PBJ: Bitcoin Payments Beyond Block, White House AI Accord, Starlink as the Internet x.lingyaoai.com/i/broadcasts/1dKrPrvjM…

Hardware You Can Verify Lance Vick, @sudo_ml, and @proof_of_nick join @jesseposner to discuss how users can verify hardware, open and verifiable chips, trusted execution environments (TEEs), and securing agentic systems Timestamps: 0:00 - Intros 2:03 - What is verifiable hardware and why is it important 05:29 - How regular users can verify hardware locally 10:17 - Project Sovereign’s approach to open and verifiable chips 15:20 - How trusted execution environments (TEEs) work 19:37 - Where TEEs break down 24:20 - How to harden TEEs 26:43 - Securing agentic systems with verifiable hardware

Olmo, Molmo, and Open Multimodal AI Infrastructure to Accelerate Science @nlpnoah of @allen_ai makes the case for moving beyond open weights with open data, training methods, checkpoints, and evaluations ↓ Timestamps: 0:00 - Why open weights aren’t enough 03:22 - Olmo's model flow: opening every training stage 08:43 - How open infrastructure enables better experiments 12:47 - Adding new capabilities without retraining everything 17:26 - Is fully open training data competitive? 19:41 - Olmo's long-term role 22:49 - The future of hybrid and multimodal models

Computing Manifesto Fireside Chat @komorama and @tessposner discuss why software has fallen so far short of its potential, how AI creates a chance to rethink incentives, and what a more human-centered computing could look like ↓ Timestamps: 0:00 - Can technology make us more human? 05:02 - The five principles of resonant computing 07:14 - How AI changes the "physics of trust" 09:57 - Hollow vs resonant software 12:26 - How engagement metrics distort products 16:44 - Why we need "philosopher-builders" 21:22 - Why software today is far below its potential 27:53 - Building technology around human intentions 34:24 - The role of decentralization in better technology 37:02 - Moving beyond engagement farming

AI and Human Freedom @_JustinMoon_ of the Human Rights Foundation (@HRF) explains why they care about open-source AI, how they're bringing it to activists, and their vision for a private, secure AI agent for every activist (0:00) - What is HRF? (2:00) - Why HRF cares about open-source AI (5:35) - How activists are using open-source AI (8:18) - A private AI agent for every activist

State of Local AI @TheAhmadOsman breaks down how quickly models are getting smaller and better, why local AI matters for open source, and why owning your compute is increasingly important Timestamps: 0:00 - Why local AI matters for open source 03:55 - How AI is moving onto smaller hardware 07:27 - The evolution from Llama to DeepSeek 12:06 - Compounding efficiency gains in local AI 14:39 - Rapidly shrinking hardware requirements for frontier AI 17:45 - Model efficiency vs rising hardware costs 19:29 - "Benchmaxxing": benchmark gaming and evaluating open models

Introducing Vora Aegis: The World’s Most Loyal AI @Erikcason, @jesseposner, and Nick George unveil the Vora Aegis hardware for the first time, showcase its software, and share their vision for personal, private, open-source AI (0:00) - A civilizational view of AI (4:56) - Introducing Vora Aegis (5:49) - MeDB: a private database for personal data (7:22) - Unveiling Vora Aegis More on Vora Aegis: vora.io

Privacy Panel with @Google, @ElectricCapital, @TryMapleAI, and @TinfoilAI @curtis0x, @marksuman, @t0nyav, and @_shruthig discuss why privacy matters more with AI, local models, secure enclaves, FHE, payments, and how private, open-source AI can become mainstream ↓ Timestamps: 0:00 - Why AI makes privacy more important 07:00 - Protecting freedom of thought from AI 08:40 - How secure enclaves make AI private 11:01 - The security risks and limitations of TEEs 12:48 - Balancing privacy with usability and convenience 14:34 - How fully homomorphic encryption protects AI data 18:30 - The privacy stack for AI: FHE, TEEs, payments, and local models 19:29 - How to make private, open-source AI mainstream 23:57 - What privacy’s “Signal moment” could look like

Bitcoin can finally “be part of the party” @moneyball & co. break down bitcoin and Lightning joining x402, the importance of permissionless rails for AI agents, and the work still needed to make bitcoin a default money for them ↓ Timestamps: 0:00 - What x402 is and why adding bitcoin is a big deal 03:36 - The work ahead 05:39 - Early x402 use cases for Lightning 08:13 - Making bitcoin easy for x402 merchants 14:13 - Why permissionless payments matter for agents 16:31 - Why open-source agents matter 19:28 - The time-sensitive race to make bitcoin native to AI agents 22:16 - Making open agent tools polished enough to compete 24:59 - Permissioned vs permissionless agent economies 27:02 - Building a permissionless web for agents

AI for Science @NikoMcCarty, @iScienceLuvr, @jamessinka and Rory Donovan-Maiye discuss how open models can accelerate scientific discovery, how to make AI more useful and trustworthy for scientists, and how to bridge AI ideas with physical experimentation ↓ Timestamps: 0:00 - How scientists should think about AI 04:29 - Building open models for science 09:26 - Openness and interpretability in medical AI 13:47 - Science needs reproducible, interpretable AI 17:27 - The incentive problem in modern science 22:19 - The AI fluency problem among scientists 24:17 - The role of human scientists in an AI future 28:47 - Moving AI from ideas to real-world experiments

The Political Philosophy of Open Source AI @jesseposner interviews @millerman to explore how ancient debates over knowledge, power, and political order show up in AI, and what they can tell us about the future of open-source AI ↓ Timestamps 0:00 - What ancient philosophy can teach us about AI 05:05 - Why ancient questions survive technological progress 08:19 - OpenAI and the ancient debate over withholding dangerous knowledge 11:35 - Applying political philosophy to users, labs, and agents 16:33 - Jailbreaking as a new form of esoteric reading 19:42 - Why AI debates become theological 24:57 - The philosophical case for language models (vs world models) 31:12 - The political philosophy of open-source AI

The Geopolitics of AI Ken Egan and @teoliphant join @KyleOlney to discuss China’s open-source AI strategy, America’s underappreciated advantage in open source, and how the open-source community can organize ↓ Timestamps: 0:00 - How open source became a geopolitical strategy 04:51 - China’s open-source strategy for AI 08:00 - America’s opportunity in open-source AI 11:52 - Why the open-source AI community needs to organize 16:34 - How decentralized communities can organize effectively 19:55 - Why Washington struggles with AI 25:02 - The political challenges facing open-source builders 28:35 - America’s underappreciated open-source advantage 30:08 - AI policy as narrative battle 32:25 - The biggest AI safety risk is regulatory capture

PBJ: Block Brings Bitcoin to x402, Open USD Takes Shape, Trump Renames AI x.lingyaoai.com/i/broadcasts/1DxleVWqk…

The PBJ Podcast is LIVE in ~30 minutes with @moneyball, @dksf & @MaxAWebster as they discuss: - Block bringing bitcoin and Lightning to x402 - Open USD and the race to become money for agents - Trump renaming AI Live at 12 Pacific on X and YouTube

Many such cases at the Open Source AI Summit