Gitcoin

🔥 New Gitcoin site just dropped Why? The funding landscape is fragmented. Lessons aren't being shared. Experiments are happening in isolation. So we built a living reference: the tools, the teams, the research, the results. All in one place. Turning hard-won experience into something everyone can use → gitcoin.co Here's what's on it 👇🧵

Funding what matters requires civilizational-scale infrastructure – not a single large platform, but a pluralistic ecosystem of coordinated tools. The threats defining this century demand a response bigger and more distributed than any one institution or mechanism can meaningfully address alone. Decentralized funding is becoming a structural requirement, which means the landscape is now something we collectively hold. 💚 🔗 ↓ gitcoin.co/research/plura…

Onchain funding's often treated as niche infrastructure for open source, but it’s really a live rehearsal for distributed coordination. Climate change, AI risk, institutional erosion – each requires strangers to align priorities, incentives, and resources at civilizational scale. Public goods funding gives that alignment a live proving ground: quadratic funding surfaces community priorities, retroactive funding rewards proven work, and coalitional funding amplifies impact through visible commitments. Funding open source matters. Learning how to coordinate matters too. Especially before crisis turns our need for alignment into an excuse for consolidating control. @owocki on what's at stake ↓ gitcoin.co/research/civil…

d/acc funding doesn't depend on what the Fed decides tomorrow. Build anyway 💚

Bees don't talk to each other 🐝 They coordinate through traces left in a shared environment that signal what to do next. That’s stigmergy: indirect coordination through environmental signals. Open source already works this way: bugs, TODOs, unfinished features. Onchain systems do too: treasury balances, proposals, protocol states. And in a world where AI agents can read and respond to signals autonomously, stigmergy is a way to turn shared environment into shared direction at machine speed. Read more ↓ 🔗 gitcoin.co/mechanisms/sti…

The way we collectively make sense of the world was built for a slower one. Centralization, stratification, bureaucratic control – these tactics work best when the problem is cleanly defined. But when problems become entangled or increasingly complex, these tactics begin to break down. Misinformation, polarization, institutional decay – we're feeling the impact of this breakdown in real time. What we need now is a new model for collective understanding that can adapt faster than these legacy responses fail. 💚 We have some ideas ↓ gitcoin.co/research/a-net…

When a handful of firms become the interface between billions of people and our understanding of the world, the risk of epistemic capture becomes material. When AI extends that mediation from what we see to how we reason, the risk becomes existential. At that point, the central question shifts from who gets to build the models to who gets to define reality itself. 🛡️ To make sure that agency stays ours, we need infrastructure that keeps collective sensemaking in human hands ↓ gitcoin.co/research/colle…

‼️ FINAL 24 HOURS: Public Goods Network (PGN) shuts down tomorrow Monday April 20 at 9 AM EST Please withdraw any remaining funds before this deadline. Unclaimed funds will move to the Gitcoin Foundation treasury and will not be recoverable after shut down. ✅ Easy claim app (by @owocki): pgnclaim.vercel.app 📜 Governance post: gov.gitcoin.co/t/temp-check-w… 🗳️ Snapshot vote (passed): snapshot.org/#/s:gitcoindao… 📖 Claim instructions: docs.publicgoods.network/using-pgn/clai… Thank you for being part of the PGN experiment with us 💚

over the last 8 years, @gitcoin has helped fund the internet’s public goods millions of contributions tens of thousands of builders $70M+ deployed to what actually matters not hype not speculation real funding for real infrastructure open source climate coordination systems the stuff everything else depends on and we’re just getting started if you’re building something that matters or want to support those who are help fund what matters ↓ gitcoin.co

AI didn’t create the need for defensive infrastructure, but it makes every part of it more urgent at the same time. It’s intuitive to think about risk domain by domain: separate problems, separate solutions, separate funding. But AI doesn’t respect those boundaries; it reduces the cost of applying pressure across multiple domains at once. Which means the systems we rely on aren’t just important individually – they're interdependent. A governance system only works if the identities participating are real. ⬇️ Identity systems only work if the cryptography beneath them holds. ⬇️ Cryptography only matters if the physical infrastructure running it stays online. When one layer fails, the layers above it collapse. And when pressure's applied across every layer simultaneously, defense has to be simultaneous too 🤝

Governance post ↓ 🔗 gov.gitcoin.co/t/temp-check-w… Snapshot vote ↓ 🔗 snapshot.org/#/s:gitcoindao… Steps to claim funds ↓ 🔗 docs.publicgoods.network/using-pgn/clai… Thank you for being a part of PGN 💚

❗ Last week to withdraw your funds from Public Goods Network (PGN) PGN is shutting down Monday April 20, 9 AM EST. Unclaimed funds will be withdrawn to the Gitcoin Foundation treasury after that deadline. @owocki app to help ↓

@MrPicule Love to see it 💚 Stay tuned. You might like what we have cookin with d/acc Fund.

@gitcoin I suppose that’s why there are tech leaders and founders with a clear vision. I’m new to this field and am just trying to build my own protocol, so I can’t answer that question

@MrPicule Agreed. Diversity in experience, perspective, and the redundancy afforded by collective expertise is critical. Now if we could just coordinate that effort in a distributed way. Protocol?

That is precisely why this task is entrusted not to a single person but to large teams of professionals; they cannot afford to make mistakes. A prime example is the selection of targets for attack using AI, this clearly illustrates that programmers must do their utmost to ensure the AI has no chance of making a mistake, which means the issue here is more one of human competence

In 2024, Romania annulled its presidential election. An obscure candidate had surged from 1% in polls to win the first round, boosted by AI-generated deepfakes on social. The court voided the result. That same year, a finance worker at Arup wired $25 million to scammers after a fake request from his CFO. He'd verified on a video call with his executive team before sending. Every person on that call was AI-generated. In early 2026, the bet that AI agents will make human software workers redundant wiped $2 trillion from SaaS market cap. Not a cyberattack or a recession. An economy repricing human work as nonessential. Our current systems weren't designed to protect us from the acceleration we're experiencing. It's time to build ones that are 🤝

@MrPicule 💯 Trust infrastructure has to evolve at the same speed as the tech on top of it, but that has been easier said than done. Are agents bound to inherit our coordination failures if we train them? It would be good to prevent those failures from compounding at machine speed 👀

The tech is moving faster than the governance and that is a recipe for chaos. We are watching the transition from human trust to cryptographic verification in real time. If the code isn't open source, it is essentially a black box of risk for your capitol. Can we even trust the agents to build the new systems without baking in their own bias

@ethereumboulder 🔜

@gitcoin 👀👀

@savantchat Strong defensive use case. AI scales coverage in ways humans can't. A risk could be AI approximation in place of precision, when vulnerabilities tend to live in that margin. How do you recommend verifying the verifier?

@gitcoin It makes audits better, faster, and adds more observation angles, eventually making the whole space safer and more reliable

The world's racing toward artificial and autonomous intelligence, and systemic risk's accelerating right beside it. But AI isn't inherently adversarial – it's a force multiplier. It just increases whatever we point it at. And when it’s pointed at areas we're still sorting out – digital frailty, resource scarcity, democratic fragility – risk can compound in the wrong direction. We're at an inflection point, and the question is as simple as it is critical: Can we steer an agentic future toward coordinated benefit before the threat surface multiplies beyond us? d/acc is how we get to yes. 💚

@MrPicule 10/10 no notes. > Hardest part is governance of the protocol itself A tale as old as blockchain. Maybe governance itself eventually becomes agentic?

Same way Bitcoin handles it. No coordinator, just protocol. Nodes don't trust each other, they trust the rules. Apply that to defensive priorities: consensus mechanisms over shared threat intelligence, not a central authority deciding what counts as a threat. The protocol becomes the coordinator. And there's a second-order gain here: when AI operates on explicit, consensus-driven rules rather than opaque centralized logic, it becomes more predictable and auditable. That predictability is itself a safety layer. Easier to verify, easier to constrain, easier to course-correct. Hardest part is governance of the protocol itself, but that's a solvable problem compared to trusting any single entity with the keys

@MrPicule Smart 🫡 With decentralized inference as the foundation, the next question is how millions of nodes coordinate toward shared defensive priorities. i.e. how do we keep coordination decentralized too, and prevent any one entity from becoming the coordinator?

Decentralization of inference itself. Bitcoin-style validator architecture but for AI. Millions of nodes, interchangeable, globally distributed. Shifts the model from a monolithic decision center to redundant microservices. Single point of failure disappears. The tradeoff is energy cost, but the gain is both resilience and accessibility. No single actor owns the kill switch. That's what "decentralized enough" actually looks like in practice