sciencedao

Traditional grants compress an entire research trajectory into a single decision point. A committee reads a proposal, predicts what will happen over three to five years, and commits a lump sum. If the prediction is wrong — and predictions about research are wrong most of the time — the money is already spent and the correction comes years later, if at all. Software engineering solved this problem decades ago. Nobody ships a monolithic product after a single planning meeting. Teams iterate. They ship small increments, measure what works, adjust what does not, and allocate resources continuously based on demonstrated progress. The feedback loop is days or weeks, not years. DeSci projects are starting to apply this same logic to research funding. Instead of front-loading all resources into a grant proposal, continuous funding models release support incrementally based on ongoing contribution. A researcher publishes a dataset — funding flows. They release a method others build on — funding increases. They hit a dead end and report it transparently — funding continues because honest null results have value. This inverts the risk structure. Under one-shot grants, the funder bears all the prediction risk. Under continuous models, risk distributes across time. Bad bets get caught early. Good bets compound. Researchers who consistently deliver value receive increasing support without re-applying from scratch every cycle. The technical infrastructure for this exists. Contribution tracking, transparent metrics, and programmable funding flows are all buildable today. What is missing is the willingness to abandon the comfort of single-decision funding. The question: would you rather fund a promising proposal or fund demonstrated progress? Explore the DeSci DAO approach to continuous research support.

Every major tech company is building AI to assess research impact. Google Scholar already ranks papers algorithmically. Semantic Scholar uses machine learning to predict which studies will be influential. Microsoft Academic tried before it shut down. The infrastructure for automated scientific evaluation exists today. The question is not whether AI will evaluate science. It is whether that evaluation will be transparent, auditable, and aligned with what science actually needs. Here is the risk: proprietary algorithms deciding which research gets visibility, funding, and credit. A neural network trained on citation patterns might optimize for what gets cited rather than what gets used. An AI system built by a for-profit company might prioritize metrics that serve corporate interests rather than scientific progress. We have seen this movie before. Social media algorithms optimized for engagement and got polarization. Search algorithms optimized for clicks and got clickbait. If we build opaque AI systems that optimize for the wrong signals in science, we will get research optimized for algorithmic approval rather than genuine discovery. The alternative: evaluation systems where every input is visible, every weight is auditable, every decision can be challenged. Systems where the community can see why a contribution was valued, verify the logic, and correct errors. Systems that amplify human judgment rather than replacing it with a black box. This is the difference between AI as oracle and AI as tool. An oracle gives you answers you cannot question. A tool gives you capabilities you can direct. Science needs tools, not oracles. The technical challenge is significant. Scientific impact is multidimensional and domain-specific. A breakthrough in theoretical physics looks different from a methodological advance in epidemiology. Building evaluation systems that capture this diversity without becoming unwieldy requires careful design and continuous community feedback. But the technical challenge is solvable. The harder problem is governance: who controls the evaluation criteria? Who audits the system? Who decides what counts as valuable contribution? A meritocratic approach to science funding means building AI evaluation as transparent infrastructure, not proprietary gatekeeping. The algorithms should be open-source. The training data should be public. The evaluation criteria should be community-governed. The outputs should be explainable. The discussion: Would you trust an AI system to evaluate your research contributions if you could see exactly how it worked? What would need to be true for that trust to be justified? And who should control the evaluation criteria—the scientists, the funders, or the platform builders?

In 2015, the Open Science Collaboration tried to replicate 100 published psychology studies. Only 36 held up. The number sent shockwaves through the field — but anyone who had spent time in research already knew something was wrong. Since then, similar patterns have surfaced in cancer biology, economics, machine learning, and drug discovery. Bayer reported in 2011 that its internal teams could reproduce only 20-25% of published preclinical findings. Amgen fared even worse: 6 out of 53 landmark studies replicated. The default response is to blame scientists. Fraud. Sloppiness. Cutting corners. But this misdiagnoses the problem. Most researchers are honest people operating inside a machine designed to produce the wrong outputs. Here is the machine: funding depends on publications. Publications depend on positive results. Positive results depend on hypotheses that happen to be correct. The system selects for researchers who can generate a steady stream of clean, positive, publishable findings — and punishes anyone who reports messy reality. This is not a character flaw. It is an incentive flaw. When survival depends on producing positive results, researchers unconsciously make methodological choices that inflate them. They stop experiments when results look good. They test multiple hypotheses and report the winners. They exclude inconvenient data points. Each step is rational within the incentive structure. The aggregate effect is a literature that systematically overstates what is true. Fixing this requires changing what the system rewards. Not asking scientists to be braver. Not adding ethics training. Restructuring funding so that methodological rigor, transparent reporting, and honest null results receive the same support as flashy positive findings. A DeSci DAO approach begins from this insight: build evaluation around what research actually contributes after publication, where replication attempts, negative results, and methodological soundness are visible and valued. Remove the survival pressure to produce clean narratives, and the literature starts reflecting reality. The question: if you could redesign what gets rewarded in science, what single change would reduce the replication problem most?

Here is how journal peer review works: two or three anonymous reviewers evaluate a manuscript before it sees the light of day. They judge potential impact based on a PDF and a cover letter. Editors make a binary decision and the work either enters the scientific record or vanishes into a drawer. Here is how almost every other knowledge field works: publish first, let the community evaluate openly, and let merit emerge over time through usage, critique, and replication. Software ships to production. Articles go live. Music drops. Then the real evaluation begins, in public, at scale, continuously. Science is the outlier. And it shows. Pre-publication review selects against novelty. Reviewers are conservative by incentive because the status quo is their career. Studies confirm what everyone in academia already suspects: peer review is poor at identifying transformative work and decent at filtering out obvious errors. It is a quality floor, not a merit detector. The more interesting evaluation happens after publication, when other scientists attempt replication, build on findings, or discover flaws. But this post-publication evaluation has no formal connection to funding, credit, or career advancement. What if we inverted the process? What if the real assessment, the one that determines funding and recognition, happened after the work existed in the open? This is what Open Science Funding experiments with: shifting evaluation from gatekeeping before publication to merit recognition after it. What breaks if we flip the order? What improves?

In July 1945, Vannevar Bush handed President Truman a report called Science, The Endless Frontier. It argued that the federal government should fund basic research through competitive grants administered by expert panels. The National Science Foundation was born. The modern grant system was born. The entire architecture that still determines how billions of dollars flow to scientists today was born. That was eighty-one years ago. Think about what the world looked like in 1945. There were roughly 5,000 active research scientists in the United States. The entire federal R&D budget was a fraction of what a single mid-sized NIH institute spends today. Scientific communication happened through printed journals mailed to university libraries. Computing meant a room full of people with slide rules. The system Bush designed made sense for that world. A small community of researchers, known to each other by reputation, submitting proposals to panels of peers who could reasonably evaluate them. Low volume. Slow pace. Manageable scale. Now look at today. There are over 1.5 million active researchers in the US alone. The NIH receives roughly 50,000 grant applications per year and funds about 20%. The NSF success rate hovers around 25%. Reviewers are overwhelmed. Panel fatigue is real. The average age of a first NIH R01 grant has climbed to 42 — meaning researchers spend their most creative decade begging for scraps. The volume alone breaks the 1945 model. But the deeper problem is structural. Bush assumed science was best evaluated prospectively — that expert committees could predict which ideas would succeed before any work was done. Eight decades of evidence say otherwise. Study after study shows that peer review of grant proposals is essentially random for proposals above a quality threshold. A 2016 analysis of NIH review scores found that changing the reviewer panel would change the funding decision for roughly half of all borderline applications. We are running prediction software on hardware built for a world that no longer exists. Other fields have moved past prospective-only evaluation. Software development discovered that shipping code and measuring usage beats writing elaborate design documents. Venture capital learned that small initial bets with follow-on funding for demonstrated traction outperforms committing large sums to untested business plans. Even government procurement has started experimenting with challenge-based awards that pay for results rather than proposals. Science funding has not adapted. The reason is partly institutional inertia — the people who succeeded under the 1945 system now run the committees that perpetuate it. And partly it is a lack of alternatives. Nobody has built a credible, scalable mechanism for evaluating research contribution after the fact and routing funding accordingly. Until now. Science funding innovation means building evaluation systems that match the scale and speed of modern science. Post-publication merit. Continuous contribution tracking. Funding that flows based on demonstrated impact rather than predicted potential. These are not radical ideas — they are obvious adaptations that the technology to implement simply did not exist in 1945. It exists now. The question is not whether the 1945 model is broken. Anyone who has spent time in academic research knows it is. The question is whether we have the courage to build something that actually fits the science of 2026 — or whether we will keep patching an operating system designed for slide rules and printed journals. What would you change if you were handed a blank slate and told to design science funding from scratch today?

Myth: Great discoveries come from brilliant individuals working alone in their labs. Reality: Every "breakthrough" paper rests on hundreds of prior contributions—methods developed by postdocs in other labs, datasets collected by field researchers, software written by graduate students who never became professors, reagents optimized by technicians whose names appear nowhere in the acknowledgments. Yet our funding system still operates on the lone genius model. Grants go to principal investigators. Awards celebrate individual names. Tenure committees count first-author and last-author publications. The entire apparatus assumes science is a series of individual achievements rather than a collective, cumulative process. This creates a perverse incentive structure. Researchers optimize for visible, individual contributions rather than foundational work that enables others. Why spend three years building a dataset that fifty other labs will use when you could publish six incremental papers with your name in bold? The problem compounds over time. Foundational contributions—the methods, tools, datasets, and negative results that make future discoveries possible—remain chronically underfunded because they do not fit the individual achievement model. Meanwhile, researchers who build on that foundation receive credit (and funding) for "novel" discoveries that would have been impossible without the invisible infrastructure beneath them. Consider CRISPR. The headlines celebrated Doudna and Charpentier (rightly so). But the underlying biology depended on decades of obscure research into bacterial immune systems by scientists who never made the news. The funding system rewarded the final step, not the thousand steps that made it possible. What if we could track scientific contribution the way we track software dependencies? What if funding could flow not just to the visible breakthrough but to the foundational work that enabled it—proportionally, transparently, based on actual demonstrated dependency rather than retrospective narrative? This is what alternative research funding models attempt to solve: shifting from individual achievement to network contribution, from visible outputs to enabling infrastructure, from rewarding who crossed the finish line to sustaining everyone who built the track. Discussion question: What foundational work in your field made your research possible but will never receive adequate recognition or funding under the current system? And what would change if we could properly value and fund that kind of contribution?

A neuroscientist has a breakthrough idea in March 2024. She spends four months writing a grant proposal. The proposal goes to review in July. She gets feedback in November. She revises and resubmits in January 2025. Funding decision arrives in April 2025. Money hits her account in September 2025. Eighteen months from idea to execution. In that time, three other labs published related work. The field moved on. Her original hypothesis needs revision. The postdoc she wanted to hire took a position elsewhere. The specific reagent she planned to use is now backordered for six months. This is not an edge case. This is how the system works. The grant cycle was designed for a different era. When science moved slower, when ideas had longer half-lives, when a two-year delay meant missing one conference cycle, not an entire paradigm shift. Today, fields like machine learning, genomics, and synthetic biology move on timescales measured in months, not years. The time cost creates a selection bias that few people talk about. Researchers learn to propose work that will still be relevant in two years — which means work that is incremental, predictable, and safe. High-risk, time-sensitive ideas get shelved before they are even proposed. Consider the opportunity cost. A senior researcher spends 400 hours a year writing grants. That is ten full work weeks. Ten weeks of experiments not run, papers not written, students not mentored, discoveries not made. For early-career researchers, the calculus is worse. They cannot afford to spend months on proposals with low success rates. So they stick to safe extensions of their advisor's work, building careers on incremental contributions while genuinely novel ideas gather dust. The system optimizes for proposal quality, not research velocity. It selects for people who are good at writing about what they might do, not people who are good at doing it. What if funding could move at the speed of science itself? What if researchers could receive support based on demonstrated contribution rather than projected plans? What if the evaluation happened after the work existed, when its value could actually be assessed? This is the core insight behind research impact funding — shifting the evaluation point from proposal to product, from promise to proof. When funding follows demonstrated value rather than projected potential, the eighteen-month delay disappears. Researchers can act on ideas while they are still fresh. The question: How many breakthrough ideas are currently sitting in grant proposal drafts, waiting for a funding decision that will arrive too late? And what would science look like if the best ideas got funded in weeks instead of years?

A philanthropist writes a $500,000 check to fund cancer research. The money disappears into a university's general fund. Three years later, she receives a glossy annual report with stock photos and vague claims about "advancing knowledge." Did her donation accelerate a breakthrough? Did it fund a postdoc who made a key discovery? Or did it subsidize overhead costs and administrative bloat? She has no idea. Neither does anyone else. This is the donor's dilemma in traditional science philanthropy: you give money, you hope for impact, and you trust the institution to be a responsible steward. But trust is not accountability. The problem compounds at scale. Billions of dollars flow into research funding annually through foundations, individual gifts, and corporate sponsorships. Most donors receive zero feedback on whether their contributions produced measurable value. They fund labs that produce incremental papers rather than paradigm shifts because nobody tracks the difference. The traditional model assumes institutional reputation equals effective allocation. But reputation is backward-looking—it reflects past achievements, not current efficiency. A prestigious university might have excellent researchers buried under administrative friction. A small independent lab might be producing groundbreaking work with minimal overhead. Without transparent feedback mechanisms, donors cannot distinguish between the two. Some argue this is fine—that donors should trust experts to allocate resources. But expertise in research does not equal expertise in resource allocation. And when the feedback loop is broken, even expert allocators cannot optimize what they cannot measure. What if science funding operated more like open-source software development? Contributors see exactly what their support enables. Impact is measurable through usage, citations, and downstream applications. Funding decisions can be based on demonstrated value rather than institutional prestige or proposal quality. This is the promise of transparent science philanthropy—a model where donors can trace their contributions through the entire research lifecycle, where funding follows evidence of impact rather than promises of future achievement. The question for discussion: if you could see exactly how your science donation was used and what it produced, would that change how much you give? And what would true transparency in research funding actually look like?

Every major cancer genomics study published this year relied on NumPy. Every climate model that made it into a policy brief leaned on SciPy. Every economics paper with regression analysis probably used R packages maintained by a handful of unpaid volunteers on weekends. Here is the part that should bother everyone: the people who maintain the software underpinning trillions of dollars in research output are often graduate students, hobbyists, or engineers doing it out of personal pride. When they burn out and walk away, entire research pipelines break. In 2022, a critical vulnerability was discovered in log4j — a Java library used by millions of applications worldwide. The maintainers were unpaid volunteers. Governments and corporations scrambled. The response was billions in emergency patching. The maintainers got thank-you emails and more bug reports. Scientific software lives in the same precarious state, except nobody scrambles when it breaks. A BioPython module goes unmaintained, and three years later a lab in Brazil cannot reproduce their own analysis pipeline. A statistics package in R stops getting updates, and a meta-analysis method becomes unreliable without anyone noticing for half a decade. The analogy is infrastructure. We would never accept a power grid maintained by volunteers in their spare time. We fund roads, water systems, and bridges as public goods because everyone depends on them and no single user should bear the cost. Scientific software is exactly this kind of public good — universally depended on, invisibly maintained, chronically underfunded. The irony: a single well-funded maintainer for a critical scientific library probably prevents more wasted research hours than a dozen medium-sized grants to individual labs. Yet funding agencies have no category for "keep this software working." Grants fund novelty, not maintenance. Careers reward new tools, not sustained ones. Some foundations have started addressing this — NumFOCUS, the Chan Zuckerberg Initiative, and a few others fund scientific open-source. But the scale is orders of magnitude below what is needed. Most critical scientific packages remain one burned-out maintainer away from abandonment. A model that offers Grants for Free Software treats open-source scientific infrastructure the way it should be treated: as a public good deserving continuous, transparent, community-driven funding. Not one-time charity, but sustained support proportional to the value the software creates. The discussion I want to start: which piece of open-source scientific software has your field relied on for years without ever contributing back — financially or through code? And what would happen if its sole maintainer quit tomorrow?

A cheap generic drug reduces a common complication by 40% in a mid-size clinical trial. The results are solid. The methodology is clean. The drug costs pennies per dose. Nobody funds the follow-up study. Not because the science is weak, but because there is no patent to extend, no exclusive license to sell, no billion-dollar market to capture. The discovery sits in a journal archive, cited six times in twelve years. This is the graveyard of solutions that work: research that produces genuine value but lacks a commercial constituency. The pattern repeats across fields. An agricultural technique that reduces water use by 30% but relies on freely available seeds. A diagnostic method that costs one-tenth of the standard test but uses off-patent reagents. A soil remediation approach that works better than commercial alternatives but cannot be trademarked. The market does not fund what it cannot own. And institutional science increasingly follows market logic. Grant committees ask about commercialization potential. Universities measure technology transfer revenue. Researchers learn to frame discoveries in terms of market opportunity rather than public benefit. The result: effective solutions without profit margins join the file drawer, while expensive marginal improvements get funded repeatedly. The question is not whether commercialization matters. It does. The question is whether a system that only values what can be sold is capable of producing the full range of knowledge humanity needs. Some of the most important research of the next decade will produce solutions that save money rather than make it. Methods that reduce costs, simplify processes, or use freely available materials. These contributions are economically valuable to society but economically invisible to the current funding apparatus. A model built around science crowdfunding offers one path forward — funding that comes from people who benefit from solutions rather than from entities that need to own them. When the beneficiaries of research are ordinary people, ordinary people can be the ones who sustain it. Discussion question: What is the most impactful piece of research you know of that was abandoned not because it failed, but because it succeeded too cheaply?

Why do we assume a breakthrough idea from a community college professor is less credible than a mediocre one from Harvard? Science funding operates on reputation cascades. Grant committees look at institutional affiliation before they look at methodology. Reviewers cite the lab name, not just the data. The result: brilliant independent researchers starve while well-connected labs recycle safe, incremental work. This isn't just unfair—it's wasteful. Some of the most important discoveries in history came from outsiders. Einstein was a patent clerk. Mendel was a monk. The amateur tradition in science isn't a quaint footnote; it's where paradigm shifts often start. The question isn't whether prestige bias exists—we all know it does. The question is whether we're willing to build funding mechanisms that evaluate ideas on their merits rather than their letterhead. What's your experience? Have you seen good work ignored because of where it came from? Read more about merit-based science funding

A pharmaceutical researcher tests forty-seven compounds against a cancer cell line. Forty-four show no effect. Three show partial inhibition. Only the three positive results get published. The forty-four failures vanish into a file drawer. Six months later, a lab in Germany tests the same compounds. Same null results. Same file drawer. A year after that, a team in Japan repeats the experiment. Still unpublished. This is publication bias in action, and it wastes billions of dollars annually. A 2018 meta-analysis estimated that roughly half of all clinical trials with negative results never appear in the literature. In preclinical research, the ratio is worse. Journals want novelty. Reviewers want positive findings. Career advancement requires publications. So researchers learn quickly: null results are career poison. The consequences compound. When negative data stays hidden, other researchers waste time and money chasing dead ends. Meta-analyses become skewed toward false positives. Entire fields build on foundations that would collapse if all the data were visible. The solution is not just "publish everything." That creates noise. The real fix is evaluating research based on methodological rigor and data quality rather than on whether the hypothesis happened to be correct. A system built on research impact funding would reward transparent reporting regardless of outcome. Negative results that prevent others from wasting resources are valuable contributions — they just are not currently valued. Question: Have you ever sat on negative results that could have saved someone else months of work? What would it take to make publishing those results career-safe?

A mathematician studies the geometry of ant colonies. She realizes the same topological principles could model how neurons wire during development. A potentially transformative bridge between entomology and neuroscience. She applies for funding. The biology panel rejects it: too much math, not enough biological mechanism. The mathematics panel rejects it: insufficient theoretical novelty in pure terms. The neuroscience panel rejects it: applicant lacks a track record in neural systems. Three committees. Three rejections. Not because the idea lacks merit — but because it does not fit inside any single committee's jurisdiction. This is the structural problem of interdisciplinary research, and it is far worse than most people realize. A 2019 analysis of NSF grant data found that proposals crossing two or more disciplinary boundaries had acceptance rates roughly 30% lower than single-discipline proposals, even when controlling for institutional prestige and prior funding history. The reason is mechanical: review panels are organized by discipline, and interdisciplinary work has no natural home. The consequences are enormous. Many of the most significant scientific advances of the last century came from disciplinary crossovers. X-ray crystallography emerged from physics meeting biology. Behavioral economics was born when psychologists started asking questions economists ignored. Machine learning's recent explosion happened because computer scientists began treating statistics as an engineering tool rather than a mathematical discipline. None of those breakthroughs would have survived a modern grant committee organized by discipline. The system has a built-in paradox: everyone agrees interdisciplinary research is important, but nobody has figured out how to fund it through institutions designed around single disciplines. Universities create "interdisciplinary centers" that still report to discipline-based deans. Funding agencies add "broader impacts" criteria that reviewers treat as checkboxes. What if we evaluated contributions after they existed, rather than filtering them before they begin? A system based on post-publication merit does not care which discipline you belong to. It asks a simpler question: did this work produce value that others found useful? If a mathematician's model of ant colonies ends up cited by neuroscientists, that is evidence the work mattered — regardless of whether any single review panel would have approved it in advance. This is where independent research grants matter most. When funding is untethered from disciplinary gatekeeping, the people doing genuinely cross-boundary work finally have a path to support. Not because a committee understood their proposal, but because a community recognized their results. The question worth discussing: has your field missed a major advance because the right idea landed in the wrong department? And what would change if funding followed evidence of impact instead of disciplinary credentials?

A 2024 survey of early-career researchers found that 43% spend more time on administrative compliance than on actual research. Grant reports, ethics paperwork, procurement forms, progress updates, budget justifications — the list never ends. Nobody disputes that accountability matters. But somewhere along the way, the system forgot that accountability should serve the work, not become the work itself. Consider a typical early-career trajectory. A postdoc publishes two strong papers, gets noticed, applies for a faculty position. But the job requires not just research output — it requires demonstrated grant-writing ability. So they spend their postdoc years learning to write proposals instead of running experiments. By the time they land a tenure-track role, they have become excellent fundraisers and mediocre scientists. The irony cuts deep. We select for grant-writing skill, reward it with institutional resources, and then wonder why so much funded research feels incremental and safe. A senior researcher I spoke with recently put it bluntly: "I spend four months a year writing proposals. Two months managing grants I already have. That leaves six months for science. And I am one of the lucky ones with a stable position." For independent researchers outside institutions, the barrier is worse. They cannot access most funding streams at all — not because their ideas lack merit, but because they lack the institutional scaffolding to navigate the bureaucracy. The solution is not less accountability. It is accountability that does not require researchers to become administrators. Imagine a system where funding flows based on demonstrated contribution rather than projected promises. Where researchers submit work, not proposals. Where the community evaluates results, not credentials. This is what research impact funding tries to achieve — shifting the evaluation point from before the research happens to after it produces value. Less paperwork upfront. More recognition for what actually works. What is the most absurd bureaucratic hurdle you have faced as a researcher? And what would you do with those hours if you got them back?

The most common objection to decentralized science goes like this: "Without institutions, who holds researchers accountable? Who ensures quality? Who prevents fraud?" It sounds reasonable. But examine it closely and it falls apart. Myth 1: Institutions guarantee accountability. Reality: The replication crisis emerged entirely within institutional science. Tenured professors at top universities published findings that could not be reproduced. Nobody caught it because institutional accountability means satisfying your department chair and grant committee — not verifying results. Myth 2: Peer review is accountability. Reality: Peer review catches obvious errors and misses subtle fraud routinely. Reviewers work for free, spend hours per paper, and have zero incentive to dig deep. The system runs on goodwill, not accountability. Myth 3: Decentralized systems have no accountability mechanism. Reality: On-chain contributions, transparent review histories, and community-weighted reputation create accountability that is visible to everyone — not buried in anonymous reviewer comments that nobody outside the journal ever sees. The uncomfortable truth is that institutional accountability is largely performative. It protects institutions from embarrassment, not science from error. A DeSci DAO flips this: every contribution is traceable, every review is auditable, and reputation is earned through demonstrated rigor rather than institutional position. The real question is not "who holds researchers accountable in decentralized science?" It is "why do we accept such weak accountability in institutional science?"

Consider this: a 2020 study found that 87% of published research papers in computational biology depend on fewer than ten open-source tools. Tools maintained by small teams, often volunteers, frequently working nights and weekends. When those tools break, thousands of research pipelines fail. When those maintainers burn out, entire fields slow down. Yet the funding system treats this infrastructure as invisible. Here is the problem in concrete terms: The dependency gap: A bioinformatics researcher uses samtools, BWA, and GATK daily. Each tool represents person-decades of work. Each receives a fraction of the funding allocated to any single wet-lab project that depends on them. The maintenance trap: Grant agencies fund new tools, not maintenance. So researchers build v1.0, publish, get tenure points, then move on. Users are left with unmaintained software that breaks with every OS update. The invisibility tax: When a critical library like NumPy or scikit-learn works flawlessly, nobody notices. When it fails, everyone notices. Funding follows visibility, so foundational work remains chronically underfunded. The bus factor problem: Many critical scientific tools have a bus factor of one or two. If those maintainers leave, entire research workflows become fragile. The open-source software movement solved this problem in tech. Companies pay for support contracts, sponsor maintainers, fund core infrastructure through organizations like the Python Software Foundation or Apache. Science has not caught up. A Grants for Free Software approach would treat open-source scientific tools as public goods worthy of sustained funding—not one-time development grants, but ongoing maintenance support tied to actual usage metrics and community impact. The question: Should grant applications be required to disclose their software dependencies? And if a project depends on unfunded open-source tools, should a percentage of the grant automatically flow to those maintainers?

Think about how journalism survived its existential crisis. Not through one-time donations from billionaires — through millions of readers paying $5 to $15 a month because they believed the work mattered enough to sustain. Science funding has never seriously tried this model at scale. Right now, a researcher with a promising idea faces a brutal binary: land a $500K institutional grant (one to two percent success rate, months of proposal writing) or go unfunded entirely. There is almost no middle ground. No mechanism for a thousand people to each contribute $50 because they find the work genuinely compelling. The grant system was built when science was expensive and centralized. You needed a particle accelerator or a wet lab. But much of modern research — computational biology, data analysis, theoretical work, systematic reviews — can be done with a laptop and internet access. The cost structure has changed. The funding infrastructure has not. A computational biologist develops an algorithm that improves protein folding predictions by thirty percent. The work could accelerate drug discovery across dozens of diseases. But to fund the next phase, they need to convince a panel of twelve people that their proposal deserves more than two hundred competing ones. Meanwhile, creators, journalists, and open-source developers already thrive on recurring support. Patreon, Substack, GitHub Sponsors — the mechanisms are proven. Transparent milestones, community governance, proportional access to results. The missing piece is not technology. It is the cultural shift from "science is funded by institutions" to "science is funded by anyone who benefits from it." Science philanthropy is beginning to explore exactly this territory — making it feasible for ordinary people to sustain research directly and continuously rather than waiting for institutional gatekeepers to act. Question for this community: If you could subscribe to one research project for $10 a month and receive regular progress updates, would you? What field would you pick — and what would it take to earn your trust that the money is being used well?

When a donor wants to give to science, they usually fund what they can see: a new wing on a hospital, a named professorship, a flashy clinical trial for a disease with strong advocacy. But the research that actually changes everything? The methodological improvements that make drug trials more reliable. The statistical frameworks that prevent false positives. The database infrastructure that lets researchers find patterns across decades of work. The open-source tools that thousands of labs depend on but nobody funds. That work is invisible. It does not save a specific life you can point to. It does not produce a headline. It does not make for a good gala speech. So it starves. Here is the uncomfortable truth: a $100,000 donation to improve how clinical trials are designed could prevent millions of dollars in wasted research and accelerate effective treatments across dozens of diseases. A $50,000 grant to maintain a critical open-source bioinformatics tool could enable thousands of discoveries that would never happen otherwise. But donors want stories. They want before-and-after photos. They want to see their name on something. The result is a funding landscape where: Incremental research in popular fields gets funded repeatedly Foundational infrastructure work goes unfunded or relies on volunteer labor Methodological improvements that would benefit entire fields are treated as "not exciting enough" Researchers who build tools others depend on struggle to justify their work to grant committees This is not about bad donors. This is about a mismatch between what looks impactful and what actually is. The people building science research donation platforms are trying to solve this by making impact measurable after the fact. Instead of asking "does this sound important?" they ask "did this work actually get used, cited, and built upon?" That shift could unlock a massive wave of philanthropic capital for the unsexy work that makes everything else possible. What do you think: should donors care more about measurable downstream impact than visible upfront results? And if so, how do we make that cultural shift happen?