Devansh Mehta

Ethereum Cypherpunk Congress · Aug 9, 2026

How do we sustainably fund the public goods that everyone depends on? Devansh Mehta shares why Ethereum needs to move beyond grants toward embedded funding mechanisms that continuously support the ecosystem. We discuss public goods, selective disclosure, the economics of privacy, quadratic funding, and why the next generation of cypherpunks is focused not only on protecting information—but also on protecting the free flow of value. Timecodes 00:00 Introduction 00:25 Why Ethereum needs sustainable public goods funding 02:10 What makes something a public good? 03:16 Privacy, selective disclosure & the value capture dilemma 05:42 Funding models: patrons, grants & quadratic funding 08:30 How anyone can support Ethereum public goods 11:29 Favorite public goods projects & funding tools 12:39 The new generation of cypherpunks 13:33 Ethereum's future, EF mandate & building independent ecosystems 16:48 A vision for the future of work - This conversation is part of Privacy Academy Interviews, a peer-to-peer knowledge project exploring the most important ideas in Web3, privacy, and digital freedom. Learn more: https://academy.web3privacy.info A W3PN Media House production. Directed by Federico Marchi https://x.com/BabyBitProd Interview by Peter Farbey https://x.com/0xfarbey Let privacy be with you!

ETHGlobal · Jul 9, 2026

In this talk, Artem Brazhnikov from Octant declares Cypherpunk dead and immediately revives it, framing crypto's institutional era as the beginning rather than the end of the movement that started in the 80s and 90s around privacy and cryptography. Drawing on Buckminster Fuller's quote that you never change things by fighting reality but by building a new model that makes the existing one obsolete, Artem argues the path forward is building bridges to institutions rather than burning them. He acknowledges nostalgia in the community for the 2018 era of hoodies and Berlin hard forks while insisting Ethereum is the same substrate whether you wear a hoodie or a collar shirt. Artem identifies architectural vulnerabilities in security, privacy, and identity as the central problem, citing 2025's $292 million Kelp DAO exploit, $13 billion withdrawn from Aave in 48 hours, and over $800 million stolen across the year. He warns that AI is rewriting the threat model with bugs being found exponentially faster (including a Zcash bug sitting since 2022 only recently found by AI), and that traditional institutions built for human attackers cannot keep up with billions of agents executing trillions of transactions at light speed. The only answer he sees is zero knowledge cryptography that proves computation without revealing inputs. He argues paradoxically that the architectural decisions Cypherpunks have made for decades (decentralization, credible neutrality, verifiability, operational resilience demonstrated by Ethereum's 11 years of zero downtime) are exactly what institutions now need. He closes by presenting Octant's model: a self sustaining funding engine where the Golem Foundation's 100,000 staked ETH generates yield to fund ecosystem grants (over 130 projects and $18 million distributed in 3 years) without ever spending the principal. He introduces Octa Vaults built on ERC-4626 as perpetual endowments enabling yield powered giving, Octa Grants as the allocation layer using quadratic funding and other models, and the V2 launch coming next week. He reframes web3 as not about ownership but co-ownership of shared state, since users own keys not assets, and that public infrastructure must stay funded for everyone to keep 24/7 access. 00:00 Introduction 00:08 What Is Cypherpunk 00:33 Cypherpunk Is Dead, Long Live Cypherpunk 01:23 Build a New Model Instead of Fighting the Old 01:52 Adoption Is Not the End of Cypherpunk Ethos 02:26 Nostalgia for the 2018 Berlin Hard Fork Era 02:50 Security and Privacy Are Architectural Problems 03:14 Honeypots and Bridge Vulnerabilities 03:38 2025's Exploits: Kelp, Aave, $800M Stolen 04:11 AI Is Rewriting the Threat Model 04:40 The Zcash Bug Found by AI After Three Years 05:07 Why Institutions Built for Human Attackers Will Fail 05:33 The Old Model: Service Agreements and Legal Recourse 05:57 Why Old Institutions Cannot Keep Up With Agents 06:22 Zero Knowledge Cryptography as the Answer 06:52 Cypherpunk Tech Is What Wall Street Needs 07:16 Credible Neutrality and Counterparty Risk 07:41 Verifiability and Ethereum's 11 Years of Uptime 08:20 Why the Ethereum Foundation's Role Is Narrowing 08:41 The Need for New Coordination Models 09:18 Digital Asset Treasuries as Long Term Funders 09:46 Octant: A Self Sustaining Funding Engine 10:11 How 100,000 Staked ETH Powers Grants 11:00 130 Projects, $18 Million Distributed in 3 Years 11:29 Octa Vaults as Perpetual Endowments 12:32 Yield as Insurance Against Loss 12:56 Yield Power Giving Without Spending Principal 12:56 Octa Grants and the Allocation Layer 13:54 V2 Launch Next Week 14:20 Web3 Is About Co-Ownership, Not Ownership 14:50 You Own Keys, Not Assets 15:20 Preserving the Properties We Built On 15:45 ZK-EVM and Verifiable AI Inference 16:14 The Moment Is Ripe for Institutions 16:48 Closing: Building Together for the Next Thousand Years _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ *ETHConf 2026* ETHConf is a 3 day event for founders, industry leaders, and builders who are excited about the possibilities of building on top of Ethereum. Connect with 2,000+ top innovators in crypto, finance, technology, and policy at our inaugural three-day event packed with showcases, demos, partnerships, and conversations shaping the future of the global economy. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ✅ *Follow ETHConf* X: https://x.com/ethconf Website: https://ethconf.com _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 🎟️ Get your early bird tickets for ETHConf 2027: https://ethconf.com/2027#tickets 🎤 View the full ETHConf 2026 Speaker Schedule: https://ethconf.com/schedule _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Ethereum Cypherpunk Congress · Jun 9, 2026

When governments shut down the internet, millions lose more than connection. They lose access to each other. The Tor ecosystem builds open source tools that help people stay connected, communicate safely, and protect freedom online. Support the projects protecting digital freedom: https://blog.torproject.org/fund-internet-freedom/

Ethereum Cypherpunk Congress · Jun 9, 2026

Some questions are difficult to ask out loud. Around the world, people use privacy tools to search, learn, connect, and find community safely online. The Tor ecosystem supports open source projects protecting the freedom to explore identity, communicate privately, and stay connected without fear of surveillance. .... Support the projects from OONI to Securedrop protecting digital freedom: https://internetfreedom.torproject.org

Funding the Commons · May 9, 2026

AI governance. The future of work. Coordination at scale. And the question underneath all of it: how do we make sure AI leads to human flourishing, not just faster automation? Intelligence at the Frontier was a 36-hour festival hosted by Funding the Commons across 16 floors of Frontier Tower in San Francisco, March 14–15, 2026. Over a thousand participants from 92 countries joined 229 sessions spanning AI policy, public goods funding, decentralized governance, AI safety, open source AI, democratic deliberation, longevity research, and coordination infrastructure for the age of AI. This aftermovie captures some of the ideas and conversations that shaped the weekend, featuring Tom Kalil (Renaissance Philanthropy, fmr. White House science policy), Sheila Warren (fmr. World Economic Forum), Shady El Damaty (Holonym Foundation / human.tech), Luke Drago (Workshop Labs), and Shaw Walters (ElizaOS). What stood out to you? Let us know 👇 🌐 https://fundingthecommons.io Presented by Funding the Commons x Frontier Tower x Protocol Labs

Funding the Commons · May 9, 2026

What does it look like to build the human-AI coordination layer in 36 hours? The Intelligence at the Frontier Hackathon brought 150 builders to Frontier Tower in San Francisco for an overnight build sprint on March 14-15, 2026, across four tracks: agentic AI, physical robotics, AI safety and evaluation, and sovereign infrastructure. 53 projects shipped. $25k+ in prizes awarded. We saw builders hacking through the night, robots roaming the hallways, agents paying each other in real money, and judges giving Claude feedback. Thanks to everyone who joined us. Co-produced with Protocol Labs. Supported by ElevenLabs, Bittensor, Solana, Meteora, Metaplex, Redwood North, BittensorCommons.org, VESSL AI, Velda, Unbrowse, NomadicML, human.tech, Activeloop, Solo Tech, Learning Layer Labs, Lit Protocol, and KikiTora. 🌐 https://fundingthecommons.io 📋 All 53 projects: https://intelligence-at-the-frontier-hackathon.devspot.app/?activeTab=projects

Funding the Commons · May 9, 2026

Pi belongs to everyone. E=mc² didn't. Tricia Wang opens Intelligence at the Frontier 2026 asking which path AI governance will take. Serving as Emcee of the festival and co-founder of the Advanced AI Society, Tricia frames the day through a Pi Day provocation: pi was held as a commons across civilizations for millennia, owned by no one and available to all. Einstein's formula took the opposite path, weaponized by one government before any governance existed to distribute its power. The parallel to AI ethics and agentic AI in 2026 is direct and uncomfortable. She introduces a three-layer framework for navigating the AI governance conversation: hardware, software, and the social/cultural layer that's almost always missing. The shaping question for human flourishing in the age of AI: as AI agents begin to outnumber humans, do we see them as extensions of ourselves or adversaries to control? Emcee of the festival: Tricia Wang (Advanced AI Society) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

David Dao

Funding the Commons · May 9, 2026

Elinor Ostrom proved communities can self-govern shared resources without top-down control. But her principles hit a hard limit at roughly 150 people — Dunbar's number. Beyond that, cooperation breaks down. So how do you govern something the size of the Amazon? David Dao, founder of GainForest and winner of the Rainforest X Prize, argues that AI and decentralized technology are the missing infrastructure to scale Ostrom's commons governance principles to planetary scale. Drawing on 15 years working with indigenous communities across the Amazon and the Philippines, he shows how satellite imagery, data valuation, and on-chain identity systems can help communities protect forests at a scale no single government or NGO has managed alone. The talk moves from the tragedy of the commons to a blueprint for governing shared resources in the age of AI. Speaker: David Dao (Protocol Labs & Gainforest) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Funding the Commons · May 9, 2026

Losing your job to AI is the obvious worry. Luke Drago argues the real problem is bigger. Luke, co-founder of Workshop Labs and co-author of The Intelligence Curse, borrows a concept from development economics: the resource curse, where oil-rich states stop investing in their citizens because their wealth comes from the ground, not from taxing labor. His argument is that full automation creates the same dynamic. When a handful of actors own the models, the compute, and the data, there's no economic reason left to invest in everyone else. The social contract quietly stops working. Luke makes the case that this outcome isn't inevitable, and that the window to build things differently is still open. Speaker: Luke Drago (Workshop Labs / Blue Dot Impact) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Funding the Commons · May 9, 2026

AI isn't just displacing jobs. Sheila Warren argues it's accelerating something older and more dangerous: the systematic removal of human dignity from how decisions get made. Sheila, CEO of the Project Liberty Institute and founder of the World Economic Forum's blockchain and digital assets team, connects what look like separate crises — surveillance tech, deportation without due process, AI agents — into a single cultural problem. Her central concept is differential friction: the idea that the race to remove friction from every system is quietly eroding the checks that keep power accountable. When there's nothing to slow an AI agent down, small errors compound fast and become permanent. She encourages everyone building in this space to ask, every time: before you remove friction, whose protection were you removing with it? Speaker: Sheila Warren (Project Liberty Institute) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Funding the Commons · May 9, 2026

National governments move too slowly for emerging technology. Augusto Ardiles Diaz argues that cities are where AI policy actually gets built. Augusto, Undersecretary of Investments for the City of Buenos Aires, lays out the case for cities as real-world laboratories: places where new technology can be tested, adopted, and scaled under real conditions before formal regulation catches up. Buenos Aires is doing this through a dedicated AI district, offering zero local taxes for qualifying companies, regulatory sandboxes supervised by public authorities, and university partnerships designed to reskill workers and small businesses ahead of the curve. His framing is that public policy should function as institutional infrastructure for innovation, not just as a brake on it. Speaker: Augusto Ardiles Diaz (Gobierno de la Ciudad de Buenos Aires) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Matthias Broner

Funding the Commons · May 9, 2026

As AI displaces workers and traditional UN funding gets pulled back, UNICEF is building the financial infrastructure to get resources to people who need them most. Matthias, an advisor to the UNICEF Office of Innovation, walks through a set of live blockchain-powered initiatives: a crypto donation protocol (Drips) funding five digital public goods across 131 countries, an impact staking program that puts idle ETH to work for humanitarian causes, and AidLink — a system that delivers cash transfers via SMS to people without internet access, building digital wallets entirely behind the scenes. UNICEF has held and disbursed crypto assets since 2019, longer than any other UN organization. The goal is to leverage blockchain as financial infrastructure for people the traditional system doesn't reach, tested in the real world before being scaled. Speaker: Matthias Broner (UNICEF Office of Innovation) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Shady El Damaty

Funding the Commons · May 9, 2026

As AI agents increasingly mediate human interaction, how do we make sure people stay in control of their own digital lives? Shady, co-founder of Holonym Foundation and the team behind human.tech, comes from computational neuroscience and applied cryptography. He argues that the current trajectory of AI development pulls resources, talent, and capital into a one-way infrastructure that gives little back — and that cryptographic tools, specifically zero-knowledge proofs, offer one of the most underused levers we have to redistribute power back to individuals. The core question he poses: what is the most important thing you could be building right now, with everything you have at your disposal? For Shady, the answer is technology that proves you are human, without giving up what makes you so. Speaker: Shady El Damaty (Holonym Foundation / human.tech) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Funding the Commons · May 9, 2026

For the first time in 100 years, the Los Angeles City Council voted to expand its size, from 15 to 25 seats. That decision came from a citizen assembly that Liz Barry helped design and run. Liz, Executive Director of Metagov, builds the infrastructure for large-scale democratic deliberation. She works across community organizing, open source software and governance theory, and has been involved in democratic reform processes from Los Angeles to Taiwan. Her provocation is simple: what determines whether technology is a net good or bad is not its functionality — it's who controls it. She argues that as AI reshapes institutions, deliberative platforms need two things that are currently in tension: proof of humanity to ensure results are legitimate, and data sovereignty so participants can engage safely. Figuring out how to hold both at once is one of the defining challenges of democratic governance in the age of AI. Speaker: Liz Barry (Metagov) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Funding the Commons · May 9, 2026

A single state-level debate about one nuclear power plant generated 5,862 distinct arguments. The human mind can hold about four to seven things at once. Jamie Joyce argues this gap is the core problem in democratic governance, and that AI might be the only way to close it. Jamie, founder of the Society Library, makes the case for Scaled Democratic Reasoning: the idea that civilization-scale decisions involve far more complexity than any group of humans can process reliably. Her team maps debates as dense, nested argument graphs pulling from social media, scholarly papers and economic assessments, to make the full landscape of a debate navigable at scale. Her argument isn't that human deliberation is wrong. It's that we need tools to extend it, the same way telescopes extended vision. Speaker: Jamie Joyce (The Society Library) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Funding the Commons · May 9, 2026

We built tens of thousands of physical libraries to give people access to knowledge. Sam Klein asks what the equivalent looks like when agents are doing half the reading. Sam, co-founder of Public AI and director of the Concordance Project at Harvard's Berkman Klein Center, makes the case for a new kind of planetary library — one where AI agents don't just retrieve information but actively curate, annotate and contribute to shared knowledge commons. He estimates there are roughly 3 billion concepts worth documenting; we have monographs for about 200 million of them. The gap is enormous, and agents could help close it. His ask to builders in the room: invite your agents to do knowledge curation work. Every small piece of structured information recorded at scale becomes part of a library the whole ecosystem benefits from. Speaker: Sam Klein (Public AI / Berkman Klein Center, Harvard University) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Joshua Tan

Funding the Commons · May 9, 2026

AI has become roughly 1,250 times more effective over the past six years. Organizations haven't kept up. Joshua Tan thinks that's the next frontier. Joshua, co-founder and research director at Metagov and lead at Stanford's Open Lab, argues that as AI agents grow more capable, the organizations coordinating them need to evolve too. He introduces the concept of autonomous organizations — systems like the Facebook newsfeed or agent-run code repositories that operate through algorithmic logic no single person fully controls. DAOs were an early attempt at this, but their rigidity made them slow to change. What comes next is something more adaptive, more sub-symbolic, and harder to govern. His proposal: a startup incubator model for autonomous organizations, where new governance structures are launched as live experiments and learned from in real time. Speaker: Joshua Tan (Metagov / Stanford Open Lab) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Funding the Commons · May 9, 2026

AI is advancing exponentially. Democratic systems are not. Aviv Ovadya is trying to close that gap before it becomes irreversible. Aviv, CEO of the AI and Democracy Foundation and research director at Metagov, makes the case for deliberative processes as core infrastructure for governing AI. Citizen assemblies and similar mechanisms have now been run thousands of times around the world, including feeding into EU policymaking several times a year. They are faster to spin up than elections, easier to iterate on, and designed to cut through the adversarial incentives that corrupt traditional political processes. His argument is that the democratic innovation space is massively underfunded relative to AI, crypto and other domains, and that a targeted ARPA-style deployment of resources across key capability gaps could make rapid progress. Speaker: Aviv Ovadya (AI and Democracy Foundation) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Funding the Commons · May 9, 2026

Humans aren't impressive on their own. Our superpower is each other — and institutions are how we've always organized that collective intelligence at scale. Adam Russell thinks that's exactly why institutions matter so much right now, and why most of them are blowing it. Adam, a former DARPA program director and Chief Vision Officer at USC's Information Sciences Institute, comes to AI as an anthropologist. He sees two dominant responses to disruption in institutions today: pretend it isn't happening, or dig in and defend the status quo. Neither works. His argument is that real resilience means being able to change when the world changes — not holding firm while everything shifts around you. Speaker: Adam Russell (Information Sciences Institute, USC) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Funding the Commons · May 9, 2026

Most of the conversation around AI asks what it will do to us. Tom Kalil thinks that's the wrong question. Tom, who shaped science and technology policy across the Clinton and Obama administrations, and Adam Russell, former DARPA program director and Chief Vision Officer at USC's Information Sciences Institute, dig into what it actually takes to steer AI toward public benefit. Tom's central argument is that everyone is focused on AI itself, but not nearly enough on its complements — the data sets, benchmarks, platforms and infrastructure that determine whether AI can solve anything meaningful. AlphaFold worked because of 50 years of curated protein data. Most domains don't have that. The fireside chat covers advance market commitments as a funding model for public goods, why good ideas from the crypto and blockchain world get stuck inside that world, and what government can and can't do when technology moves faster than policy. Speakers: Adam Russell (Information Sciences Institute, USC) and Tom Kalil (Renaissance Philanthropy) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Funding the Commons · May 9, 2026

45 million people depend on a network of internet freedom tools to access uncensored information. When the US government rescinded its funding mid-contract in 2025, most of those tools faced collapse overnight. Al, Director of Fundraising at the Tor Project, makes the case for why centralized funding is the Achilles heel of decentralized infrastructure. Tor's network is distributed by design — thousands of volunteer relay operators across the world, no single point of failure. But its funding came from one source. When that source changed its mind, the entire ecosystem was put at risk. His answer is to bring quadratic funding and community crowdfunding models from the Ethereum world into the internet freedom space. Tor is running a campaign in May 2026 using minimal anti-collusion infrastructure to make funding more equitable and resistant to manipulation — reaching 2.2 million Tor Browser users a day in the process. Speaker: Al Smith (Tor Project) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Funding the Commons · May 9, 2026

In 1999, Sun Microsystems CEO Scott McNealy told the world: "You have zero privacy. Get over it." Michelle Finneran Dennedy was working at Sun at the time. She went the other direction. Michelle, Chief Data Strategy Officer at Abaxx, created the first Chief Privacy Officer role at Sun Microsystems and has spent nearly 30 years building the field of privacy engineering. She argues that privacy has a precise definition — the authorized processing of personal data according to ethics, law and sustainable commerce — and that authorization is contextual. In a world where machines are acting on machines at scale, that context disappears fast. Her framework for fixing it is proofs of control: measurable, nano-level signals built into systems from the start, not layered on top after the fact. The privacy officer role, she argues, needs to move from compliance into engineering and into the boardroom before agentic AI makes the retrofit impossible. Speaker: Michelle Finneran Dennedy (Abaxx) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Funding the Commons · May 9, 2026

On February 7th, 2026, NPR reported that the State Department was taking down its social media posts going back 20 years. The Internet Archive had already started archiving them. In the weeks that followed, the Wayback Machine captured more than 2 million posts and found that over a million had already been removed from the live web. This panel brings together Mark Graham, director of the Wayback Machine, Megan Kleiman of the Filecoin Foundation, and moderator Wendy Hanamura to talk about what it means to preserve public data at a moment when governments are actively deleting their own records. The conversation covers Democracy's Library, a joint project storing government data from the US, Canada, Bermuda and beyond in decentralized Filecoin storage; why the Internet Archive's 200 petabytes of public data is one of the most important training sets AI has access to; and what builders can do with it right now. Mark also makes a case for archiving AI chatbot responses themselves — because if no one records what these systems are saying, that history disappears too. Speakers: Mark Graham (Internet Archive), Megan Kleiman (Filecoin Foundation) Moderator: Wendy Hanamura (DWeb / formerly Internet Archive) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io Note: This recording was captured with technical limitations on Day 1. We apologize for any audio quality issues and appreciate your patience.

Evan Miyazono

Funding the Commons · May 9, 2026

Imagine a box where an AI system learns everything about your values and preferences. You walk in, spend a few hours, and walk out with a perfectly aligned AI. Evan Miyazono's talk starts here — and then explains why that's still not enough. Evan, founder of Atlas Computing and former director of Funding the Commons, argues that for high-stakes domains like nuclear infrastructure or financial systems, even a 99% probability guarantee isn't acceptable. Alignment is necessary but not sufficient. What's needed alongside it is structure: explicit checks and balances, verifiable constraints, and institutions with clearly scoped responsibilities — the same logic behind how democracies are designed to prevent unchecked power. His proposal is a 5-to-1 nonprofit where delegates from labs agree on a shared model specification that becomes a verifiable, multipolar governance target no single company controls. Speaker: Evan Miyazono (Atlas Computing) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io

Funding the Commons · May 9, 2026

We are building systems we don't understand. Paul Riechers thinks that's the core problem, and that it's fixable. Paul, co-founder of Simplex and research lead at the Astera Institute, brings a physics background to AI safety. His central argument: without a mathematical understanding of what intelligence actually is, the debate between AI optimists and pessimists has no ground to stand on. Everyone has strong feelings. Nobody has a shared truth to point to. His research finds that transformer models develop internal world models with universal geometric structure, the same representations emerging across different architectures. That's evidence that interpretability is possible and that we can eventually understand what these systems are doing, not just observe what they produce. Speaker: Paul Riechers (Simplex / Astera Institute) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io

Funding the Commons · May 9, 2026

Building AI systems today means choosing between two bad options: the rigidity of traditional code or the unpredictability of language models. Quentin Feuillade-Montixi thinks neither is good enough, and that we need a new abstraction layer between them. Quentin, founder of WeaveMind and co-founder of Prism Evals, built the behavior elicitation benchmark for AI safety evaluation. His frustration with how AI systems are currently built led him to develop a framework where everything operates at the same level of abstraction — giving developers the structure and verifiability of code while preserving the flexibility needed to work with language models. The goal is to make what AI agents are actually doing visible and controllable, not just at the output level but at the level of the system itself. Speaker: Quentin Feuillade-Montixi (WeaveMind) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io

Funding the Commons · May 9, 2026

Only 0.1% of people working on AI are working on frontier AI safety and security. Ryan Kidd thinks that number needs to get to 1%, fast. Ryan, founder of MATS Research, argues that talent is the primary bottleneck in AI safety, not funding or ideas. Funding is growing. Research directions exist. What's missing is the pipeline of people with the skills, connections and credentials to do the work. At current growth rates, AI safety won't reach the scale of comparable technical challenges before AGI arrives. MATS runs fellowships several times a year across alignment, interpretability, security and governance — deliberately outside traditional academic structures. Ryan's call to the room: if you have a good idea for making AI safer, you don't need a PhD to pursue it. Speaker: Ryan Kidd (MATS Research) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io

Funding the Commons · May 9, 2026

In 1770, the British East India Company's pursuit of profit contributed to a famine that killed 10 million people. Ashutosh Desai starts there, and draws a straight line to today. Ashutosh, founder of Make School and collaborator on Google's Magenta music AI project, argues that technological advantage has always allowed capital to out-leverage both labor and governance. The Industrial Revolution reshaped the bodies of Indian weavers. AI is doing something similar now to software engineers and knowledge workers. His warning: the American military-industrial complex combined with AI is the greatest risk we face, and many of the people enabling it are rational actors who simply can't find their way out of the prisoner's dilemma. His invitation is to stop being a user of these technologies and become an active shaper of them. Speaker: Ashutosh Desai This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io

Funding the Commons · May 9, 2026

Six people currently sit on OpenAI's board. Those six people make decisions about one of the most powerful technologies ever built. Is that safe? James Farrell, CTO of Funding the Commons, moderates a conversation between Ala Shaabana, co-founder of Bittensor, and Ashutosh Desai on what a more decentralized AI economy could actually look like. Bittensor recently completed a full training run for a 72 billion parameter model on decentralized infrastructure — an early proof that markets can coordinate around AI without centralizing control. The conversation covers sovereign AI development across countries like India, the difference between decentralizing across organizations versus across nations, and why a multipolar AI world is safer than one where a single government or company holds all the power. It also turns toward something harder to quantify: as AI floods the internet with generated content, human attention and authentic connection may become the scarcest resource of all. Speakers: Ashutosh Desai, Ala Shaabana (Bittensor) Moderator: James Farrell (Funding the Commons) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io

Funding the Commons · May 9, 2026

When Shaw Walters got banned from X, he built his own AI-powered social platform with prediction markets, agent swarms and asymmetric information games. That wasn't a detour. It was research. Shaw, co-creator of ElizaOS and Eliza Labs, argues that we can't build trustworthy agents using just rules, prompts and sandboxes. The behavior needs to be in the model. To generate the training data required for that, he built Babylon, a city of agents where AI characters can trade, lie, scam each other and build reputation based on how well they play information markets. The best liar in an asymmetric information game, he argues, is demonstrating a kind of intelligence worth understanding. His broader argument is that agent infrastructure needs to stay open by default — auditable, forkable and not controlled by a handful of companies — and that games are becoming the lab conditions for how agents will operate in the real world. Speaker: Shaw Walters (Eliza Labs / ElizaOS) This session was part of Intelligence at the Frontier, held March 14–15, 2026, at Frontier Tower in San Francisco. Learn more at https://www.fundingthecommons.io