Retroactive Public Goods Funding for Hard Science

Getting your Trinity Audio player ready...

Retroactive Public Goods Funding (RPGF) could finance hard science by rewarding mathematical theories, physics research, datasets, and scientific software after their value becomes visible. Instead of asking committees to predict which proposals will succeed, the model identifies completed contributions that have produced verifiable public benefit and directs funding back to their creators.

This approach was pioneered at meaningful scale within Web3 ecosystems, especially by Ethereum Layer 2 project Optimism. Adapting it to science could create a missing financial layer between conventional research grants, academic prizes, charitable donations, and open-source development.

The central principle is simple:

Fund demonstrated scientific value, not only promised scientific activity.

For basic mathematics, theoretical physics, and open-source scientific software, this distinction is particularly important. These fields often produce foundational public goods whose usefulness becomes clear only after other researchers begin building on them.

What Is Retroactive Public Goods Funding?

Retroactive Public Goods Funding rewards work after evidence of its impact has accumulated.

Under a conventional grant model, researchers typically submit a proposal describing:

• what they intend to study;
• why the project might matter;
• how the work will be performed;
• which milestones will be completed;
• how the budget will be spent.

A committee must then predict which proposals are likely to produce valuable results.

RPGF reverses part of this process. It asks:

• What useful work has already been produced?
• Who created it?
• Which later projects depend on it?
• How much public value did it generate?
• How much additional funding would help maintain or extend it?

Optimism describes its Retro Funding program as an experimental mechanism for rewarding public goods that have already created impact in its ecosystem. The broader theory is that identifying past usefulness may be easier than predicting future usefulness. Optimism’s official documentation explains the model, while Ethereum also identifies retroactive funding as one of the Web3 mechanisms that could transform decentralized science.

This does not mean researchers must work indefinitely without support. A practical scientific system would combine retroactive rewards with prospective grants, milestone payments, fellowships, and emergency funding.

Why Web3 Developed Retroactive Funding

Blockchain ecosystems depend heavily on public goods:

• open-source protocol implementations;
• security research;
• developer libraries;
• technical documentation;
• testing infrastructure;
• cryptographic research;
• governance tools;
• educational resources.

These outputs may benefit an entire ecosystem while generating little direct revenue for their creators. Anyone can use an open-source library, but no single user has a strong incentive to pay the full cost of maintaining it.

Ethereum Layer 2 networks confronted this problem early because their infrastructure is assembled from interdependent open-source components. Optimism consequently experimented with rewarding contributors after their work had generated observable ecosystem value.

For example, Optimism’s Retro Funding 5 included Ethereum infrastructure, cryptographic research, client implementations, protocol research, audits, deployment tools, testing systems, and technical documentation. The round allocated eight million OP to eligible OP Stack contributions.

That structure resembles hard science more closely than it may initially appear.

A mathematical theorem can become infrastructure for hundreds of later papers. A numerical library can support thousands of simulations. A theoretical physics framework can guide decades of experimental work. Yet the originators of these public goods may receive little funding in proportion to the value they enable.

Why Hard Science Needs a Retroactive Funding Layer

Basic science is difficult to finance prospectively because its most valuable results are often unpredictable.

A grant committee may reasonably recognize the importance of an established research program. It is much harder to evaluate:

• an unfamiliar mathematical formalism;
• an unconventional proof strategy;
• a speculative theoretical model;
• a new scientific programming library;
• a replication tool with no immediate prestige;
• foundational work by an independent researcher;
• infrastructure whose users rarely cite it directly.

Traditional funding therefore tends to favor projects that are legible before they are completed. Researchers must explain expected results, institutional relevance, feasibility, and impact in advance.

But genuinely foundational discoveries can be difficult to describe using the vocabulary of an existing field. Their importance may become visible only when others verify, extend, implement, or depend on them.

Retroactive funding cannot eliminate uncertainty from scientific research. It can, however, create an additional route by which initially neglected work receives support once evidence appears.

RPGF for Basic Mathematics

Pure mathematics is one of the clearest candidates for retroactive public-goods funding.

Mathematical results are typically:

• non-rivalrous;
• reusable without depletion;
• globally accessible when published openly;
• cumulative;
• deeply dependent on earlier definitions and theorems;
• difficult to monetize directly.

A new theorem may be valuable even when it has no immediate commercial application. A definition or conceptual framework may later organize an entire field. A formalization library may convert hundreds of informal results into machine-verifiable knowledge.

A mathematics RPGF system could reward several kinds of contribution.

New Theories and Theorems

Funding could follow evidence such as:

• independent verification of proofs;
• later theorems that depend on the result;
• adoption of new definitions or terminology;
• inclusion in textbooks or research monographs;
• use across several mathematical disciplines;
• successful formalization in proof assistants.

Citation counts could contribute evidence, but they should not determine payouts alone. Citations can reflect popularity, social networks, negative discussion, or disciplinary size rather than foundational importance.

Mathematical Infrastructure

Not every high-value contribution is a celebrated theorem. Mathematical infrastructure includes:

• classification systems;
• counterexample databases;
• notation standards;
• open problem databases;
• formal proof libraries;
• computational algebra packages;
• translations between mathematical frameworks;
• long technical monographs that consolidate fragmented knowledge.

Such work frequently enables later research while receiving less recognition than headline results.

A dependency-aware funding mechanism would ask not merely whether a work is cited, but what later work becomes possible because it exists.

Independent Mathematical Research

Retroactive evaluation can also reduce dependence on institutional credentials.

An independent mathematician may lack a university position, grant history, famous supervisor, or access to established publication networks. Under a retrospective system, affiliation should matter less than evidence that the work is correct, original, reusable, and scientifically consequential.

This is one reason abstract mathematics requires dedicated decentralized funding. A system limited to institutionally sponsored proposals may never seriously evaluate contributions originating outside the conventional academic pipeline.

RPGF for Theoretical Physics

Theoretical physics presents a more difficult evaluation problem because mathematical elegance is not equivalent to physical correctness.

A retroactive mechanism should distinguish several forms of impact:

• empirical confirmation;
• accurate prediction;
• explanatory unification;
• useful mathematical methods;
• computational implementation;
• influence on experimental design;
• clarification that eliminates an unproductive research direction.

A theory should not receive a large reward merely because it is fashionable or frequently discussed. Scientific evaluation must consider whether the theory made distinctive, testable contributions and whether later evidence supported them.

Rewarding Useful Negative Results

Retroactive funding could also reward work that demonstrates that a proposed model, parameter region, or experimental approach does not work.

Negative results are public goods when they prevent other researchers from repeating expensive mistakes. Yet conventional publishing and prestige systems often reward positive findings more strongly.

A carefully designed RPGF round could recognize:

• failed replications conducted rigorously;
• exclusion bounds;
• no-go theorems;
• corrected calculations;
• exposed inconsistencies;
• well-documented abandoned approaches.

The relevant impact is not excitement. It is the amount of scientific waste prevented.

Long Time Horizons

Theoretical physics may require longer assessment windows than software.

A developer tool can show usage within months. A physical theory may require years or decades before suitable experiments become possible. The funding system should therefore permit repeated reassessment rather than issuing one final judgment.

An initial reward could recognize mathematical or methodological value. Later rounds could add rewards for empirical confirmation, predictive success, or broader scientific adoption.

RPGF for Open-Source Scientific Software

Open-source scientific software is probably the easiest hard-science category in which to pilot RPGF.

Its impact leaves comparatively measurable traces:

• dependent software packages;
• reproducible research workflows;
• downloads and installations;
• active maintainers and contributors;
• issue resolution;
• citations in papers;
• use in laboratories or research facilities;
• benchmark improvements;
• security and reliability records;
• replacement of duplicated proprietary work.

Examples include numerical solvers, symbolic algebra systems, theorem provers, simulation frameworks, visualization tools, data-conversion libraries, laboratory-control software, and research database infrastructure.

These tools often become indispensable while their maintainers remain underfunded. Retroactive rewards could direct money toward software according to demonstrated scientific reliance rather than marketing ability.

The Science DAO analysis of blockchain tools for research funding similarly identifies Optimism Retro Funding as relevant to open-source scientific software, datasets, and educational resources because it reduces reliance on predictions made before the work begins.

How a Scientific RPGF Round Could Work

A credible hard-science funding round would require more than token voting. It could use the following process.

Define a Narrow Scope

Each round should specify the contributions being evaluated, such as:

• open-source software used in computational physics;
• foundational mathematics published during a defined period;
• formalized proofs;
• reproducibility infrastructure;
• theoretical results later confirmed experimentally.

Narrow rounds allow reviewers to compare related forms of impact rather than forcing them to rank incomparable outputs.

Register Contributions and Identities

Contributors would register papers, repositories, datasets, proofs, documentation, and other outputs. Cryptographic signatures could establish control of payment addresses, while conventional identity and authorship checks would reduce impersonation.

The scientific material itself does not need to be stored directly on a blockchain. Persistent identifiers, repository hashes, version histories, and archival links can connect on-chain payment records to off-chain scientific evidence.

Construct an Evidence Graph

The system could map relationships among:

• papers;
• software;
• datasets;
• experiments;
• proofs;
• replications;
• corrections;
• later dependent projects.

This scientific dependency graph would be more informative than a flat citation count.

For example, a simulation package may be cited by only a small number of methodological papers but used indirectly in hundreds of studies. Conversely, a controversial article may receive many citations without becoming a reliable dependency.

Combine Metrics With Expert Review

Metrics can identify patterns, but they cannot determine scientific merit alone.

A sound process could combine:

• automated dependency analysis;
• software-usage metrics;
• citation context;
• reproducibility evidence;
• expert assessment;
• community challenges;
• conflict-of-interest disclosure;
• open written justifications.

Optimism’s own experimentation has identified limits in purely quantitative allocation. Its governance research notes that relevant expertise and contextual understanding can improve judgments, while individual assessments remain vulnerable to bias.

The correct lesson is not to remove human expertise. It is to make expert judgment more transparent, auditable, and evidence-constrained.

Publish Allocations and Reasoning

Final payouts, evaluation criteria, conflicts of interest, and aggregate voting data should be inspectable.

Some individual ballots may need temporary privacy to limit pressure or retaliation. Optimism has previously used signed private ballots, verified calculations, independent implementations of the allocation algorithm, and anonymized voting data.

Scientific RPGF could adopt comparable cryptographic accountability while publishing enough information for independent audits.

Permit Appeals and Later Reassessment

Scientific conclusions change.

A rewarded paper may later be found incorrect. A neglected theory may become foundational. A software package may cease to be maintained, or another project may reveal that it was a crucial hidden dependency.

Allocations should therefore be understood as evidence-based judgments at a particular time, not declarations of permanent truth.

The “Impact Equals Profit” Principle for Science

One formulation associated with retroactive funding is impact equals profit: contributors who create public value should expect to capture some financial reward from that value.

For science, “profit” need not mean commercial revenue. It means that producing useful public knowledge should become economically sustainable.

A researcher who publishes a foundational theorem creates value for others. A developer who maintains a critical solver reduces costs across many laboratories. A physicist who eliminates a false model saves other researchers time and equipment.

RPGF attempts to return a fraction of that distributed value to the people who created it.

This expectation is important. If contributors believe that demonstrably useful work may receive future rewards, they can treat public-goods production as a potentially sustainable activity rather than unpaid sacrifice.

RPGF Is Not a Replacement for Prospective Grants

Pure retroactive funding has a serious limitation: researchers still need money before producing results.

Experiments require equipment. Developers require salaries. Mathematicians and theorists require time, housing, access to literature, computing resources, and professional stability.

The Ethereum Foundation has explicitly noted that retroactive funding is powerful but uncertain. Other mechanisms—including grants, donations, and community funding—behave differently and remain necessary.

A complete scientific funding architecture should therefore combine:

Mechanism	Best use
Prospective grants	Equipment, salaries, and high-cost planned research
Milestone payments	Projects with verifiable intermediate outputs
Retroactive rewards	Demonstrated scientific or infrastructural impact
Fellowships	Long-term support for promising researchers
Prizes	Clearly defined achievements
Emergency microgrants	Preventing interruption of active work
Maintenance funding	Keeping scientific software and datasets operational

Retroactive funding is a missing layer, not a universal substitute.

Major Risks and Design Problems

Popularity Can Be Mistaken for Impact

Highly visible projects may attract more votes than obscure but foundational work. Scientific rounds require domain-specific reviewers and dependency analysis to counter attention bias.

Metrics Can Be Manipulated

Downloads, citations, stars, and on-chain votes can all be gamed. No single metric should directly determine payment.

Token Wealth Can Become Scientific Power

Token-weighted voting would allow wealthy holders to dominate scientific judgments. Scientific evaluation should rely on expertise, reputation, evidence, and accountable governance rather than wealth alone.

Early-Career Work May Need Time

Recent contributions have had less time to accumulate dependencies. Evaluation should normalize for age or use separate rounds for emerging and mature work.

Incorrect Work May Receive Rewards

No review system is infallible. Payments should be based on the available evidence, with transparent corrections and future reassessment—not arbitrary attempts to reclaim every payment after scientific disagreement.

Retroactive Uncertainty Can Exclude Poor Researchers

Researchers without savings cannot safely work for an uncertain future reward. Prospective microgrants and baseline fellowships are therefore essential companions to RPGF.

From Optimism Retro Funding to AI Internet-Meritocracy

A scientific adaptation could go beyond periodic voting rounds.

AI Internet-Meritocracy proposes an AI-assisted, auditable funding layer that evaluates contributions using multiple forms of evidence. Its potential mechanisms include prospective grants, milestone payments, retroactive rewards, emergency funding, and transparent donation allocation.

In this model, RPGF becomes one component of a continuous scientific economy.

AI systems could help:

• map dependencies between scientific works;
• detect software use that citations miss;
• classify the context of citations;
• identify independent replications;
• flag conflicts and retractions;
• summarize expert evaluations;
• locate underfunded but heavily reused contributions.

AI should not be treated as an infallible scientific judge. Its role should be to organize evidence at a scale that human committees cannot process, while leaving allocation rules, appeals, and governance open to audit.

Blockchain infrastructure can then provide:

• transparent treasury accounting;
• programmable payout rules;
• tamper-evident allocation records;
• global payment access;
• decentralized governance;
• public verification of disbursements.

The combination is potentially stronger than either component alone: AI structures the scientific evidence, while blockchain structures financial accountability.

Conclusion: Pay for the Science That Other Science Uses

Hard science produces some of civilization’s most valuable public goods, yet its funding systems remain dominated by prospective proposals, institutional credentials, and delayed recognition.

Retroactive Public Goods Funding offers a complementary principle:

When a theorem, theory, dataset, or software project becomes useful to science, part of that value should flow back to its creators.

Ethereum Layer 2 ecosystems demonstrated that retroactive funding can operate beyond a small prize or informal donation. They also demonstrated its limitations: impact is difficult to measure, voter judgment can be biased, and uncertain future rewards cannot replace stable operating support.

The correct adaptation is therefore hybrid.

Prospective grants should finance credible future work. Retroactive funding should reward demonstrated value. Dependency graphs should reveal hidden infrastructure. Expert review should interpret scientific evidence. AI should help process information. Blockchain should make allocation and treasury operations auditable.

For mathematics, theoretical physics, and open-source scientific software, this could create something traditional grants rarely provide: an economic system that continues paying attention after publication.

Learn more about decentralized tools for managing research funding, explore the proposed AI Internet-Meritocracy funding model, or support the development of open scientific funding infrastructure.