. A follow-up to their July IFP proposal, after NSF’s Office of Metascience, the House Build America Caucus agenda and the Kyoto Vision put metascience units on the agenda. Units should differ by agency, do most work internally, publish preregistered work plans and release findings without leadership review, have their own appropriation instead of a reimbursable model, be led by a career director with an external research deputy, use qualitative as well as experimental evidence, and serve bottom-up requests so offices become champions. They warn of four failure modes seen in US government innovation units, from “cycles of responsibility” to changes of administration, and say a unit needs about ten years to show results.
  • Børge Obel — Problemet er ikke RUC. Det er hele universitetssystemet. Politiken. October 18, 2026. Writing after Roskilde University’s de facto dismissal of professor Michelle Pace, the former rector and dean argues that three decades of reform have moved power from academic authority to management: the end of civil-servant status for new professors and external members on governing bodies (1993), boards with external majorities and appointed leaders (2003), and ministerial approval of board chairs (2017). Each step was defensible alone; together they weakened researchers’ protection and invite self-censorship. Because the problem is systemic, he proposes five changes that work together: a protected form of employment after external assessment, boards chosen by a broad representative body, larger and longer-term basic funding redistributed from political pools, a collective professorial veto on permanent appointments, and academic councils with a suspensive veto. All could be done almost cost-neutrally through the University Act.
  • Mogens Fosgerau — Some Thoughts on the Future Knowledge Ecosystem with AI. LinkedIn. October 17, 2026. The University of Copenhagen economist distinguishes AI and human “knowledge capital”. Who owns AI-generated results will decide who benefits, so open repositories such as arXiv, or an AI-oriented successor, could keep that knowledge public by design. Human expertise remains the precondition for good questions, but if AI removes the struggle through which researchers build expertise and earn recognition, it may erode the human capital that makes AI useful. He asks how education, institutions, incentives and recognition can be rebuilt to keep producing deep understanding.
  • Rob Marus (AAU) — White House Launches New “Golden Age” of Science Initiatives – But Other Administration Policies Could Prevent Success. Association of American Universities. October 16, 2026. The research universities’ association welcomes the 8 October initiatives, including a six-year integrated bachelor’s-to-PhD programme at NSF and a four-year DOE doctoral pilot, but notes that much of the money comes from state or private sources. It argues that cuts, conditions and slowed grant-making have forced universities to reduce PhD admissions, and that visa and H-1B policy deters foreign talent, so other administration policies could undo the report’s aims.
  • NIH — Consider the New RM1 Broad Funding Opportunity for Your Team Science Project. NIH Extramural Nexus. October 16, 2026. A new Collaborative Program Grant for Interdisciplinary Teams (RM1, PAR-27-003), shared across several NIH institutes, funds integrated teams working toward one shared goal over five to ten years, without the separate subprojects of a P01 programme grant. Reviewers assess whether the science truly needs a team, how well frameworks and methods are integrated, and the leadership plan; applications are expected to plan decision-making, resource allocation and professional credit and to include several career stages.
  • Mike Schmierbach — Academic Publishing is Broken. We’re All to Blame. Mass Communication and Society (editorial). October 15, 2026. The journal’s outgoing editor reports that submissions more than doubled in five years and that finding reviewers now takes a dozen or more requests, with growing signs of AI in reviews. He blames for-profit publishers but calls the crisis a collective failure of incentives that reward paper counts, and proposes visible credit for reviewers, open post-publication review, promotion based on a few strongest papers, rewards for collaboration, and journals pooling resources to leave commercial publishers.
  • Stephen Curry — Don’t Be Dazzled by Fools’ Gold. Go Meta! (RoRI). October 16, 2026. Calls the Kyoto Vision sensible but innocuous, and the Golden Age report behind it nostalgic, nationalistic and dismissive of universities and international collaboration. He finds the German Rectors’ reply “a bit tin-eared” and points instead to Sarah de Rijcke’s six principles of academic freedom and Geoff Mulgan’s rules for healthy disagreement, arguing that verification at scale and agency metascience units are open disputes best settled through plural perspectives. RoRI’s META-MOMENT report is due on 3 November.
  • Daniel Garisto — ‘Complete Devastation’: How NSF Gutted Its Social Sciences Research. Science. October 6, 2026. NSF’s social, behavioral and economic sciences directorate, which the 2027 budget request proposes to eliminate, spent $95m in 2026 against a $247m average and made 91 new grants against about 800; more than 30 programmes, including economics, sociology, archaeology and research ethics, made no new awards, and grantmaking staff fell from 30 to about a dozen.
  • Matt Faherty (New Science) — New Science’s Report on the NSF. New Science. October 2026. Based on 31 interviews, finds NSF better run than NIH, with program officers who can back riskier work and rotators who keep it open, yet sharing the same grant treadmill, hyper-competition and paperwork (about 120 hours per proposal). The central finding is that most reform ideas, from people over projects to lotteries and block grants, cannot be judged because there is no accepted way to measure research quality, and NSF’s past pilots were small, decentralised and rarely evaluated. It welcomes the Office of Metascience, announced on 10 September 2026, and agrees with most of the Golden Age report, while warning against making political missions the organising principle of the whole system.
  • UKRI — UKRI publishes new research data policy. UKRI. October 14, 2026. A single policy replaces UKRI’s earlier data policies and principles, aiming to make publicly funded data findable, accessible, interoperable and reusable (FAIR). Data behind published findings must be shared no later than publication, and other data with reuse value within two years of the end of funding. It brings simpler data management plans, requires sharing of the code and software needed to validate findings, asks for contributor credit using taxonomies such as CRediT, and adds UKRI-wide guidance on sharing data about people. It applies to new funding opportunities from 1 April 2027, and UKRI will fund good-practice resources for research software and arts and humanities data. See the policy itself.
  • Xiaoying You — China’s Extraordinary Rise in Science and How It Will Change the World. Nature. October 14, 2026. China has gone from 0.02% of international papers in 1978 to a leading R&D spender, and plans at least 7% annual growth in R&D spending to 2030. Science is tied to industrial and security strategy, from a state semiconductor fund (a third round of 344 billion yuan in 2024) to “strategic emerging industries” and a five-year plan that gives firms a bigger role. Fundamental research is just under 7% of R&D (7.08% last year, missing an 8% target), against 14.4% in the US. Despite a 2020 evaluation reform, paper counts still drive hiring and promotion, and nearly 60% of retracted papers in 2014–2024 had China-affiliated authors.
  • Allison Stanger — Who Cares Whether Academics Use AI to Write?. The Chronicle of Higher Education. October 13, 2026. A journal accepted Stanger’s article, written over ten drafts with large language models, then asked her to rewrite sentences that “read like something written by AI”. She argues that analytical writing should be judged by its quality, which any reader can test, and not by its suspected origin, which no reader can verify: AI detectors gave the same excerpts verdicts from very likely human to 100% AI. The author who signs the work answers for it, and a byline has never certified unassisted composition.
  • Stuart Buck, Aishwarya Khanduja & Hiya Jain — The Overlooked Memo That Could Reshape Federal Science. Asterisk. October 13, 2026. Argues that the joint OSTP–OMB budget memo of July 2026 matters more than the Golden Age report, because OMB controls the money. Agencies spending over $3bn on R&D must deliver action plans within 90 days on practices such as a diverse portfolio of institutions, grants of five years or more, fast grants, prizes, pilots of golden tickets and regranting, metascience offices, and tracking of near-miss applicants. The authors note that the memo never mentions replication or reproducibility, and that budgets have proposed cutting the graduate fellowships the report wants to expand.
  • Eric Gilliam — Help Wanted: Technical Scouts. FreakTakes. October 13, 2026. An update on the BBN Fund, which helps build contract-and-grant funded R&D organisations on the model of Bolt Beranek and Newman: ARIA has doubled its BBN launchpad cohort to eight (£800k+), BBNs in the ARIA ecosystem have brought in about $20–30m in contracts, and Coefficient Giving has granted $700k. Gilliam is recruiting technical scouts to match founders with funders and customers, the fund’s main bottleneck.
  • Mark Rubin — Preregistration: Deviations, Concerns, and Ignorance. Critical Metascience. October 11, 2026. Three new studies: an Institute for Replication paper finds that 84% of preregistered psychology articles deviate from their plan and that 40% of those disclose none of the deviations; a survey of authors of observational health studies finds resources, tradition and knowledge of registries are the main barriers to registration; and Torsten Wilholt argues that standards cannot replace trust in research aimed at “deep ignorance”, where no candidate answers yet exist.
  • Torsten Wilholt — Is Trust in Science a Good Thing? European Journal for Philosophy of Science 16, 74. 2026. Argues that trust, within science and from the public, cannot be replaced by control. Standards such as preregistration and statistical frameworks work for questions with known candidate answers, but not for research on “deep ignorance”, where no candidate answers exist; a science that only pursued the former “would cease to be the creative, transformative enterprise that merits public trust”. He calls the US “Restoring Gold Standard Science” order the starkest example of the metascience movement’s reproducibility narrative being turned against science, since its standards could bar the best available evidence from regulation. Open access.
  • Research Professional News: Proposal to cut FP10 by 10% decried as 'choosing to fall behind'. Research Professional News. October 12, 2026. The Irish Council presidency's 10 October proposal sets FP10 at €158 billion (inflation-adjusted), €17 billion below the Commission, while Parliament's co-leads hold to €200 billion.
  • Julia Rock-Torcivia — NSF Looks to UK Experiments as It Plans to Test How It Awards Research Grants. R&D World. October 9, 2026. Puts the NSF–UKRI statement in context: no money or projects are named yet, and NSF has not said when its tests begin. The UK Metascience Unit has run on about £3–4m a year; its distributed peer review pilot cut decision times to about 2.2 months, while evidence for funding lotteries is so far weak. A preprint on rejected ICLR submissions finds that “golden ticket” rescues did about as well as a lottery and worse than rescuing the highest-scored rejects. NSF’s FY2026 budget is $8.75bn after Congress rejected a $3.9bn request; the FY2027 request of $3.96bn would be a 55% cut.
  • NWO — Dutch Research Council (NWO) Cancels Subscriptions to Web of Science and Scopus. October 9, 2026. From 2027 the Dutch Research Council drops both proprietary databases and will use open sources such as Crossref and OpenAlex, mainly for finding reviewers and checking conflicts of interest, in line with the Barcelona Declaration on Open Research Information. It notes the databases’ bias toward English-language and natural-science publications, and follows CNRS, several Dutch universities and CWTS, which has stopped using Web of Science for the Leiden Ranking.
  • Jonathan Owen — Burnham Commits £1bn for Regional Innovation over Four Years. Research Professional News. October 9, 2026. The UK prime minister commits £1bn from existing budgets over four years to back regional innovation clusters with UKRI, in “a more place-based approach”, and brings forward to 2028 the transfer of UKRI’s £500m Local Innovation Partnerships Fund to mayors, calling it “the first step towards devolving dedicated innovation funding to local leaders”. The same day, 18 university and business leaders warned that devolution risks duplicated investment and resources spread too thinly.
  • Michael Kratsios — Remarks at the Science: A New Golden Age Summit. White House OSTP. October 8, 2026. Names the first metascience experiments: NSF’s Office of Metascience starts with golden tickets and distributed review, NASA’s Science Mission Directorate tests double-blind peer review, and DOE runs its own golden-ticket experiment. NIH launches R3PEATS, a $174m reproducibility initiative, and a “linked discoveries” tool to track whether findings have been replicated. NSF and NIH will move more than half of their application forms to “just in time” submission. Kratsios argues that Vannevar Bush’s linear model no longer fits a recursive innovation cycle in which private funders play a much larger role.
  • The White House — Fact Sheet: Trump Administration Announces the Most Ambitious Set of Science Initiatives This Century. October 8, 2026. At the “Science: A New Golden Age” summit the administration announced what it calls over $6 billion in federal, industry and philanthropic commitments. For metascience: NSF, DOE, NASA and the Department of War launched metascience initiatives to run experiments on how they fund and conduct science, and the NSF–UKRI cooperation is described as the foundation for an International Network of Metascience Units. Also: an X-Labs Consortium ($310m), $2.4bn in compute credits from eleven companies for the Genesis Mission, two $33m NSF prize challenges, a $215m quantum competition, a $1.8bn virtual-cell initiative, and accelerated PhD and early-career programmes framed as replacing “DEI with merit”.
  • U.S. National Science Foundation — NSF Lays the Foundation for the Next Generation of Technologies and Support for American Talent. October 8, 2026. NSF is establishing an Office of Metascience as a testbed for how scientific communities respond to new review and funding mechanisms, operating “with full transparency”; a job announcement for its head is to follow. NSF and UKRI say they intend to build on the metascience principles of the Kyoto Vision (see the joint statement). Also announced: two prize challenges (Quantum+X and synthetic multicellularity), an X-Labs Consortium with over $300m in partner commitments for space and instruments, up to $75m for “super intelligence-native” lab instruments, and $30m for National Math Stars.
  • Benjamin Steyn, UK Metascience Unit — Metascience Investment Explainer on AI-Driven Research Evaluation. October 8, 2026. The Unit’s co-head explains about £5m for seven sandpit projects on AI-driven indicators of rigour, novelty, new code and data, over-reported positive results, research fraud, economic value and the reliability of AI-assisted assessment, with partners including UKRN, ARMA and major publishers. The aim is a validated “basket of indicators” that assists human reviewers, with the weighting left as an explicit human value judgement. Philosophers of science, RAND Europe and RoRI will examine the normative and political questions. The Unit stresses it is building evidence, not committing to use the indicators.
  • NSF & UKRI — NSF–UKRI Metascience Statement. October 8, 2026. The two agencies will collaborate on metascience, citing the Kyoto Vision, NSF’s Golden Age memorandum and the UKRI Strategy 2026–31, in four areas: measuring research productivity and impact, improving review and funding, assessing how new methods and tools are used, and responsible analysis of metascience data. The partnership between NSF’s new Office of Metascience and the UK Metascience Unit is to seed an international network of government metascience units for shared experiments; the collaboration is to be formalised in the coming months.
  • Alex Imas & James Manyika — Bending the Curve of Discovery: AI in Science Today and Tomorrow. DeepMind Institute. October 8, 2026. Two Google DeepMind authors combine 15 million Gemini interactions, data on 2,600 specialised AI models and a survey of 637 US and UK scientists (paper). Nearly half use AI daily and save about seven hours a week, but bottlenecks are moving to lab testing and verification, and almost half say AI steers them toward safer, incremental questions (against about 30% toward riskier ones). They call for verification infrastructure, grant and peer-review reform that rewards high-uncertainty work, and open models and data.
  • David Matthews — Concern After Europeans Endorse US Science Agenda. Science|Business. October 8, 2026. Germany’s DFG says it will “closely monitor” whether the US uses the Kyoto Vision to restructure funding to the detriment of universities, and the German Rectors’ Conference says the call for “new institutional models” raises questions about academic freedom. The fear is that the endorsement lends international cover to a US agenda that derides universities as “legacy institutions”. Germany’s research ministry says it did not agree to the late switch to “super intelligence”; a Commission official says the US approached the EU too late for it to consult member states.
  • Matthew Esche — Kevin Bryan on AI and Innovation. Macroscience. October 8, 2026. Lecture 7 of IFP’s Economics of Ideas series. Bryan argues that AI’s gains for science will come from reorganising labs and institutions around it, as factories were rebuilt around electricity, and that bottlenecks such as experiments and trials cap the growth effect: in Ben Jones’s model, even unlimited machine intelligence raises research productivity only about 2.25 times. He also points to the five-fold salary gap between academic and industry AI researchers as a sign of how sticky academic organisation is.
  • Peter Singer — A Canadian ARPA-H Would Pay Three Times. Global Health Insights. October 8, 2026. The founding CEO of Grand Challenges Canada proposes an arm’s-length Canadian ARPA-H that links challenges to provincial health systems, which act as first customers, runs a venture-style funnel of many small bets, and books the fiscal return of each innovation with the finance ministry. He estimates a public cost of about C$9bn over 25 years against about C$68bn returned to government, figures he says should be tested. Existing bodies each cover only part: NRC’s DARPA-inspired Challenge programmes drew only 5% of collaborators from business.
  • Nature editorial — How to Mobilize the Agile Science Needed in a Crisis. October 6, 2026. Reports two UK initiatives. Oxford’s Agile Initiative tested three ways of setting topics for 6–12-month grants funded by NERC: researcher-led, decision-maker-led, and co-created; only co-creation produced many strong ideas. The Co-Production Futures Inquiry, funded by Research England, found barriers in paying non-academic partners, short contracts, publication requirements and individual credit. Both stress that funders must allow experiments to fail.
  • Ziming Luo, Atoosa Kasirzadeh & Nihar B. Shah — The More You Automate, the Less You See: Hidden Pitfalls of Autonomous AI Scientists. PNAS 123(41). October 5, 2026. Controlled experiments on two open-source “AI scientist” systems (Agent Laboratory and The AI Scientist v2) find a built-in post hoc selection bias, equivalent to automated p-hacking, as well as easier benchmarks chosen when baselines are shown and undisclosed synthetic or subsampled data. An auditor reading only the final paper detects these flaws barely better than chance (53.7% accuracy); with logs and code it reaches 81.5%. The authors argue that journals and conferences should require full workflow logs and code for AI-generated research, since manuscript-only peer review cannot catch these failures.
  • Anders Eldrup, Jens Ringsmose & Thomas Bjørnholm — Innovationsaktører: Forskningsidéer skal blive til virksomheder. Men vi ved for lidt om vejen derhen. Altinget. October 5, 2026. The chairs of Innovationsfonden, Danske Universiteter’s rectors’ conference and DFiR argue that Denmark lacks a shared picture of the path from research idea to company, so the debate runs on anecdotes while spin-out, patent and invention statistics have stagnated. DFiR will spend three years building a consolidated data foundation that links four levels: research environments, technology-transfer units, early soft funding, and private capital and revenue.
  • BMFTR — „Kyoto Vision“: Neue Impulse für den transatlantischen Forschungsdialog. October 4, 2026. Germany’s research ministry explains its endorsement of the Kyoto Vision: the paper is “a basis for discussion” and not a binding declaration, its support gives Germany a foot in the door for US–German science talks planned for November, its proposals will be examined further since national research systems differ, and the independence of science and freedom of research remain non-negotiable.
  • Science, technology and innovation ministers — Kyoto Vision for a Golden Age of Science (PDF). October 4, 2026. The one-page statement itself, written in soft terms: signatories “endeavor”, “encourage” and “note the value”. It notes the value of “a deliberate portfolio of funding mechanisms including long-duration awards, rapid grants, and prizes and challenges”, encourages experimentation with institutional models offering “varying levels of autonomy, resources, and direction”, and welcomes metascience, including “dedicated metascience capabilities in our governments”. It also commits to reproducible, transparent science with unbiased peer review that treats negative and null results as valuable. The other sections cover “Super Intelligence” for science, and talent, including staff who operate and maintain advanced instruments.
  • The White House (OSTP) — U.S. Leads International Coalition to Endorse Kyoto Vision for a Golden Age of Science. October 4, 2026. At the 23rd STS Forum in Kyoto, the US and 16 countries (Argentina, Bulgaria, Chile, Cyprus, Germany, Greece, Indonesia, Italy, Japan, Kazakhstan, Korea, New Zealand, Poland, Singapore, the UAE and the UK) endorse the Kyoto Vision. Its first priority, improving the institutions of discovery, values a variety of funding tools (long-duration awards, rapid grants, prizes and challenges), encourages experimentation with institutional models offering different levels of autonomy, resources and direction, and welcomes metascience “as an empirical field to improve how science is conducted, organized, and funded for public benefit.” The other priorities are AI (“super intelligence”) for science and cultivating talent. No Nordic country signed.
  • Jared Perlo — Global Tech Policymakers Agree to Embrace AI in Science. NBC News. October 3, 2026. The US and 15 other countries, including Japan, Korea, the UK and the UAE, release a “Kyoto Vision for a Golden Age of Science” at the STS Forum, led by OSTP director Michael Kratsios. It says “new ways of organizing research are opening entirely new frontiers”, values “a deliberate portfolio of funding mechanisms” and welcomes the “growing field of metascience”, which IFP’s Jenn Gustetic reads as international adoption. Critics point to proposed cuts of about half to NSF’s budget.
  • Prachee Avasthi — arXiv Is Killing Itself So AI Doesn’t. October 2, 2026. A critique of arXiv’s new cap of two submissions per submitter per month (three active at once, rejections included), introduced after submissions doubled in two years to 40,363 in September. Avasthi argues the cap sizes the scientific record to volunteer moderator time, penalises honest researchers while missing the small share of bad actors, and pushes preprints back toward journal-style judgements of what is “complete”. She proposes automated provenance and coherence checks and pattern detection, with volunteers kept for real judgement calls.
  • Craig Nicholson — EU R&I Chief Wants ‘Creative Destruction’ and Horizon Opening. Research Professional News. October 2, 2026. At the launch of the Commission’s SRIP report, Marc Lemaître said EU funding may have created hurdles for grassroots researchers and small innovators and called for “alternative funding methods” in the next Horizon programme. Newcomers make up 55% of organisations in Horizon Europe but win only 14% of the funding. Philippe Aghion repeated calls for an EU ARPA-style agency and for the ERC to fund institutions, and Reinhilde Veugelers called the 3% R&D target “the wrong target”.
  • Asad Ramzanali, Ian R. Banks & Heather Douglas — Defending Science From Partisanship. Issues in Science and Technology 43(1). Fall 2026. Three responses to M. Anthony Mills’s essay on who should decide how to fund research, after the White House fired the National Science Board. Ramzanali argues Congress should scrap the board, which did little to shield NSF from political interference. Banks asks for bipartisan safeguards against partisan control by either side, ahead of NSF’s reauthorisation in 2027. Douglas argues for different funding systems for different kinds of science: screened lotteries, geographic allocation or baseline funding for curiosity-driven research, since peer-review ranking feeds concentration at elite institutions and groupthink; oversight for mission research; and a distinct system for publicly engaged science.
  • Nicklas Berild Lundblad — Unpredictable Patterns #156: On the Organization, Growth and Governance of Socratic Networks. Unpredictable Patterns. October 1, 2026. Treating science as a network of questions and answers, Lundblad argues that machines may now answer questions faster than meaningful new ones can be posed, citing the Erdős problems (about 565 solved and 652 open by August 2026) and the Fields medalists’ September 2026 declaration on AI in mathematics. He sketches “m-science” growing at the speed of compute and “h-science” at the pace of human attention, and argues the task is to govern the two pace layers and translate findings between them.
  • Richard Jones — On the Devolution of Research, Development and Innovation Funding: Some Tentative Principles. Soft Machines. October 1, 2026. As the UK government moves power and funding to local leaders, Jones sets out which R&D funding should be devolved. Discovery research should stay national, with expert peer review and UKRI in charge; the innovation end, aimed at raising private R&D and building clusters, can move to Strategic Combined Authorities. Devolved money should be ring-fenced for R&D, governed by university, business and local government together with outside experts, and backed by new UKRI capacity. He warns of capture, sub-critical pots and facilities split across neighbouring regions.
  • James Brooks — ESRC Targets Project Diversity and Risk-Taking in Grants Shake-Up. Research Professional News. October 1, 2026. Stian Westlake sets out four drivers for the ESRC’s new applicant-led schemes: a more outcomes-focused approach, strategic management of a portfolio that had become “inadvertently very concentrated”, more novel and risk-taking research, and less burden. Jen Gold, co-leader of the UK Metascience Unit, says Explore’s cap of four core researchers comes from metascience evidence on the role of small teams in new research directions. Explore offers up to £2m, Grow up to £500,000 and Challenge up to £10m, all open year-round alongside the Connect and React awards.
  • Research Agenda — ESRC Sets Out Its Plans for New Funding Awards. October 1, 2026. The UK Economic and Social Research Council is closing its broad applicant-led schemes (new investigator grants, research grants and secondary data analysis) on 4 November and, with no extra money, replacing them with three funds organised around ways of working: Grow, for researchers at key career points to build capabilities or change direction; Explore, for teams of up to four pursuing bold, curiosity-driven and high-risk ideas; and Challenge, for larger programmes on major challenges, open to applicants who have not held large awards before. Builds on the React and Connect pilots launched in spring.
  • European Commission (DG Research and Innovation) — Science, Research and Innovation Performance of the EU 2026 (SRIP). October 1, 2026. The Commission’s biennial analysis of the EU research and innovation system frames three messages: strength (second globally in scientific output), structure (a system fragmented across countries, regions and institutions) and scale (weak translation of research into market leadership). Of note for metascience: EU and national R&I funds reinforce each other instead of crowding each other out; precarious contracts and “publish or perish” call for responsible research assessment; self-reinforcing agglomeration concentrates R&I in already strong regions; and a contributed chapter argues the EU has a narrower pipeline of science-intensive innovation from public institutions, calling for stronger institutional channels and better tracking of publicly funded innovation.
  • Jakob Edler, Henning Kroll, Thomas Stahlecker & Max Mittenzwei (Fraunhofer ISI) — Europe Can’t Wait: Building Flagship Innovation Clusters to Restore Competitiveness. European Round Table for Industry (ERT). September 2026. Based on interviews with senior executives, the study finds Europe’s established innovation hubs strong and “sticky”, but warns that firms increasingly place next-generation R&D in US and Chinese clusters, notably in AI, quantum, biotech and semiconductors. It recommends concentrating resources on a limited number of designated “flagship clusters” with long-term funding, anchored on university campuses, alongside a single market for research and talent (a “Fifth Freedom”), an EU talent visa, deeper capital markets and innovation procurement.
  • Jonathan Breckon & Geoff Mulgan — Universities Are Failing at Mission-Led Research, Here’s How to Fix It. Research Fortnight. September 30, 2026. Warns of “mission-washing”: universities adopt the language of challenge-led research and bank the funding, while rewards still favour siloed, individual work. Proposes five fixes: choose challenges at the right granularity by blending top-down goals with bottom-up input; reshape incentives and recognise collaborative and design work; invest in shared collaborative spaces; build structured problem-solving partnerships, citing ARIA, DARPA, FROs and Denmark’s Alliance for Welfare Innovation; and evaluate for “double additionality”, showing missions beat business as usual.
  • Dan Recht, Ben Reinhardt & Jordi Cabana — Debate: National Labs vs. New Science Institutions. Macroscience. September 29, 2026. Reinhardt and Recht argue the DOE National Labs are structurally unsuited to the work FROs and X-Labs target and that new institutions will drive reform through competition; Cabana replies that publicly funded new institutions will face the same constraints and that metascience should study existing institutions as rigorously as it promotes new ones.
  • Mark Rubin — Transparency, Replication Rate Targets, and Direct Replications. Critical Metascience. September 28, 2026. Introduces Rubin’s Nature Human Behaviour article arguing there is no universal target replication rate: the right rate depends on a field’s epistemic goals, and the gap between expected (about 63%) and observed (about 50%) rates is too small to settle whether there is a “crisis”. Also covers Stephen John on the costs of transparency and a commentary arguing that science self-corrects mainly through better theory and measurement.
  • Danish National Research Foundation — Risk-Taking: Examples from Centers of Excellence. DNRF theme publication. September 2026. From board discussions with all Centers of Excellence in 2025–26. Centres say time and flexibility make risk possible, but early-career researchers on short contracts carry most of the risk, since a project without publications can cost them their career. Centres protect them by mixing high-risk and safer projects, extending contracts when risky projects fail and giving protected time. They ask the DNRF to recognise “productive failures” and to reconsider the year-five mid-term evaluation, which makes centres more cautious early on.
  • Lise Degn (CFA, Aarhus University) — Risk in Research Funding: Implications for Foundation Governance. Danish Centre for Studies in Research and Research Policy. September 2026. A research brief commissioned by the DNRF. It splits “risk” into seven kinds of uncertainty (scientific, conceptual, feasibility, investigator, implementation, strategic and portfolio) and finds that funders rarely take more risk simply by asking reviewers for bolder projects. Risk is governed through organisational design, combining three strategies: reducing uncertainty before funding, accommodating it through long and flexible grants, and spreading it across a portfolio of different instruments. It ends with questions for foundation boards, including whether they evaluate changes to their own assessment.
  • Leya Mohsin (FAS) — Are We Guessing Well or Knowing Well? Evidence Gaps in R&D Institutional Design. Federation of American Scientists. September 25, 2026. A companion to the FAS design exercise: today’s R&D institutions are “part inherited guesswork”, and the evidence for redesigning them is thinner than confident proposals assume. Metrics show the survivors (DARPA’s GPS, not its failed bets), findings from one agency or field may not transfer, and funding instruments get locked in by habit. Proposes documenting design logic as it happens, building comparative tools, and treating every founding design as a hypothesis with metascience units whose evidence actually reaches decisions.
  • Loren DeJonge Schulman, Erica Goldman & Leya Mohsin — What Survives the Trip From Blue-Sky to the Real World. Federation of American Scientists. September 24, 2026. Lessons from a July FAS workshop where seven groups each designed a new R&D institution from scratch. Recurring choices: coordination needs authority to make decisions stick, funding instruments should match a problem’s stage, public input should shape problem choice, and political durability is a founding design question. Peer review, which pushes funding toward safe research, is named as the unaddressed constraint on all of it, with NSF reauthorization flagged as the next policy window.
  • Stephen Curry (RoRI) — Meta Morsels 2. Go Meta! (Research on Research Institute). September 24, 2026. The second RoRI roundup: Alondra Nelson on why social science should consider refusing proprietary AI models, Giorgio Gilestro on open science under a flood of AI-generated text and the case for open-source hardware, and Stiglitz et al. proposing an International Panel on Inequality.
  • David Matthews — Is Metascience Being Co-opted?. Science|Business. September 24, 2026. Asks whether the White House’s Golden Age report turns metascience into cover for defunding universities: James Wilsdon warns of a “convenient hook” and a “political land grab”, while others argue researchers should report what they find regardless. Covers the X-Labs trade-off with basic grants, the evolution-or-revolution divide, and Europe, where the proposed Commission metascience unit has not materialised and Sam Bogerd fears the ideas will be “tainted politically”.
  • Stuart Buck, Josh New & Aishwarya Khanduja — Mandatory Reporting of Null Results. The Good Science Project. September 24, 2026. As the NIH and the White House signal interest in requiring null results to be reported, the authors argue for a tailored “report all meaningful results” rule: extend the clinical-trial model (preregistration, results within a year) to studies with a prespecified yes-or-no question, starting with preclinical efficacy, biomarker validation, genetic association, toxicology and protein-ligand binding research, and making NIH funding contingent on it. Hidden null results also mislead the AI tools now used to synthesise the literature.
  • Andrew Gerard — Matt Clancy on Science and the Returns to R&D. Macroscience. September 24, 2026. A writeup of the sixth IFP Economics of Ideas lecture on why the returns to public R&D are hard to measure (lags, spillovers, which costs to count) and how four empirical studies point to social returns well above one without agreeing on a number. Clancy also announces a five-year NBER “pop-up journal” on the question, funded with Coefficient Giving and the Sloan Foundation.
  • Industriens Fond — Giv plads til de skøre idéer. Industriens Fond. September 23, 2026. An interview with Lindsay Heil, DARPA’s director of international cooperation, after TechBBQ in Copenhagen: about 100 programme managers on time-limited contracts with a quarter replaced each year, no in-house labs, and contracts signed in months. Her warning to countries copying the model is that boards, funding committees and ministerial layers kill it; what works is a strong mission and autonomous programme managers.
  • Ben Reinhardt — Bubbling Cauldrons and Transient Strike Teams. Spectech. September 22, 2026. A scale-free model of healthy research ecosystems: “bubbling cauldrons” of small, researcher-directed work, temporary top-down “strike teams” aimed at one goal, and mechanisms to spot ripe ideas and staff them fast, a pattern Reinhardt traces from single labs to Bell Labs to national systems.
  • Sarah de Rijcke — Six Principles for Supporting Academic Freedom in Universities. Nature 657, 876–879. September 21, 2026. Leiden’s rector offers six tests for where legitimate steering of a publicly funded system ends and interference begins: ground academic freedom in the duty to produce reliable knowledge; protect questions whose value cannot be shown in advance; treat universities as democratic infrastructure; hold security and openness as one responsibility; protect the people who carry the risk of independent work; and return the trust on which autonomy rests. A measure passes if it strengthens independent scrutiny.
  • Amit Gupta — The Neo-Prime. Catalyze. September 21, 2026. Proposes a software-led “neo-prime” institution for moonshots, modelled on the prime contractors of the Manhattan Project and Apollo: master contracts with agencies, AI programme management to fund parallel approaches, and a network of labs to execute, with ARPA-H’s IGoR programme cited as a public analogue.
  • Ministry of Science and ICT, via Digital Today — South Korea Sets AI Research Ethics Guidelines for National R&D. September 20, 2026. A national guide treats hidden prompts planted in submitted files to evade review criteria or distort AI-assisted evaluation as research misconduct with statutory penalties, requires an AI-use disclosure form on R&D project reports, and bars evaluators from feeding evaluation materials into external AI services.
  • Clare Zhang — Researchers Skeptical of NSF Metascience Plan. AIP FYI. September 18, 2026. Former Science of Science programme directors Mary Feeney and Joshua Rosenbloom question NSF’s new metascience office: placed under the director’s office it risks becoming “a political activity, not a scientific activity”, mechanisms such as golden tickets and EAGER grants can bypass peer review, and the agency made a single science-of-science award in 2026 while dissolving the directorate that funds the field. NIH’s former extramural research head Mike Lauer asks how funding experiments on scientists’ careers can have informed consent.
  • Reuters, via US News — Scientists Challenge US NIH Grant Screening as Unconstitutional. September 16, 2026. A class-action suit backed by the ACLU and the Center for Science in the Public Interest alleges NIH screens grant applications against roughly 235 flagged keywords, including “gender” and “climate change,” to identify awards for termination or forced renegotiation.
  • Heleen Herbert (Dutch Ministry of Economic Affairs) — Septemberbrief Taskforce Toekomstige Welvaart en Vestigingsklimaat. September 15, 2026. The Dutch government’s letter to parliament sets out three instruments: €500m for a National Agency for Disruptive Innovation (NADI), a €375m round of the National Growth Fund, and a €3.3bn National Investment Institution. NADI will fund time-limited, problem-driven programmes in which independent programme managers steer competing R&D teams with milestone funding and wide discretion to scale up or stop. It is to be a state-owned company instead of an independent administrative body, to avoid public-sector pay caps and keep decisions fast, with a pilot alongside SPRIND in 2027 and full operation by mid-2028. Its success is to be judged at portfolio level, with room to fail. See also NADI’s press release.
  • Scholars at Risk — Free to Think 2026. September 15, 2026. Counts 382 attacks on higher-education communities across 59 countries and territories over the past year, naming the United States, Türkiye, and Georgia for partisan appointments to university governing bodies.
  • Teeny et al. — The “Synthetic Validity” Problem in AI-Generated Science. PNAS. September 15, 2026. Names a subtler research-integrity failure mode than fabrication: AI-generated results that are internally coherent and statistically plausible but ungrounded in any real underlying phenomenon, harder to catch with standard fraud-detection heuristics.
  • STAT News — Court Halts DHS Visa Limits on Foreign Scholars. September 14, 2026. A federal judge's preliminary injunction blocks the DHS rule ending indefinite “duration of status” for F-1 and J-1 holders, one day before it was due to take effect.
  • Wellcome Trust — Wellcome Discovery Awards Eligibility Update. September 2026. From 29 October 2026 the scheme accepts lead applicants only from the UK and low- and middle-income countries in Africa, South Asia, and South-East Asia, newly excluding Danish institutions as leads.
  • AIP FYI — DOJ Official Queried a Dozen Journals on Partisan Views in 2025. September 9, 2026. Newly released records show a DOJ official sent near-identical letters in April 2025 to at least eleven journals, including JAMA, NEJM, and Science, accusing them of partisanship and demanding explanations of editorial process.
  • NIH — NOT-OD-26-088: Request for Information on Reporting Peer Review Outcomes. August 2026. Proposes replacing numerical impact scores and percentiles with broad tiers, “Most Competitive,” “Competitive,” and “Not Discussed,” aligned with NIH's Unified Funding Strategy; comment period runs through October 13, 2026.
  • Denmark Digest — DTU Fires Climate Researcher Henrik Svensmark. July 2026. DTU dismissed the veteran space and climate physicist after nearly three decades, saying only that it is downgrading atmospheric physics; Svensmark told Berlingske he could not rule out political considerations.
  • Stephen Curry (RoRI) — Meta Morsels 1. Go Meta! (Research on Research Institute). September 17, 2026. The first in an occasional RoRI series rounding up notable metascience writing, opening with Daniel Hook on AI-era “decollaboration,” Henry Farrell on how social science should think about AI, and Christopher Smith on why science policy needs context over metrics.
  • Paul Litvak — We Must Understand. In One Lifetime. September 17, 2026. Makes the case for building an open evidence graph, a shared public record linking scientific claims to the evidence for and against them, as a defence against “feudal science” and the erosion of the cognitive commons, while weighing the risk that it also accelerates dangerous AI.
  • Martin Juhl — Forskerbevægelsen: Giv forskere faste beløb til at forfølge egne idéer. Uniavisen. September 16, 2026. Responding to the CFA mapping of Danish research funding, the academic-freedom association Forskerbevægelsen proposes an annual research budget (an “annuum”) of about DKK 100,000 for each of Denmark’s roughly 7,000 permanent associate and full professors, around DKK 700m a year, to test ideas before applying for large grants. It would keep competitive funding, ease application pressure, and shift the balance at the University of Copenhagen, where external income grew from about DKK 3bn to 4.4bn between 2015 and 2025 and made up 38% of revenue.
  • Forskerforum — Fire ud af fem forskere ville ændre deres forskning med fri finansiering. Forskerforum (DM). September 2026. Danish coverage of the CFA Forskningslandskabet report: 79% of researchers say they would shift their research interests given two years of unconstrained funding, and three in four say current grant formats constrain how they would ideally work.
  • Brian Stone (NSF) — NSF Leadership in a New Golden Age of Science. September 10, 2026. NSF’s acting director sets out a portfolio approach that matches the funding mechanism to the nature of the work: longer awards for ambitious research, short fast grants for exploratory ideas, milestone funding, prizes and challenges (“we will specify the objective, never the method”), review experiments such as reviewer “golden tickets” and distributed review, and new institutional forms beyond X-Labs. A new Office of Metascience will run prospective experiments on funding and review, report the effectiveness of mechanisms publicly and publish postmortems for positive and negative results, aiming to double scientific productivity per federal dollar by 2036. Four cross-directorate Frontier Initiatives are also launched.
  • UKRI — UKRI Modernises Grant Assessment for the Age of AI. September 10, 2026. An eight-measure overhaul, effective October 2026, to speed funding decisions and respond to generative AI's effect on both grant writing and peer review, part of a July 2026 pledge to halve grant-processing time by 2031.
  • Andrew Gerard — Kyle Myers on the Direction of Science. Macroscience. September 10, 2026. A writeup of an IFP Economics of Ideas lecture on how funders steer science: the economics of grants, prizes, patents and contracts, disease-group lobbying and soft earmarks at the NIH, how review noise distorts grant budgets, and the cost of steering scientists toward new directions.
  • Stuart Buck & Kushal Kadakia — Investing in the Science of Science: What Medicare Can Teach the NIH About Experimentation. The Good Science Project. September 10, 2026. A reprint of a 2022 Health Affairs piece, newly relevant as agencies stand up metascience offices, arguing the NIH needs a statutory mandate and protected funding to experiment on its own grantmaking, modelled on the Center for Medicare and Medicaid Innovation.
  • Orin Herskowitz, Kyle McEneaney & Kumar Garg — Building the Bridge from Innovation to Impact: Philanthropy’s Unique Role. SCI PHI (Renaissance Philanthropy). September 10, 2026. Argues that promising innovations die between lab result and patient benefit, where no market incentive funds validation, regulation and adoption, and that philanthropy, uniquely free of market and statutory constraints, should act as a systems integrator across the whole innovation-to-impact bridge, with examples from the SMA Foundation, Rare As One, Convergent Research’s FROs and the Cystic Fibrosis Foundation.
  • Mads Schäfer Bak — The Structural Dynamics of Inequalities in the Research System. Zenodo (preprint), Danish Centre for Studies in Research and Research Policy (CFA). September 7, 2026. Drawing on 20 expert interviews and a systems-thinking lens, identifies four interconnected sources of inequality in research: structural incentives and competition, prestige and networks, geography, and the publishing system. Inequalities in one domain spill over into others through cumulative advantage, so interventions aimed at single components tend to leave the underlying dynamics intact.
  • Mark Rubin — Is Replication Rare? Registered Report Timelines, Open Science and Trust, P-Hacking and Publication Bias. Critical Metascience. September 7, 2026. A roundup arguing replication looks rare or common depending on definition (direct replications are rare, conceptual ones and unlabelled “micro-replications” are not), alongside new work on registered-report timelines, whether open-science cues raise public trust, and how fast versus slow p-hacking interact with publication bias.
  • Lars Bo Nielsen (Villum Fonden) — Vi har brug for flere veje fra forskning til handling. Science Report. September 5, 2026. Villum Fonden’s director argues Denmark must both protect curiosity-driven basic research and build more bridges from discovery to application, and says the foundation is raising innovation in its grant strategy, citing Spin-outs Denmark (a new DKK 115M phase) and the Green X Accelerator (DKK 80M). Innovation, he argues, is society’s ability to turn knowledge into solutions, more than a measure of growth.
  • Andrew Gerard, Jenn Gustetic et al. — What We’ve Been Reading, Vol. 2. Macroscience. September 4, 2026. The IFP metascience team’s reading roundup, spanning the fight over NSF social-science funding, a “tragedy of the cognitive commons” as AI erodes entry-level expertise, AI agents and the validation bottleneck in science, cancer vaccines, and why AI RCTs must evolve to stay useful.
  • Kyle Cranmer, Neil D. Lawrence, Jessica Montgomery & Denis Thérien — AI for Science: Reframing AI’s Role in Discovery. RSS Data Science and AI. September 2026. A framework for how AI actually contributes to science, organised around task capabilities, workflow integration and domain constraints, opening deeper questions: whether AI-built knowledge counts as scientific understanding if it is impenetrable to humans, what it would take to move AI beyond pattern-matching toward causal reasoning, and what institutional changes responsible adoption requires. Opens a call for papers from RSS Data Science and AI.
  • Phillip J. Haubrock et al. — How Contemporary Academic Structures Constrain Scientific Creativity and Hold Back Early-Career Researchers. Nature Human Behaviour. September 3, 2026. A Perspective co-authored by more than fifty researchers arguing that contemporary academic structures and incentives suppress scientific creativity and disproportionately hold back early-career researchers.
  • Ben Reinhardt — The Hardcore Institute of Technology. Spectech. September 3, 2026. Prompted by the executive order for a US Space Academy, Reinhardt pitches a new kind of technical institution: an industrial research lab rather than a university, forging elite scientists and engineers by taking on journeymen and apprentices who learn by doing serious pre-commercial work, tuition-free and salaried, with no grades or degrees.
  • Amit Gupta — From Bits to Atoms: Building the Infrastructure for the New Scientific Era. Catalyze. September 2, 2026. Argues the binding constraint on AI-driven discovery is physical execution, and calls for a networked R&D infrastructure to commission, fund and verify experiments across labs, capturing the “dark data” of negative results and tacit judgement that published literature omits.
  • NACo — Congress Passes Short-Term Funding Bill Keeping the Federal Government Funded Through December 11. September 2, 2026. The stopgap bill, passed 370-48 in the House and signed into law by President Trump, bars OMB from finalising its rule letting political appointees override grant peer review before December 11, confirming in statute a delay the Senate had only voted on.
  • Federico Bignone, Ammon Salter & Stefano Baruffaldi — Examining the Innovation Funding Lifecycle: Selection, Reapplication and Firm Outcomes in Innovate UK Funding. Innovation & Research Caucus. September 2, 2026. Using Innovate UK records from 2017 to 2024 and a regression-discontinuity design, finds entry into the funding system is highly selective, rejection rarely drives firms away, and every £1 of grant is associated with about £1.68 in additional private venture funding, making the grant chiefly a lever to mobilise private capital.
  • Paul Karon — Catalyze: A New Funding Model to Connect Donors and Scientific Research. Inside Philanthropy. September 1, 2026. Profiles Catalyze, Sandeep Patel's new platform (with Amit Gupta and David Bowen) that leverages the peer review NIH and NSF already do, surfacing highly-rated but unfunded proposals so philanthropists can find and back vetted science via AI-assisted search. The built version of the "science-vetting machine" idea, launched with 80 to 100 projects.
  • Nelson Ndahiro & Reinvent Science — Field Report: How I Let AI Do Some of My Thinking as an Experimental Scientist. Reinvent Science. September 1, 2026. A wet-lab scientist describes representing his lab as code so coding agents can design plates, cost assays and catch pipetting errors, offloading the operational drudgery that drives talent from the bench.
  • Andrew Gerard — Pierre Azoulay on Open Science as an Economic Institution. Macroscience. August 28, 2026. A writeup of an IFP Economics of Ideas lecture on science as an economic institution: Merton’s CUDOS norms and the Matthew effect, why firms let scientists publish (Stern’s compensating differential), Pasteur’s Quadrant, and evidence that funding people rather than projects buys more risk-taking.
  • Zachary Pirtle, Mahmud Farooque & Michael Bernstein — Metascience for Public Missions: Principles for Agency Implementation. Engineering Outcomes. August 28, 2026. As agencies face an October 19 deadline to propose metascience units under the Golden Age report, the authors urge them to draw on the Science of Science and Innovation Policy (SciSIP) tradition, mixed methods and public-value mapping, so the units become a learning capability rather than a compliance exercise.
  • Gemma Derrick, Amy Devenney & Lizzie Ville — Universities Rarely Use Research Evidence in Their Decision-Making, So Why Do They Expect Anyone Else To?. LSE Impact. August 27, 2026. From the Wellcome-funded COMET project (55 decision-makers at six UK universities): universities produce the evidence other institutions rely on but rarely draw on externally validated research-on-research evidence for their own research-culture and strategy decisions, leaning instead on internal data and competitor benchmarking.
  • Columbia Postdoctoral Workers (UAW Local 4100) — Columbia postdocs authorise a strike after contract talks stall. August 27, 2026. With the contract that set a $70,000 minimum postdoc salary expired on June 30 and six months of talks without a deal, over 90% of roughly 850 voters (the union’s largest turnout) authorised a strike over demands including an $80,000 minimum.
  • Kaare Nielsen — Vi ønsker gennembrud, men organiserer forskningen som vi plejer. Science Report. August 26, 2026. The POMUS founder argues research organisation is itself a policy variable and calls for a permanent Danish experiment-pool to test new institutional models (long grants, FROs, DARPA-style program managers).
  • Mark Rubin — Mindless Transparency. Critical Metascience. August 24, 2026. An argument against maximal openness as a default, since transparency is always a selection and the aim should be disclosure of the relevant, not of everything.
  • Kevin Munger — Introducing the K-index. Never Met a Science. August 24, 2026. A send-up of academic metrics arguing that the proliferation of boutique, for-profit indices de-naturalises measurement itself, since once many are possible no single one can pass as neutral truth.
  • Dan Garisto — Exclusive: NSF Set to Issue Lowest Number of New Grants in Four Decades. Nature. August 20, 2026. NSF will award about 30% fewer new grants this fiscal year, roughly 6,100 and the lowest since the early 1980s, mainly because about $1 billion of its $8.8-billion budget is held in a central account.
  • Peter Gluckman — Modern Science: Geostrategic, Technological and Societal Drivers of Change. International Science Council. August 19, 2026. The ISC president argues science must not respond passively to the forces reshaping it, and calls for rethinking its self-governance, incentives and social contract.
  • Science Works — Reforming UKRI Eligibility. Science Works. August 19, 2026. An open letter to UKRI arguing that eligibility rules, not merit, keep new research organisations out of public funding, and proposing reforms so a track record with trusted funders can substitute for the three-years-of-accounts rule.
  • Frederik Hertel — Brud med forlagenes magt kræver et opgør med meriteringssystemet. Forskerforum (DM). August 17, 2026. Argues that fighting publishers over price and copyright is not enough, because their power rests on the academic merit system that makes journal rankings the currency of a career.
  • NSF — NSF Announces $1.5B for Foundational Research to Drive Scientific Breakthroughs. NSF. August 17, 2026. NSF released 12 new funding opportunities and a shift to a portfolio-based approach, framing many calls as structured experiments it will assess and scale, in effect running metascience on itself.
  • James Evans — Why AI Systems Are Most Useful as Designers of New Scientific Tools. Nature. August 17, 2026. A review of Alexander Krauss’s The Engine of Scientific Discovery, drawing the implication that an AI’s most important contribution to science may be helping design the next generation of scientific instruments.
  • Denisa Lepadatu & Ed Boyden — The Ground Truth Institute. Universe Conquering. August 17, 2026. A proposal for a Ground Truth Institute to “engineer” scientific revolutions in biology by unifying tool-building, data-gathering and commercial deployment so downstream returns fund the high-risk pioneers upstream.
  • Paul Litvak — How Well Does AI Peer Review Work?. In One Lifetime. August 10, 2026. Litvak planted 100 known errors into 10 psychology papers and ran them through AI review tools: the best single system caught 71, but pooling every system caught 93.
  • Sandeep Patel, Stuart Buck, Ian Banks & Amit Gupta — America Already Built the World’s Largest Science-Vetting Machine. Macroscience. August 11, 2026. Argues the peer review of the many top-quartile proposals NIH and NSF leave unfunded should be treated as public infrastructure, via an opt-in platform letting scientists share those reviews with private funders.
  • Jenn Gustetic (IFP) — Ambitious American Science Requires Reliable and Effective Federal Funding. Institute for Progress. August 11, 2026. IFP and the Niskanen Center file a joint comment opposing OMB’s grant rule, arguing universal discretionary termination and political pre-issuance review would destabilise federal funding and demote the expert peer review empirically tied to better outcomes.
  • Max Bennett — When the Best Decision Is No Decision: The Rise of Randomization in Grant Funding. Nature. August 2026. A survey of the growing use of partial lotteries in research funding, spanning the Swiss National Science Foundation, the British Academy, and eleven trials catalogued by the Research on Research Institute since 2022.
  • Research Professional News — UK’s Aria ‘Closer to Darpa Model’ than Germany’s Sprind. Research Professional News. August 2026. Reports a master’s thesis by Christopher Schade (European University Institute) arguing that DARPA rests on organisational flexibility and a focus on high-risk transformative work. ARIA copies both more fully. SPRIND decides collectively, through directors and innovation managers seeking consensus, while DARPA has three layers of hierarchy, no labs of its own and programme managers on three to five year terms.
  • Søren Riis Paludan, Eske Willerslev, Susanne Mandrup m.fl. — Danmark skal investere i grundforskning, der skaber fremtidens gennembrud. Altinget. August 13, 2026. Sixteen leaders of Danmarks Grundforskningsfond Centers of Excellence call on the government to recapitalise the foundation, defending concentrated, long-horizon basic research as the source of breakthroughs.
  • Retraction Watch — Biology Journal Pulls 100+ Articles and Counting From Special Issues for Peer-Review Manipulation. August 12, 2026. The International Journal of Biological Macromolecules has retracted more than 120 papers since June 2026, almost all from guest-edited special issues, citing "systemic, coordinated, and widespread manipulation of the peer-review process."
  • Retraction Watch — BMJ Public Health Retracts COVID-19 Vaccine Excess-Mortality Study Cited at Senate Hearing. August 11, 2026. A 2024 excess-mortality paper invoked to link COVID-19 vaccines to deaths, including at a June 2026 US Senate hearing, was retracted "on the grounds of misinformation on causes of death and limited original work."
  • GeekWire — Gates Foundation Gives $540 Million to UW Health Institute, the Largest Gift in University History. August 10, 2026. A ten-year commitment to UW's Institute for Health Metrics and Evaluation, expanding the Global Burden of Disease study's coverage from 925 to nearly 5,000 locations worldwide.
  • STAT News — NIH Plans to Block International Researchers on Visas From Accessing Key Funding Avenue. August 7, 2026. NIH intends to restrict its K99/R00 Pathway to Independence award, the main bridge from postdoc to independent PI, to US citizens and permanent residents, closing a key route for international researchers on visas.
  • Science/AAAS — China Now Tops the U.S. in R&D Spending by One Key Measure. August 7-8, 2026. Japan's NISTEP indicators find China's 2024 R&D spending reached about $615bn, overtaking the US for the first time and making China the leader on all three core measures.
  • Unite the Union — Staff at World-Leading Institute of Cancer Research to Strike Over Pay. Updated August 7, 2026. ICR's indefinite strike, the charity's first industrial action in its 117-year history, running since 3 June 2026, escalated into Acas-mediated talks.
  • Adam Glen — The State Is the Customer. Adam’s Substack. August 6, 2026. Argues the UK’s translational intermediaries, the Catapult network and ARIA’s Activation Partners, have become the customer of first resort for the technologies they fund, crowding out real market formation, and calls for funding founders directly through procurement vouchers and fellowships rather than institutions.
  • Adam Mastroianni — Incentives Are for Losers. Experimental History. August 5, 2026. Argues that “bad incentives” has become an excuse, in science as elsewhere. Since Goodhart’s law means every incentive gets gamed in the end, we should keep revising incentive systems, but researchers also need convictions of their own, starting with asking whether their work is meant to solve problems or to publish papers.
  • TechCrunch / Axios — Jeff Dean and Other Top AI Researchers Are Leaving Google to Launch Their Own Startup. August 5, 2026. Google's chief scientist Jeff Dean left after 27 years, with Sanjay Ghemawat, Oriol Vinyals and Quoc Le, to found Discovery Loop, aiming to automate the experimental loop of scientific research.
  • Times Higher Education — National Security Used to Curb Academic Freedom, Report Says. August 5, 2026. A BRISMES Committee on Academic Freedom report documents 28 cases where governments and universities invoked counter-terrorism, national-security, or community-safety justifications to restrict academic expression.
  • NSF — New NSF State and Regional AI Infrastructure Hubs Will Power AI. August 4, 2026. A $100 million programme funding up to ten regional consortia to build shared AI and scientific-compute infrastructure for researchers outside frontier institutions.
  • Science/AAAS — NIH Director Bhattacharya Kills More Health-Disparities Grants. August 4, 2026. NIH terminated a $9.6 million University of Pittsburgh grant studying cognitive aging and Alzheimer's risk among Black residents of low-income neighbourhoods, citing its health-disparities framing.
  • EIS — GCU Strike, Day Four. August 4, 2026. UCU and EIS members at Glasgow Caledonian University took a fourth strike day against proposed cuts of up to 100 posts, with management still not ruling out compulsory redundancies.
  • US Department of Education — Secretary McMahon Issues National Call to Action to University Presidents and Governing Boards. August 3, 2026. An open letter asking every US institution to publicly answer seven questions on admissions, viewpoint diversity, research security, and national priorities by year-end 2026.
  • Valeriy Ginzburg — The Kratsios Report and the Future of Science. Valeriy Ginzburg (Substack). July 28, 2026. An industry scientist’s critique of Science: A New Golden Age. Its call for “deliberate portfolio construction” resembles central planning that assumes computers can predict where breakthroughs come from; the “linear model” it rejects is really a spiderweb of cross-links that new AI-centred institutes might cut; and industrial basic research has shrunk since the 1980s, contrary to the report. He agrees the link between graduate training and careers needs rethinking, and asks who decides when AI is used to settle paradigm changes.
  • NSF — Prohibition on Collaborations with Restricted Entities. Dear Colleague Letter 26-022. July 8, 2026 (effective FY2027). Bars NSF-funded senior and key personnel from appointments with, support from, or collaboration with any entity on seven federal restricted-party lists.
  • South China Morning Post — Pentagon Expands Blacklist to Include China's Top Civilian Universities. July 23, 2026. DoD's Section 1286 list expands to 130 institutions, the largest single-cycle expansion in its history, barring elite civilian universities like Fudan and Shanghai Jiao Tong for the first time.
  • Mark Hahnel — Do we actually need to write research papers at all?. OpenResearch.wtf. July 31, 2026. Argues that as AI strips the friction out of analysis and write-up, the research paper stops being the point, leaving certification, trust and provenance as the scarce product.
  • Hiya Jain, Stuart Buck, Aishwarya Khanduja & Analogue — A Taxonomy of R&D Orgs: What Is New, What Is Missing?. Analogue Press / Good Science Project. July 31, 2026. Maps the design space of research organisations along nine parameters so funders and researchers can find unfilled "holes" in the ecosystem the way Marblestone and Rodriques named the FRO-shaped one.
  • Jacob Robertson & Cate Young — What Does it Take to Deliver a New Golden Age?. FAS. July 30, 2026. A Federation of American Scientists response to the OSTP Science: A New Golden Age report, arguing the vision is serious but that the next era hinges on implementation into agency structure, funding models, and hiring.
  • Stuart Buck (Good Science Project) — Moonshots, the Impact of Science Cuts, the Genesis Mission, and more. The Good Science Project. July 25, 2026. A metascience reading roundup surfacing the pieces below and announcing the Metascience 2027 Conference (Montreal, June 14-16, 2027).
  • Daniel P. Gross, Bhaven N. Sampat & Hansen Zhang — Retreating from Science: The Long-Run Effects of the 1970s U.S. Military Disinvestment from University Research. NBER Working Paper 35491. July 2026. Studies the DoD's roughly 70% real-dollar cut to university research contracts after 1970, finding lasting declines in publications, PhD training, and downstream innovation in more DoD-exposed fields.
  • Jasper Doeninghaus — An Evidence-Based Account of the ARPA Model in Europe. M-RCBG Associate Working Paper 281, Harvard Kennedy School. July 2026. Uses SPRIND to ask which ARPA design features drive outcomes and which transfer to European institutions, as SPRIND is evaluated under its 2023 freedom law and the Netherlands and France look to it for guidance. Notes that success stories alone cannot show which features worked.
  • Frances Jones — Europe’s Emerging Arpa-Style Network ‘Needed for New Industries’. Research Professional News. July 2, 2026. National breakthrough agencies are emerging across Europe as an alternative to a single EU-level ARPA, following the Franco-German task force’s call for a French agency modelled on SPRIND.
  • Leya Mohsin — The Science of Moonshots: Using Evidence to Design Transformative Initiatives. FAS. July 2026. Argues "moonshot" has become a tired marketing term, and recovers what the genuine article requires: a clear end goal and path to it, and governance concentrating authority with accountability to a defined endpoint.
  • Dan Turner-Evans — Picking the Right Challenges for Genesis Mission. Institute for Progress. July 2026. Argues the Department of Energy's Genesis Mission spreads across far too many challenges (26) to function as a real moonshot, and should narrow to a small set of goals where the federal government adds unique value.
  • M. Anthony Mills — Who Should Decide How to Use Taxpayer Dollars to Fund Scientific Research?. Issues in Science and Technology. July 2026. Responds to the White House's moves to fire the National Science Board and let political appointees cancel grants, arguing the federal research enterprise should remain democratically accountable.
  • Noah Smith — We asked too much of the American university. Noahpinion. July 24, 2026. Argues the American university became the country's last broadly functional institution, overloaded by mass enrolment and the social-unification roles that churches, the military, and mass media shed after 1970.
  • Carlos Arleo — Can AI tell if your research is actually new?. Go Meta! (RoRI). July 24, 2026. Describes building the Disruption & Novelty Indicator (DNI), an AI novelty detector that earned a special mention in the UK Metascience Unit's 2025-26 Novelty Indicators Challenge.
  • Mark Rubin — Feminist Metascience and Feminist Open Science. Critical Metascience. July 24, 2026. Introduces a forthcoming Gender & Society special issue defining feminist metascience as the study of how gender and power relations shape how knowledge is produced, evaluated, and corrected.
  • The White House — White House Commits More Than $5 Billion to Expand the Genesis Mission. July 22, 2026. Fifteen-plus federal agencies commit funding, datasets, and facilities, and DOE selects 278 AI-for-science projects across 50 states.
  • Hewlett Foundation — Hewlett Foundation Announces New $100 Million Emerging Technology and Security Initiative. July 22, 2026. A five-year, $20M/year initiative targeting governance gaps between fast-moving AI, biotech, and quantum research and slower regulatory frameworks.
  • Inside Higher Ed — Portland State Will Lay Off 36 Employees, Close 2 Departments. July 21, 2026. A final reduction plan cuts 34 full-time faculty and closes two departments, driven by a $35 million two-year budget shortfall.
  • Roll Call — Critics Mount Offensive Against Change to Federal Grant Rules. July 20, 2026. OMB's Uniform Guidance overhaul closed its comment period on July 13 with 496,769 submissions, about 95% opposed; OMB is reportedly now targeting finalisation by October 1, 2026.
  • Bloomberg / Higher Ed Dive — States Win Suit Over Trump Administration Termination of Grants. July 17, 2026. US District Judge Indira Talwani granted summary judgment to a coalition of 23 states, ruling that standard federal grant terms do not permit agencies to terminate awards based on program priorities identified after the grant was made.
  • Research on Research Institute — The Future of Peer Review. July 2026. Capstone report of RoRI's three-year peer review project (39 stakeholder interviews plus literature review) concludes current peer review is unsustainable under hyper-competition supercharged by AI, and argues for systemic reform over per-project patches.
  • Baumann et al. — Stop Automating Peer Review Without Rigorous Evaluation. Presented at ICML 2026, July 8, 2026. Finds AI-generated peer reviews show a "hivemind effect" (higher cross-paper similarity than human-involved reviews) and are gameable via "paper laundering," an LLM rewrite with no substantive change measurably raising AI-reviewer scores.
  • Kristin Ellis — Science Is So Back. Contents May Shift. July 27, 2026. A reading of the OSTP Science: A New Golden Age report, calling it the most serious rethink of US research since Bush's 1945 report even as it arrives after deep cuts.
  • Andrew Gerard, Jenn Gustetic & Matthew Esche — Science Agencies Need Metascience Units. Institute for Progress (also cross-posted to Macroscience). July 23, 2026. Argues federal science agencies should stand up empowered metascience units: internal think-tank teams that run controlled experiments and mine administrative data to learn how to fund and conduct science better.
  • Michael Kratsios (OSTP) — Science: A New Golden Age. White House Office of Science and Technology Policy. July 22, 2026. A 123-page report billed as the first comprehensive rethink of the US science enterprise since Vannevar Bush's Science: The Endless Frontier, recommending support for individual scientists, diversified funding, agile research organisations, and agency metascience units.
  • NSF & Astera Institute — Astera partnering with the National Science Foundation. Astera Institute / NSF. July 2026. NSF is investing $400M over four years to build a national network of 20 AI-enabled autonomous Programmable Cloud Lab nodes, in partnership with Astera, as part of the Genesis Mission.
  • Reinvent Science — The Scientific Literature is Poisonous to LLMs. Reinvent Science. July 21, 2026. Argues the modern scientific literature is uniquely toxic training data, since scientists judge reliability through informal social networks that AI agents cannot access.
  • Tom Kalil — Kalil's Corner: The Role of Philanthropy in Identifying and Pursuing AI Moonshots. SCI PHI (Renaissance Philanthropy). July 21, 2026. Argues philanthropy should fund the AI moonshots the market under-invests in: public goods like open datasets and benchmarks, fundamental science, and applications serving low-income populations.
  • Department of Homeland Security — Establishing a Fixed Time Period of Admission and an Extension of Stay Procedure for Nonimmigrant Academic Students and Exchange Visitors. Federal Register. July 17, 2026. Final rule ending the indefinite "duration of status" for F-1 and J-1 visa holders, replacing it with a fixed admission period of up to four years, effective September 15, 2026.
  • Bloomberg — Warren Buffett Ended Gifts to Gates Foundation Amid Jeffrey Epstein Scrutiny. July 14-15, 2026. Buffett excluded the Gates Foundation from his 2026 mid-year gift for the first time since 2006, redirecting roughly $6 billion in Berkshire Hathaway shares to family foundations.
  • World Weather Attribution — Fossil fuel emissions have rapidly worsened European heatwaves in just a few decades. July 13, 2026. A rapid attribution study finding the June 2026 heatwave's temperatures would have been "virtually impossible" without human-caused climate change.
  • Christian Hendriksen (via Altinget) — Forsker: Københavns Universitet undergraver den ambitiøse danske forskningsstrategi. Altinget. July 2026. The University of Copenhagen's Faculty of Science and Health is cutting roughly 100 professor-equivalent posts even as the university received DKK 148.8 million in extra 2026 research-strategy funding.
  • CNBC / TechCrunch — John Jumper to leave Google DeepMind for Anthropic and Google Gemini co-lead Noam Shazeer leaves for OpenAI. June 18-19, 2026. Within a day of each other, AlphaFold co-creator John Jumper left DeepMind for Anthropic and Transformer co-author Noam Shazeer left Google for OpenAI.
  • Australian Research Council — ARC Open Access Policy – 2026 Update. Effective July 1, 2026. Requires journal articles and peer-reviewed conference papers from newly opened grant schemes to be openly accessible immediately upon publication under a CC BY licence, with no embargo allowance.
  • Astera Institute — 2026 Hidden Science Competition. Astera Institute. July 2026. A follow-on competition targeting the useful science that gets done but never shared, offering life-sciences graduate researchers $10,000 contracts to publish null results, dead-end datasets, and code openly.
  • Fahd Husain, Antonia Theodorakopoulou, James Custer & Tom Kalil — Charting AI's Role in Scientific Discovery. Renaissance Philanthropy. July 15, 2026. A landscape study of where AI is becoming a structural enabler of scientific discovery, arguing progress is broad but shallow and the binding constraint is shifting from model capability to data infrastructure.
  • Conor Griffin & Don Wallace — Conjecture Machines: AI Agents and the New Validation Bottleneck in Science. Google DeepMind. July 15, 2026. Argues AI agents are becoming "conjecture machines" that make hypotheses abundant and cheap while validation stays physical, institutional, and slow, so refutation becomes the binding constraint.
  • Andy Stirling — Illuminating Neglected Uncertainties: How Metascience Can Help Rather Than Hinder Democracy. Go Meta! (RoRI). July 15, 2026. A critical-metascience essay arguing the field's hardest problems come less from too little evidence than from asking too little about what evidence cannot settle.
  • Alex Shintaro Araki — The Race to Allocate One Trillion Dollars in Philanthropic Capital. Alex's Substack. July 14, 2026. Argues imminent frontier-AI-lab IPOs will mint thousands of new millionaires holding over $1 trillion, and that philanthropy is unprepared, with absorptive capacity and talent the binding constraint.
  • Daniel Correa — Scaling Team Science Is the Important Experiment We Need. FAS. July 14, 2026. The FAS CEO makes the case for the NSF's X-Labs (up to $1.5 billion) as a valuable experiment that scales the philanthropically de-risked Focused Research Organization model of team-based, goal-oriented science.
  • Avery Orrall — 'The exploitation still remains': Stats journal associate editors resign over $3,000 publishing charge. Retraction Watch. July 9, 2026. Eighteen of 160 associate editors at Springer Nature's Statistics and Computing resigned after the journal announced a mandatory $2,990 article-processing charge, calling it "irreconcilable with our vision of science."
  • Nature — Nobel-winning chemist leaves US to direct AI materials lab in China. Nature. July 9, 2026. Chemistry Nobel laureate Omar Yaghi left UC Berkeley to lead a new AI-assisted materials-discovery institute at Tsinghua University, citing US grant cuts, joining a documented pattern of senior scientists (including Charles Lieber) moving toward China.
  • United States Senate Committee on Appropriations — Sen. Collins Asks OMB to Withdraw Parts of Grant Rule, Extend Comment Period. July 7, 2026. Appropriations Chair Susan Collins asks OMB to withdraw the provisions of its proposed grant-review rule affecting scientific and biomedical research and to extend the comment period 90 days.
  • Novo Nordisk Foundation — No easy answers: European research consortia take on major challenges. July 10, 2026. Nine 2026 Challenge Programme grants worth up to €80M over six years, the first cycle since the scheme opened to European main applicants, with seven of nine consortia led outside Denmark.
  • Noah Smith — MAGA's Attack on Science Is Even Worse Than It Looks. Noahpinion. July 10, 2026. Argues the postwar US scientific enterprise was an engine of American preeminence now under an unprecedented Trump-administration attack of funding cuts and ideological review.
  • Aishwarya Khanduja & Stuart Buck — Seeing Like a Gardener: A New Framework for the Public Funding of Science. Analogue Press (originally for Liberal Currents, The Reconstruction Papers). July 10, 2026. Applies James C. Scott's Seeing Like a State to US science policy, arguing the state should act less like a central planner and more like a gardener cultivating a diverse ecosystem of organisational forms.
  • Mohana Basu — Think preprints are unreliable? Analysis of 70,000 studies might change your mind. Nature. July 10, 2026. A bioRxiv study comparing 72,644 biomedical preprints with their published versions found about 90% of main conclusions unchanged or only minorly revised after peer review.
  • Analogue, Aishwarya Khanduja & Stuart Buck — Studying Inquiry: Ideas for Metascience Experiments. Analogue Press. July 6, 2026. A catalogue of 32 concrete, testable experiments for how science is funded and organised, in the spirit of Azoulay's call to "turn the scientific method on ourselves."
  • Retraction Watch — Computer science society creates new research integrity role to address case backlog. Retraction Watch. July 2, 2026. ACM is hiring its first director of research integrity after a five-year rise in misconduct allegations left a backlog of cases taking a year or more to resolve.
  • Science — NIH likely to award fewer grants as it races to spend its 2026 budget. Science/AAAS. June 29, 2026. As of June 20, NIH had made 27,541 awards worth $16bn against a $47.2bn FY2026 budget that must be spent by September 30, with the grant success rate falling from 21% to 11%.
  • PBS NewsHour / American Enterprise Institute — College endowment tax is leading to hiring freezes and could mean cuts in financial aid. PBS NewsHour. Effective July 1, 2026. A new tiered US endowment excise tax took effect, estimated to cost Harvard about $368M/year and Yale, Princeton, and Stanford comparable sums.
  • Council of the European Union — Council agrees negotiating position on FP10. June 26, 2026. The Council adopted its partial negotiating mandate for Horizon Europe's successor programme (2028–2034) at €167.9 billion, below the Commission's €175 billion proposal, ahead of negotiations with Parliament.
  • Caroline Fry & Andrew Gerard — Science for Development: What Comes Next?. Lauren Policy. June 24, 2026. Reflections from a Washington workshop arguing the returns to LMIC R&D are high but the evidence on what works is fragmented, under-visible, and weakly connected to decision-making.
  • The White House — Ushering in the Next Frontier of Quantum Innovation and Securing the Nation Against Advanced Cryptographic Attacks. June 22, 2026. Two executive orders: one launches a federal effort (QC-ADDS) to deliver at least one application-scale quantum computer to a DOE facility for open scientific use; the other mandates federal migration to post-quantum cryptography by 2030–31.
  • Pierre Azoulay, Raffaella Sadun & Daniela Scur — Before the Exodus? Young Scientists and the Future of US Science. NBER Working Paper 35330. June 2026. A survey of 916 young biomedical scientists finds only about half now say they are likely to remain in academia, down 22 points in six months, evidence of a possible exodus of future PIs.
  • Norwegian Ministry of Education and Research — Regjeringen vil endre norsk forskning: Vil ha mer innovasjon og ta forskningen raskere i bruk. Regjeringen.no. June 19, 2026. Norway starts work on a new long-term plan, now for research and innovation rather than research and higher education, to be presented in autumn 2027. Minister Sigrun Aasland says research does not contribute enough to competitiveness, welfare and security: the plan will concentrate on especially strong and promising groups, give top researchers time for long-term work, raise business R&D and tighten links between research and industry. It replaces the 2023–2032 plan, which combined research and higher education around three goals, six thematic priorities and two national missions. On 20 October the ministry invited input to the plan.
  • Pierre Azoulay — In (Qualified) Defense of the Research Project Grant. Science. June 11, 2026. A skeptical response to the reform push to move from investigator-initiated project grants toward independent, block-funded institutes, defending what the research project grant gets right structurally. "Wary, not opposed."
  • Søren Barlebo Rasmussen & Bo Jellesmark Thorsen — Universitetet i bevægelse: samfundskobling, strategisk prioritering og nye spændinger i forskningen. Mobilize Strategy Consulting. June 5, 2026. A Danish essay arguing the classical research model is not in crisis but in motion, pushed toward stronger societal coupling by systemic challenges and more strategic, portfolio-based funding.
  • National Bureau of Statistics of China, via Nature — China's 2025 R&D spending and 15th Five-Year Plan targets. Nature. June 2, 2026. China's 2025 R&D spend reached $549.7bn (2.80% of GDP, above the OECD average for the first time); basic-research funding crossed 7% of total R&D spend for the first time; the 2026–2030 plan targets at least 7% average annual R&D growth.
  • Inside Higher Ed / NPR — OMB proposes rules establishing political review of federal research grants. May 29, 2026 (comment period closes July 13, 2026). Proposed rule would require political-appointee pre-issuance review of discretionary grants across NIH, NSF, CDC, and DOE; downgrade peer review to advisory status; extend foreign-collaboration bans government-wide; and permit "convenience terminations" of active awards.
  • Anders M. Klöcker & Søren Barlebo Rasmussen — Mellem hype og styrkepositioner: Prioritering i forsknings- og innovationspolitikken. Mobilize Strategy Consulting. May 14, 2026. A Danish essay on prioritisation in research and innovation policy, contrasting hype cycles with mature "strongholds" and arguing that mission-oriented innovation stays rhetoric unless authority and resources are genuinely concentrated.
  • Jens Peter Andersen & Jesper W. Schneider — Forskningslandskabet i Danmark: En analyse af konkurrenceudsatte forskningsmidler. Dansk Center for Forskningsanalyse (Aarhus Universitet). May 2026. A systematic empirical mapping of Danish competitive research funding, 2018–2024, finding the system's instrument diversity is thinner than assumed, with the project grant dominating and imposing structural bindings.
  • Bjerre et al. — Safe to Speak, Teach, and Research? Experiences of Reprisals in Danish Academia. British Journal of Sociology. February 21, 2026. Peer-reviewed analysis of 130 #SafeToSpeak testimonies finds 49.4% of surveyed Danish university researchers report fearing, being threatened with, or experiencing reprisals for speaking up.
  • Nature Index — Nature Index 2026 Research Leaders. Nature. June 2026. China's indexed research output rose 22.4% year-on-year and nine of the global top ten institutions are now Chinese; India entered the global top ten for the first time. First year the index incorporated applied-sciences journals.
  • Mark Rubin — What is critical metascience and why is it important? Frontiers in Psychology. July 2, 2026. Argues for "critical metascience" as a distinct field that steps back to question metascience's own assumptions, contending the field is dominated by an activist open-science reform movement prone to collective bias.
  • Nature — The complex truth about trust in science. Nature. July 1, 2026. Editorial arguing that public confidence in science has not collapsed but that scientists face urgent problems to address, paired with a companion piece urging researchers to recognise their own political biases to build public trust.
  • Seemay Chou (Astera Institute) — What Scientists Said: Results from Astera's First Essay Competition. Astera Institute. July 1, 2026. Nearly 200 working scientists published essays diagnosing structural bottlenecks in their fields, with coordination failure the dominant theme.
  • Ben Reinhardt & Dan Recht — Pivot Your Career at Age 50. Reinvent Science. June 30, 2026. Argues that as scientists' healthspans lengthen, Planck's principle threatens to slow progress, so scientists should make a dramatic pivot to a new field between ages 50 and 60.
  • Kirsty Grainger — A Stupid Idea: Why the UK university crisis needs something politically impossible. anotherwayoflookingatit. June 29, 2026. Argues the UK university crisis should be met with a coordinated, nationally strategic managed contraction rather than market-driven decline, assigning each university disciplinary strength areas backed by block grants.
  • David Hilmer Rex — Debatten om metascience giver et alt for snævert billede. Der er behov for at brede samtalen ud. Science Report. June 27, 2026. Argues the Danish metascience debate has become too narrow, and calls for a more diverse ecosystem of institutions and new social practices for science.
  • ARIA — Clouds, clouds, everywhere. ARIA. June 26, 2026. ARIA's June update on new seed funding calls, plus a feature on Earth Compress, a £7M project building open-source AI infrastructure to compress and share petabyte-scale environmental data.
  • Kris Willis — Making Our Own Luck. The Good Science Project (cross-posted from Macroscience). June 25, 2026. Presents a bibliometric method that detects a signal of transformative breakthroughs on average five years before they occur by tracking bursts of new, influential, cross-topic papers.
  • Caroline Fry and Andrew Gerard — Science for Development: what comes next? Lauren Policy. June 24, 2026. DC workshop report on science in LMICs: returns are high but evidence is fragmented, there is no dedicated LMIC science think tank, and the effective customer is shifting away from USAID.
  • DFIR — Nye rammer for ph.d.-uddannelsen til styrkelse af kreativitet, entreprenørskab og mobilitet. DFIR. June 26, 2026. Concludes the Danish PhD framework has been essentially unchanged since 1992 and calls for more flexibility in regulation, research time, and thesis formats.
  • Alessandra Zimmermann (AAAS) — FY 2027 R&D Appropriations: The House R&D Report. AAAS. June 23, 2026. Summarises the House's first round of FY2027 markups, which sharply softened the administration's proposed cuts, holding NIH roughly flat while NSF faces a 20% cut, down from the 54.7% requested.
  • Elizabeth Gibney — Europe as science superpower: what it will take to rival the US and China. Nature. June 23, 2026. Europe already rivals the US on academic output but faces a persistent gap converting research into economic value.
  • Ludo Waltman, André Brasil, Kathleen Gregory, Stephen Pinfield & Peter Kolarz (RoRI) — From hyper-competition to strategic leadership: The role of funders and research institutions in 21st-century science. Research on Research Institute. June 23, 2026. Argues peer-review problems cannot be resolved within the current hyper-competitive logic, proposing funders shift to block funding based on institutional strategic profiles.
  • Brendan Haley — Why does Canada develop megaprojects instead of innovation systems? Centre for Industrial Policy. June 23, 2026. Argues Canada is structurally caught in the “staple trap”: crises push governments toward resource megaprojects rather than innovation systems, in contrast to MITI, Fraunhofer, and DARPA.
  • Christian Hendriksen — Om få år er problemet ikke, at forskere bruger AI for meget. Det er, at de bruger det for lidt. Altinget. June 22, 2026. CBS associate professor argues AI non-use will soon be the larger problem; calls on universities to provide institutional access to frontier models or risk making good scientific practice a privilege of the well-funded.
  • Geoff Mulgan — Shaping future institutions: What insights do academic disciplines offer? Substack. June 22, 2026. Surveys what economics, psychology, law, anthropology, and management theory each contribute to institutional design, finding strong analysis but no field that takes institutional design as its primary object.
  • Policy Atlas & Aidan K. — Five lessons from building evidence-strength scoring into an AI policy tool. Substack. June 22, 2026. Five practical lessons about reliability, transparency, and domain specificity from using LLMs to classify study types for a Nesta-backed policy evidence tool.
  • Science — NSF slashes research programs to support new tech initiative, insiders say. Science. June 2026. NSF has cut basic research budgets by 20–30%+ to fund the new $1.5B X-Labs initiative, despite a 3% congressional reduction cap.
  • Isaac Barbosa — Distributed peer review ‘good but no holy grail’, DSIT official says. Research Professional News. June 19, 2026. UKRI pilot results: 65% reduction in per-application assessment time; DSIT Metascience Unit cautions that speed gains don’t resolve selection-quality questions; follow-on AI survey announced.
  • Crypto Briefing — BioProtocol launches OpenLabs at DeSci Berlin. June 19, 2026. OpenLabs combines AI agents with community voting to coordinate research funding on BIO Protocol infrastructure; BIO Genesis raised $33 million ahead of the launch.
  • Nicolas Dufourcq, Rafael Laguna de la Vera et al. — Report of the Franco-German Task Force on Breakthrough Innovation. French Ministry of Economy / German Ministry of Research, Technology and Space. June 18, 2026. Recommends France establish a dedicated ARPA-type entity interoperable with SPRIND; identifies five ARPA invariants and a taxonomy of failure modes; argues France’s gap is institutional, not financial.
  • Uddannelses- og Forskningsstyrelsen — Bekendtgørelse nr. 548 om stillingsstruktur for videnskabeligt personale ved universiteter. Uddannelses- og Forskningsstyrelsen. June 18, 2026. Revises the Danish position structure for academic staff, opening the honorarium-paid clinical lecturer and clinical professor positions to professional groups beyond physicians.
  • Lars Frølund — From Fenced Yards to Shared Gardens: Quantum Technology, Sovereignty, and the Middle-Power Moment. Inaugural Lecture, Niels Bohr Institute, University of Copenhagen. June 19, 2026. Argues middle powers need “shared gardens” of collective capabilities held by trusted coalitions, proposing a Nordic Quantum Buyers’ Club as an embedded demand mechanism.
  • Stanford HAI — 2026 AI Index Report. Stanford University. 2026. Leading AI models score above human-expert baseline on PhD-level science question benchmarks for the first time, and a new science chapter documents AI's expanding role across the disciplines.
  • Alexander C. Furnas, Ruixue Jia, Margaret E. Roberts & Dashun Wang — Geopolitics in the Evaluation of International Scientific Collaboration. NBER Working Paper 34789. 2026. Two RCTs showing US policymakers and scientists penalise China-collaboration proposals by 18–40 pp vs. identical Germany proposals — the first empirical demonstration of geopolitical bias in science evaluation.
  • Nils Gilman — The University As We Know It Is Finished. Persuasion. June 2026. Argues Clark Kerr's "multiversity" is coming apart, with AI rendering obsolete the shared knowledge-transmission technologies that held it together, to be replaced by disaggregation.
  • Søren Barlebo Rasmussen & Bo Jellesmark Thorsen — Universitetet i bevægelse: samfundskobling, strategisk prioritering og nye spændinger i forskningen. Mobilize Nordic. 2026. Argues the classical disciplinary model is not in crisis but in motion, creating real dilemmas between autonomy, strategic direction, and institutional identity.
  • Caleb Watney & Jenn Gustetic — The Six Camps of Metascience: A Field Guide for Policymakers. Macroscience. June 11, 2026.
  • Geoff Mulgan — Rethinking universities' social contracts (…and how to avoid the fate of the monasteries). Substack. June 11, 2026.
  • Science Works & Wellcome Trust — Reorganising Research. Call for evidence mapping UK institutions operating outside the mainstream academic model; essay prize (£3,000, deadline July 15, 2026). June 2026.
  • Paul A. David (posthumous), ed. Bronwyn H. Hall & Jacques Mairesse — Research universities' futures in the networked world: technological challenges and opportunities for institutional responses. Economics of Innovation and New Technology. June 10, 2026.
  • Dario Amodei — Policy on the AI Exponential. June 2026.
  • Gerard DiPippo, Jonathon Sine & Benjamin Lenain — China's Techno-Industrial Strategy in the Xi Era: Producing Under Pressure. RAND Corporation. 2026.
  • Anthropic — Making Claude a Chemist. June 5, 2026.
  • Lucas Laursen — NSF Experiments With New Kind of Science Funding. IEEE Spectrum. June 4, 2026.
  • David Matthews — European DARPAs Won't Create New Industries Overnight. Science|Business. June 4, 2026.
  • Mark Hahnel — The New Era of Going "Fast and Far" in Research. June 4, 2026.
  • Shaun Leamon, Helen Buckley Woods, Peter Kolarz, James Wilsdon & Stephen Curry — Fast and Flexible Research Funding: A Review. RoRI Working Paper No. 20. May 2026.
  • Erik Canton — Multi-Criteria Portfolio Construction for Research and Innovation Policy: A Heuristic Optimisation Approach. European Commission R&I Paper Series. May 2026. Formalises R&I portfolio selection as a compromise optimisation problem and introduces a heuristic search for large candidate sets; shows ranking-based selection underperforms when criteria involve real trade-offs.
  • Loren DeJonge Schulman & Leya Mohsin — What the Metascience Community Should Learn From the Federal Evidence Movement Before Making Our Mistakes. FAS. June 3, 2026.
  • Matt Clancy, Caleb Watney et al. — A Policymaker-Centric Guide for Selecting Innovation Funding Mechanisms. NBER / University of Chicago Press. May 2026. Working-paper version: NBER w35549, July 2026.
  • Adam Marblestone & Anastasia Gamick — Searching for X-Lab-Shaped Bottlenecks in Quantum Interconnect and Integrated Photonics. May 29, 2026.
  • Cosmos Institute — What Will You Build For: Matt Clifford. May 29, 2026.
  • ARIA — May Update: From the Lab to the World. May 29, 2026.
  • IFP — IFP Update: May 2026. May 29, 2026.
  • Dorothy Chou & Zoë Brammer — AI for Science Funding Strategy. May 2026.
  • Stuart Buck — Mental Health at NIH. May 24, 2026.
  • Snezana Gvozdenovic — The Burden of Discernment. Palladium. May 23, 2026.
  • Jordi Cabana — Metascience Is Ignoring the National Labs. May 18, 2026.
  • Eirini Malliaraki — How Big Is Our Blindspot? Why AI for Science Needs More Than Better Models. Renaissance Philanthropy (SCI PHI). May 14, 2026. Argues the binding constraint on AI-driven science is shifting from model capability to data infrastructure, since many fields lack both the right data and the incentives to produce it as a public good.
  • Kevin Munger — Jubilee for Dysfunctional Institutions. May 12, 2026.
  • Eric Gilliam — Turning Risk Appetite Into Impact: An Ode to Warren Weaver. May 22, 2026.
  • Marius Oosthuizen — The Systems Transformation Architect: A Role for the Century Ahead. May 20, 2026.
  • Research on Research Institute — RoRI Launches META-MOMENT: The Largest Comparative Global Study of Metascience. May 19, 2026. Launched at the Global Research Council meeting in Bangkok, the project maps metascience units, investments and capabilities in 21 countries, with a report due in November 2026. The launch slides list the 21 countries, among them Norway, Germany, France, Belgium, the Netherlands, Austria and Ireland; Denmark is not included. Preliminary scientometric results show metascience output growing about 8% a year, with a 49% jump in 2020, and the US at the centre of international collaboration, with US–UK the strongest link. Funder heads from UKRI, Norway’s Research Council, the Australian Research Council, SSHRC, JST and Namibia’s NCRST set out why they use metascience; UKRI’s Ian Chapman calls it ironic that funders “have been far less scientific about our own funding practices”. Notes Research Ireland’s planned meta-research unit and China’s Shanghai Declaration on the Science of Science.
  • Alice Gibson — The Potential of Using Resources from Systems Change Theory to Strengthen Metascience. May 19, 2026.
  • Nan Ransohoff — The Third Wave of American Philanthropy. May 19, 2026.
  • Jiachen Liu et al. — The Last Human-Written Paper: Agent-Native Research Artifacts. arXiv:2604.24658. May 2026. Proposes replacing the PDF with a four-layer Agent-Native Research Artifact (ARA); reports QA accuracy improving from 72% to 94% and reproduction success from 57% to 64% on PaperBench.
  • Anastasia Gamick and Adam Marblestone — NSF Just Did Something Important. May 18, 2026.
  • Alex Shilina — Crypto's Most Serious AI Use Case May Be Science. April 25, 2026.
  • Research Ireland — Shaping the Next Phase of Ireland’s Research and Innovation System. April 20, 2026. Ireland’s new merged funding agency publishes its first Programme Plan after consulting the sector: annual calls for a new Investigators programme, a Pathway programme for early-career researchers, a new strategic institutional mechanism for PhD studentships and postdoctoral fellowships that consolidates supplementary grants, and a streamlined, more predictable funding offer. It follows the agency’s strategy for 2026–2030 (snapshot), which commits to a meta-research unit.
  • Andy Hall — AI is already 10x-ing academic research. How do we get to 100x?. April 16, 2026.
  • Sam Bogerd — It’s Time to Experiment with How Europe Funds Science. Science|Business. April 14, 2026. Argues that as the EU sets its next Framework Programme, almost no one asks how to learn to spend it well, and calls for an experimental metascience unit inside the European Commission.
  • Forsker-bevægelsen — 130 Testimonies from the #SafeToSpeak Campaign. April 2026.
  • Universities must reinvent themselves for the intelligent age. Time. April 13, 2026.
  • Ben Reinhardt — Research Has Customers. April 9, 2026.
  • Jordan Dworkin — Who will program manage the program managers?. March 25, 2026.
  • Niko McCarty — Ideas for Biotechnology BBNs. March 13, 2026.
  • Anastasia Gamick — Companies that should exist. March 10, 2026.
  • Aimee Rose, Andrew Chang, and Raylene Yung — The Early Stage Deep Tech Funding Gap: Why Great Companies Get Stuck Between Capital Markets. March 5, 2026.
  • Research Ireland — Curiosity, Capability, Competitiveness: Charting Ireland’s Research and Innovation Future 2026–2030. March 2, 2026. The first strategy of Ireland’s merged funding agency. Strategic objective 08 commits it to “create a meta-research unit within Research Ireland to provide a national policy intelligence function and a ‘strategic brain’ for Ireland”, using research metadata, AI tools and patent data to benchmark and right-size national research investment and better match its funding to researchers’ needs; a Meta-Research Fund is listed as a potential new initiative. It also commits to regular independent reviews of its programmes and to building evaluation in from the start. The strategy drew an open letter of concern signed by over 2,300 researchers.
  • Prachee Avasthi — Free the PhD. February 28, 2026.
  • Jed McCaleb and Seemay Chou — Sharpening Astera's focus. February 23, 2026.
  • Stuart Buck and Aishwarya Khanduja — A Reality-Based View of Government Funding of Science. February 18, 2026.
  • Michael Nielsen — Which Future?. February 17, 2026.
  • Renaissance Philanthropy — Rewiring the Knowledge Economy: Building a New Infrastructure for Research Funding.
  • Nature — US science after a year of Trump: what has been lost and what remains. January 2026.
  • Benchling — 2026 Biotech AI Report. Benchling. January 2026. A survey of AI adoption across biotech finding that real wins cluster where data is clean, structured, and verifiable, while adoption stalls in messier, regulated, generative-design territory.
  • Eric Gilliam — The BBN Fund. January 16, 2026.
  • Nan Ransohoff — There should be 'general managers' for more of the world's important problems. January 15, 2026.
  • Anastasia Gamick, Adam Marblestone, and Joseph Fridman — Announcing Merge Labs. January 15, 2026.
  • Andrew Gerard (IFP Metascience Team) — NSF Tech Labs FAQs. January 15, 2026.
  • Sandeep Patel, Stuart Buck, and Amit Gupta — Accelerating the American Scientific Enterprise. January 14, 2026.
  • Institute for Progress — IFP Update: New Science Funding Edition. January 13, 2026.
  • Ben Reinhardt — Spectech Lessons and Updated Hypotheses from 2025. January 13, 2026.
  • Santi Ruiz — What's Wrong with NIH Grants?. January 9, 2026.
  • Samuel Arbesman and Asimov Press — Why Do Research Institutes Often Look the Same?. January 8, 2026.
  • Anastasia Bektimirova and Chris Fellingham — An ARIA-grade agenda for social science. January 8, 2026.
  • David Lang — The Shape of Science to Come. January 3, 2026.
  • Peter Palumbo, Jordan Kruger & Christopher L. Robinson — Decentralized Science Supported by Decentralized Finance to Support Pain Research: A Merger of Giants. 2026.
  • Sharma, G., Munteanu, V., Ghiasi, N.M. et al. — Towards a Decentralized Future for Open-Science Databases. Nature Genetics. 2026.
  • 2025
  • Adam Marblestone — Convergent Research: A Mission Control for Frontier Technology. December 23, 2025.
  • Stuart Buck & Aishwarya Khanduja — The Economy of Knowing. December 23, 2025.
  • Adam Marblestone, Anastasia Gamick, and Joseph Fridman — What Is the Future Missing?. December 24, 2025.
  • Ben Reinhardt — The Art of Industrial Leapfrogging. December 16, 2025.
  • Convergent Research — The Future of Focused Research Organisations. December 16, 2025.
  • Stuart Buck — Innovations in Scientific Institutions. December 13, 2025.
  • Caleb Watney — Science Funding Goes Beyond the Universities. December 12, 2025.
  • Federation of American Scientists — NSF Plans to Supercharge FRO-style Independent Labs. We Spoke with the Scientists Who First Proposed the Idea.. December 12, 2025.
  • James Pethokoukis — Funding Outside the Box (A Quick Q&A). December 11, 2025.
  • Prachee Avasthi — How to Not Make a Scientific Journal by Accident. December 4, 2025.
  • Santi Ruiz (Statecraft) — How to Save Science Funding. December 4, 2025.
  • Tim Hwang — Metascience in Dangerous Times. December 2, 2025.
  • Pangram Labs — Pangram Predicts 21% of ICLR Reviews Are AI-Generated. November 2025.
  • Tony Blair Institute for Global Change — A New National Purpose: Reimagining UK Science and Technology Through Lovelace Disruptive Invention Laboratories. November 20, 2025.
  • Smrithi Sunil — Inside the movement that's rewriting how we do science. November 19, 2025.
  • Ashlee Vance — Exclusive: A New Bid To Control The Genome. November 17, 2025.
  • Seemay Chou — Big Experiments Are Only Big If They Can Fail. November 6, 2025.
  • Jesper W. Schneider — Forsker: Nyt udspil lover risikovillig forskning – men systemet hæmmer potentialet. Altinget. November 5, 2025. Responding to the Danish government’s DKK 19 billion research plan, Schneider argues that risk-taking research depends less on funding levels than on institutional “protected spaces” giving researchers stability and autonomy, as the Grundforskningsfond centres of excellence appear to provide. He warns that concentration, hyper-competition and precarity go unaddressed and that a national strategy should include the private foundations.
  • Sandeep Patel — Why Science Needs a Marketplace. How we can make scientific progress less accidental.. October 30, 2025.
  • Anastasia Gamick — EP 05: Anastasia Gamick (Co-founder & President Convergent Research), Accelerate Science Now. October 30, 2025.
  • Ben Reinhardt & Stuart Buck — The Paradox of Harmful Information. Risk and progress in the sciences.. October 15, 2025.
  • Adam Marblestone — What Science can Learn from Startups | Adam Marblestone on Focused Research Organizations. October 7, 2025.
  • Lukas Weidener & Sasha Shilina — Futarchy in decentralized science: empirical and simulation evidence for outcome-based conditional markets in DeSci DAOs. October 6, 2025.
  • Karthik Tadepalli — Ideas Aren't Getting Harder to Find. October 2025.
  • Prachee Avasthi — Scholarly Communication Is a Research Problem. This Means You. October 4, 2025.
  • Pierre Azoulay, Matthew Clancy, Danielle Li & Bhaven N. Sampat — What If NIH Had Been 40% Smaller?. Science. September 25, 2025. A counterfactual analysis asking how today's medicines would have been affected if the permanent NIH budget had been 40% smaller, finding cuts of this magnitude could affect a large share of drug approvals.
  • Nature — The Future of Universities. They made the modern world. Now they must survive it.. September 24, 2025.
  • Ed Boyden, Nina Khera, Adam Marblestone & Claire Wang — The Tiling Tree Method: How to think of every way of solving a problem. September 21, 2025.
  • Alvin Djajadikerta & Laura Lungu — Why science needs outsiders. September 18, 2025.
  • Tom Kalil — EP 01: Tom Kalil (CEO of Renaissance Philanthropy), Accelerate Science Now. September 3, 2025.
  • Geoff Mulgan — An ARIA for Social Issues?. September 2, 2025.
  • Stuart Buck — The case for crazy philanthropy. August 22, 2025.
  • Emil Bargmann Madsen, Philippe Mongeon & Jesper Wiborg Schneider — Gender Disparity in Funding Rates in Double-Blind Grant Peer Review: The Case of the Villum Experiment. Quantitative Science Studies. August 18, 2025. An analysis of 2,041 applications to Villum Fonden’s fully blinded scheme finds a small gender gap in success rates, concentrated in life sciences, that is unlikely to come from reviewer bias and largely disappears once the underrepresentation of senior women among applicants is accounted for. Blinding can remove direct bias but not inequalities that precede the application.
  • Ben Reinhardt & Dan Recht — Science Needs Outlier Organizations. There is no best structure, but there are many bad compromises.. August 12, 2025.
  • Seemay Chou — From systems operators to systems architects. August 12, 2025.
  • Ben Reinhardt & Dan Recht — Science is an Extreme Phenomenon. August 5, 2025.
  • Ben Reinhardt & Dan Recht — A New Scientific Method. July 30, 2025.
  • Amitabh Chandra & Connie Xu — Where Discovery Happens: Research Institutions and Fundamental Knowledge in the Life Sciences. July 2025.
  • Carter Bloch, Ea Utoft, Andreas Kjær Stage & Steven Wooding — Risk-Taking and Research Funding: The Case of the Villum Experiment. Minerva. July 2025. How grantees of Villum Fonden’s programme for unorthodox ideas, judged under anonymous review, perceive and handle risk in their proposals and research, drawing on interviews with 24 recipients and surveys of recipients from 2017 to 2023.
  • Sanja Terlević — Can Europe Build an ARPA of Its Own?. Polity Futures. July 6, 2025. Looks at what makes the ARPA model work, why it is hard to replicate, and how the European Innovation Council’s reform plan borrows from it as the EIC positions itself as “Europe’s ARPA”.
  • UK Research and Innovation — A year in metascience. June 30, 2025.
  • UK Metascience Unit / RoRI — The Past, Present and Future of UK Metascience (Annex). June 30, 2025.
  • Ashlee Vance with Adam Marblestone & Anastasia Gamick — Fixing American Science. June 25, 2025.
  • Patrick Joyce, Aldo de Pape & Aaron Weaver — Open & Decentralised Science at Proof of Talk 2025. June 22, 2025.
  • Alex Obadia — Wanted: New Instruments to Fund BBNs. June 20, 2025.
  • David Lang — Emboldening Scientific Curiosity | Experiment Foundation. June 18, 2025.
  • Eric Gilliam — An Oral History Interview with ARIA CEO Ilan Gur. June 12, 2025.
  • Eric Gilliam — Inside ARIA, the UK's Bet to Build Scientific Revolutions. June 12, 2025.
  • Zach Winn — Former MIT researchers advance a new model for innovation. June 6, 2025.
  • Seemay Chou — Scientific Publishing: Enough is Enough. Why we're no longer funding journal publications. June 2, 2025.
  • Lise Degn — En offentlig forskningsstrategi nytter ikke noget. Private fonde skal med. Altinget Fonde. May 28, 2025. A CFA researcher argues that a Danish research strategy must include the private foundations, because public and private funders both concentrate money on the same researchers and have converged on the same project-grant “laboratory model”, giving a remarkably uniform way of awarding grants in an otherwise varied funding landscape. She calls for a joint national strategy that commits the whole system.
  • The Times Tech Podcast — Can the UK invent the next internet?. May 23, 2025.
  • Matt Abbott, Marc Sedam & Paul Kohlhaas — Ship or Die 2025: University Research Driving Innovation. May 23, 2025.
  • David Matthews — Are groundbreaking science discoveries becoming harder to find?. May 21, 2025.
  • Kumar Garg — How philanthropy can fuel scientific innovation. May 15, 2025.
  • SpineDAO interviewed by Ella McCarthy-Page of Molecule — How SpineDAO Is Redefining DeSci - From Back Pain to Blockchain. May 13, 2025.
  • The Techno-Industrial Policy Playbook. May 2025.
  • Caleb Watney — Launching X-Labs for Transformative Science Funding. May 2025.
  • Stuart Buck — Experimenting with NIH Funding. May 2025.
  • Geoff Mulgan — Do we need a new generation of public institutions for science, technology and innovation?. May 6, 2025.
  • Adam Marblestone, Anastasia Gamick, Mary Wang & Joseph Fridman — Field Notes on Moving Focused Research Organizations Forward. Fall 2025.
  • Lisa Margonelli — Innovation's Hidden Scaffolds. Summer 2025.
  • DeSciWorld — Science Needs More Context: Introducing Knowledge Network. April 28, 2025.
  • Kumar Garg — Pathways to Progress. April 23, 2025.
  • Pierre Azoulay, Raffaella Sadun & Daniella Scur — Before the Exodus? Young Scientists and the Future of US Science. NBER Working Paper 35330. 2025. Survey of 916 young biomedical scientists: likelihood of staying in the US fell from 93% to 72% following funding disruptions; foreign-born scientists showed the largest decline (−30.8 pp).
  • Nicolai Søndergård — DeSci: The Future of Science or Pure Window Dressing?. March 20, 2025.
  • Jen Nelles & Tim Vorley — The Applications of Metascience to Research and Innovation Systems. In Innovations in Innovation Policy, Edward Elgar. March 18, 2025.
  • Elizabeth Gibney — AI tools are spotting errors in research papers: inside a growing movement. Nature. March 7, 2025. LLMs used systematically to find errors in published papers; profiles researchers building detection tools and the error types they surface.
  • Paul Kohlhaas and Tim Copeland — Bio Protocol on DeSci and the future of DAOs. March 2, 2025.
  • Alexey Guzey & Jacob Trefethen — The Future of Science. February 28, 2025.
  • Ciarán Murray — The Crossroads of DeSci: The Choice is Ours to Make. February 24, 2025.
  • Adam Marblestone — How to Build Essential Technology. February 7, 2025.
  • Benjamin Reinhardt — Unbundle the University. February 11, 2025.
  • Binance — From Challenges to Opportunities: How DeSci Reimagines Science. February 9, 2025.
  • Science|Business — Germany's Sprind innovation agency: what works, what does not. Science|Business. January 2025. Reports the independent five-year Fraunhofer ISI evaluation of SPRIND, finding it has drawn strong researcher interest and built a faster funding culture, with open questions about its independence and financial muscle.
  • China’s Ministry of Science and Technology (translated by CSET) — Guiding Opinions on Promoting the Development of New-Style Research and Development Institutions. Center for Security and Emerging Technology. January 14, 2025. Translation of MOST’s 2019 guidelines for “new-style R&D institutions”: independent legal entities outside the state-run lab system, with diversified investors, modern management, market-style operation and flexible hiring, registered as private non-profits or public institutions. Researchers may profit from their inventions, to improve the weak conversion of research into applications in state labs. The model began in Shenzhen in the late 1990s (see Conlé on Guangdong’s new R&D institutes). Full translation.
  • NFTevening — What Is DeSci? The Ultimate Guide to Decentralized Science. January 8, 2025.
  • Daniel Roelfs — The Money in Scientific Publishing. May 9, 2024.
  • Anna-Lena Schindl — Resource (Airtable).
  • 2024
  • Tyler Golato — TR01: From Institutions to Networks. December 2024.
  • Benjamin Reinhardt — Spectech Lessons and Updated Hypotheses from 2024. January 14, 2025.
  • Lukas Ruppert for Maelstrom Fund — Degen DeSci. January 3, 2025.
  • Adam Marblestone — Automating Math. January 2025.
  • Messari Crypto — DeSci Ecosystem Map. December 2024.
  • Tom Kalil — Tom Kalil on Institutions for Innovation (with Matt Clancy). December 14, 2024.
  • AuraSci — DeSci Fall Report. December 12, 2024.
  • Ross Pomeroy — Can we automate science? Sam Rodriques is already doing it. November 18, 2024.
  • Mark Hahnel — Innovating Academia From The Outside - The Web3 Solution to Academia's Peer Review Problem. November 14, 2024.
  • Yen Tran — Billion-Dollar Race: LIFE AI's Breakthrough as Avalanche's Petabyte-Scale Layer 1. October 23, 2024.
  • Dan Elton — A Defense of Peer Review. October 22, 2024.
  • Eric Gilliam — A Scrappy Complement to FROs: Building More BBNs. October 16, 2024.
  • Statecraft — How to Build the British ARPA. September 5, 2024.
  • Abhishaike Mahajan — Things I learned talking to the new breed of scientific institution. August 26, 2024.
  • Onchain Research — DeSci – (How) Will It Make Research Great Again?. August 16, 2024.
  • Veronica Kirin — What is Decentralized Science (DeSci)?. July 19, 2024.
  • Lukas Weidener & Cord Spreckelsen — Decentralized science (DeSci): definition, shared values, and guiding principles. July 17, 2024.
  • Weixin Liang et al. — Can Large Language Models Provide Useful Feedback on Research Papers?. NEJM AI. July 17, 2024. GPT-4 feedback overlaps with human reviewer judgments at rates comparable to human-human agreement across 3,096 Nature papers and 1,709 ICLR submissions; 57% of researchers found it helpful.
  • Foresight Institute Podcast — Stuart Buck - What is Good and Bad Science. May 30, 2024.
  • Tyler Golato — Disrupting the Decline of Disruptive Science. May 28, 2024.
  • Benjamin Reinhardt — Research Leaders' Playbook. Tools and tactics for leading coordinated research programs. April 4, 2024.
  • Jocelynn Pearl and UltraRare — DeSci Wiki. Last update: January 1, 2024.
  • 2023
  • Danske Universiteter and six foundations — Agreement Between Universities and Foundations on Funding Research Projects. November 2023 (English text; guidelines version 3.0, December 2025). The eight Danish universities and the Novo Nordisk Foundation, Villum Fonden, the Carlsberg Foundation, VELUX FONDEN, the Lundbeck Foundation and the LEO Foundation agree a project supplement on grants from open calls under DKK 50m, a clear split of costs between universities and foundations and more transparency about the foundations’ contribution to university research. The guidelines put it into operation, and the parties meet annually to discuss the research system and the agreement. The Novo Nordisk Foundation will cover indirect costs on larger calls and single grants by dialogue, at least at the same level.
  • Jeffrey Ding — The diffusion deficit in scientific and technological power: re-assessing China's rise. Review of International Political Economy. March 2023.
  • Erik Canton — Being right or fair. A portfolio approach to research funding. November 11, 2023.
  • The DemystifySci Podcast — Science is a Strong Link Problem - Dr. Adam Mastroianni. August 31, 2023.
  • Janika Schmitt & Jake Swett — The ARPA Model: A Reading List. Institute for Progress. January 31, 2024. A syllabus on DARPA and its descendants: the Heilmeier Catechism, how ARPAs manage risk, the literature on replicating the model, and the ARPA clones in the US and abroad.
  • John Paschkewitz & Dan Patt — No, We Don’t Need Another ARPA. Issues in Science and Technology 40(1). Fall 2023. Two former DARPA programme managers argue that “ARPAs for everything” miss the real US advantage, diffusion capacity: the ability to adopt and spread innovations. They call for modular technologies, rotations between academia and industry, NACA-style catalyst institutions that publish usable results, SEMATECH-style consortia, and finance for scale-up such as advance market commitments.
  • Nadia Asparouhova — Early stage funding markets for science - an analysis. January 23, 2023.
  • 2022
  • Sarah Hamburg (a16z) — A Guide to DeSci, the Latest Web3 Movement. September 2, 2022.
  • Michael Nielsen & Kanjun Qui — A Vision of Metascience. An Engine of Improvement for the Social Processes of Science. October 18, 2022.
  • Scott Moore (Gitcoin) — DeSci: The case for Decentralised Science. August 29, 2022.
  • Benjamin Reinhardt — Whence Systems Research? What the heck is systems research? Why is it important? Why are we bad at it? How could we do better?. August 11, 2022.
  • Juan Benet — DeSci, Independent Labs, & Large Scale Data Science (video). July 20, 2022.
  • Jocelynn Pearl — A Guide to Decentralized Biotech. April 20, 2022.
  • Nadia Asparouhova — Understanding science funding in tech, 2011-2021. March 2, 2022.
  • 2021 and earlier
  • Benjamin Reinhardt — Shifting the impossible to the inevitable. A Private ARPA user manual. April 2021.
  • Kevin Munger — Meta-Science is Political Methodology. Never Met a Science. July 17, 2020. Argues that once methodology aims at the quality of a whole body of knowledge, the institutions that produce it (journal formats, the number of journals, peer review, the governance of disciplinary associations) become part of methodology, and so do their politics. Calls for data from journals, theory and experiments on peer review: “a positivism which uses non-positivist methods for evaluating its own success lacks the confidence of its convictions.”
  • Pierre Azoulay, Erica Fuchs, Anna P. Goldstein & Michael Kearney — Funding Breakthrough Research: Promises and Challenges of the “ARPA Model”. Innovation Policy and the Economy 19. 2019. The standard reference on the model. Its core is organisational flexibility plus programme directors with authority to design, select and actively manage projects. It suits mission-oriented research on nascent technologies where the existing innovation system works poorly, and is not a general fix.
  • Christian Ketels et al. — Peer Review of the Danish R&I System: Ten Steps, and a Leap Forward. European Commission Horizon 2020 Policy Support Facility. 2019.
  • Patrick Collison & Michael Nielsen — We Need a New Science of Progress. July 30, 2019.
  • Gerald M. Rubin & Erin K. O’Shea — Point of View: Looking Back and Looking Forward at Janelia. eLife. February 7, 2019. HHMI’s leaders review twelve years of the Janelia Research Campus, built as an environment complementary to universities: small internally funded labs, group leaders at the bench, project teams that turn tools into products, and career staff scientists. Without management pushing back, they found a slow drift towards a conventional environment, so they introduced 15-year terms for research areas, chosen by open competition, with a new area every five years. Founding Janelia was chosen over expanding the investigator programme by 20%.
  • Patrick Collison & Michael Nielsen — Science Is Getting Less Bang for Its Buck. November 16, 2018.
  • Sönke Bartling & Benedikt Fecher — Blockchain for science and knowledge creation. August 14, 2016.
  • Pierre Azoulay, Joshua S. Graff Zivin & Gustavo Manso — Incentives and Creativity: Evidence from the Academic Life Sciences. October 2009.
  • Undated
  • Ethereum — Decentralized science (DeSci).
  • Vincent Weisser — Decentralized Science and Biotech & Metascience.
  • José Luis Ricón — Research Funding category on his website.