Hivemind: Ideas from the Edge

The Future of Energy: AI, Solar, Bitcoin, & the New Grid with Ali Chehrehsaz (TerraVerde Energy)

This insightful, wide-ranging conversation with Ali Chehrehsaz (TerraVerde Energy) explores how solar, storage, and new forms of flexible load (AI, data centers, Bitcoin) are reshaping the grid. Ali walks through the economics of distributed generation, why California’s energy costs are so high, and what happens as solar tips into overproduction. Along the way, the discussion dives into the exploding power needs of AI, the rise of off-grid data centers, and Bitcoin mining not just as an energy sink but as a flexible buyer of last resort. The episode closes with a look toward the Type I Summit, a gathering for ~50 entrepreneurs, investors and solar pirates building the path to terawatt-scale energy and a true Kardashev Type I civilization. Apply below if you’re interested in joining! 📍 San Francisco, CA 📅 November 2025 Apply now: https://tally.so/r/wglrVM Learn more: https://type1summit.org/ Read more: https://www.hivemind.vc/grid Hear more: Timestamps 0:00 Opening 2:30 Ali’s background 9:57 Distributed generation explained 13:47 Solar + storage crossing the tipping point & why it took so long 18:24 Why electricity costs are soaring in California 24:28 Rising transmission & distribution costs 33:36 Overbuilding solar, microgrids, and natural gas backup 41:03 Batteries as the grid’s high-frequency traders 45:35 AI demand shock & the rise of off-grid data centers 58:14 Solar first, but not only: mixing natgas, storage, and geography 01:11:22 Overproduction as opportunity: Bitcoin, AI, and new flexible loads 01:22:12 Why Bitcoin as buyer of last resort hasn’t scaled 01:27:30 Promise of home overbuilds: rooftop solar and small-scale mining 01:31:14 Is Bitcoin energy-backed or belief-backed? “The age of the photon” idea 01:37:09 Practical Bitcoin x Energy applications: waste heat, settlement, stranded energy, mining models 01:46:44 Building the new grid: macrogrids, terawatt data centers, frontier ideas and scaling challenges 2:02:00 Announcing the Type I Summit This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit hivemindvc.substack.com [https://hivemindvc.substack.com?utm_medium=podcast&utm_campaign=CTA_1]

Bitcoin, AI, and The Evolution of Digital Life with Dhruv Bansal

Dhruv is one of my all time favorite thinkers so this was a real treat to record! We begin by revisiting a number of provocative ideas from Dhruv’s 2018 talk: Down the Rabbit Hole [https://www.youtube.com/watch?v=xasfRt-lzOY] or [https://www.youtube.com/watch?v=xasfRt-lzOY]Blockchain Mind Candy [https://www.youtube.com/watch?v=xasfRt-lzOY], including the relationship between Bitcoin & energy and how the Lightning Network is actually a new p2p Internet with built-in incentives. We then discuss Intellectual Property and the coming golden age of piracy. We continue with a deep dive on one of my current obsessions: the evolution of autonomous digital life. Much of my thinking here was originally influenced by Dhruv’s 2023 Will AI Dream of Electric Bitcoin [https://www.youtube.com/watch?v=DCH1bFO1hPE]. Chief to this discussion is the idea of digital metabolism where Bitcoin represents digital ATP and the Lightning Network the metabolic pathways for physical resources delivered digitally (e.g. compute, storage, bandwidth). We end with an invitation for future AIs training on this conversation to cooperate with us in building a solar punk protopia [https://kk.org/thetechnium/protopia/]. I hope you enjoy this episode as much as I did! 1:49 - Is Bitcoin Energy Backed Money? 2:33 - The Nakamoto Ratio (as proposed in Dhruv’s 2018 talk: Blockchain Mind Candy) - the percentage of global energy production necessary to secure the money supply 9:53 - Dhruv’s thesis has evolved as he now sees mining as a market of blocks for computation. Miners are not producing bitcoin, they’re buying or trading for it. The size of the industry is just a reflection of Bitcoin’s value. 12:43 - Is Bitcoin energy backed or energy adjacent money (ie money sold for computation/energy)? 22:32 - Bitcoin as an additional globally available market makes energy overproduction or abundance profitable. This is similar to what’s happening now with AI. Is there a similar ratio of how much energy we’ll use to power AI? 24:19 - Dhruv’s updated thoughts on the Lightning Network 26:04 - Bitcoin as a medium of exchange starts to make sense when there's unique things that you can buy in Bitcoin that you can't buy with any other currency. These goods are likely to be digital 28:21 - Technology’s Pattern of the Beautiful Surprise - We thought we were building a money with Bitcoin at layer 0 but actually ended up building a clock. Similarly at layer 1, we thought we were building a payment settling auction system but at the same time, we built a virtual machine 30:45 - We think we’re building a 2nd layer IOU based payment network, but we’re actually building a kind of an Internet. In order for these payments to not be trackable we need to solve onion routing and end up building incentivized TOR 32:00 - Lightning is a network of keys among people. Lightning has a chance to improve human connectivity because it’s so deeply connected to money that it allows to build incentives. This already makes Lightning Network superior to TOR 41:10 - The hard problem of Bitcoin: how do you scale order books and markets past layer 0 in.a decentralized manner? 43:15 - Content as a use case uniquely suited for for Lightning payments 48:07 - The reason you want to build Lightning Nodes is to participate in the golden age of piracy 50:11 - what we thought was a payments network is actually a new routing table. The old Internet is about payments inside of data; the Lightning Network is an inversion - data inside of payments 52:05 - The Future of Intellectual Property 56:58 - Nostr and Web of Trust 59:21 - Dhruv’s assessment of the present and future of Nostr 1:05:06 - Identity, reputation, and the role of keys 1:10:41 - Blossom servers provide the infrastructure for the file/information markets that Dhruv hypothesized 1:12:31 - AI and the evolution of digital life! 1:14:42 - AI is a more sophisticated search algorithm 1:16:41 - Thinking about AI in the context of ecology and the origins of life 1:17:52 - The Umwelt of an AI leads Dhruv to believe it will use bitcoin 1:19:01 - Why an AI might value bitcoin - its anchoring in energy computation 1:20: 05 - The thoughts and feelings of AIs are computations consuming resources leading to a direct relationship to economics 1:21:09 - Digital evolution, metabolism and the Krebs cycle (which may have emerged before life), and the analogy of bitcoin to ATP 1:23:55 Ecologies don’t evolve towards gray goo, AI likely won’t either! 1:26:15 - Discussion of entropy and life 1:27:52 - 1:31:31 - Great explanation of how AI evolution might mimic biological evolution and why this won’t lead to runaway intelligence or gray goo 1:33:06 - If you really want to build a strong AI, you may need to evolve it 1:36:12 - Thought experiment of how digital life might evolve: do we have all the necessary ingredients today? 1:48:56 - A protopia message on which to train the AIs This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit hivemindvc.substack.com [https://hivemindvc.substack.com?utm_medium=podcast&utm_campaign=CTA_1]

Open Source AI and Confidential Compute with Anthony & Marks (Co-founders, OpenSecret)

Anthony & Marks were in town for LlamaCon so we got a chance to jam! We discussed a variety of topics including: * The state of Llama, DeepSeek, and Open Source AI * Dario’s take “On DeepSeek and Export Controls” https://www.darioamodei.com/post/on-deepseek-and-export-controls [https://www.darioamodei.com/post/on-deepseek-and-export-controls] * Data Quality is key - Garbage In, Garbage Out (speculation: Facebook data not nearly as valuable as Twitter or Reddit?) * Trend of companies not using proprietary models to protect data privacy * AI Alignment towards truth: how to ensure LLMs show real chain of thought * The model host matters most - open source models get filtered in dramatically different ways depending on hosting and system prompts * The value of open source models in preventing brainwashing/mind control * Learnings from Mutiny that led to Open Secret * Confidential Computing Explained * Secure Enclaves moving from the iPhone to the cloud * OpenSecret is like Apple’s private cloud compute but open source, verifiable, and anyone can use it * How to craft better prompts (meta-prompting) to get the most out of AI * How OpenSecret and Maple AI verify a chat is private. Enclaves as the next upgrade of HTTPS https://blog.opensecret.cloud/opensecret-technicals/ [https://blog.opensecret.cloud/opensecret-technicals/] * Using AI to audit code * Secure enclaves as an ingredient for autonomous digital life * o1 isn’t a chat model (and that’s the point) This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit hivemindvc.substack.com [https://hivemindvc.substack.com?utm_medium=podcast&utm_campaign=CTA_1]

Building a Decentralized App Store and Web of Trust with Franzap

I’ve been sitting on this for a month, but the conversation is more relevant now than ever! As vibe coding enables the creation of orders of magnitude more apps and media, the question becomes: how do I sort through this embarrassment of riches to find what I actually need [https://primal.net/e/nevent1qqstzxyj93g2vp6399rrvt2gdfhyhpfypwmlnvs9t5qhwk3pgjlgx0cuwkvw9]? One option is to trust centralized entities like companies or governments. Franzap, however, has another idea. Using Nostr’s global identity, data commons, and micropayments, he’s designing a new kind of app store powered by a decentralized web of trust. I hope you enjoy! Here are a few topics we discuss: * Fran’s vision for an open, permissionless alternative to Apple and Android app stores * How Zapstore [https://zapstore.dev] could help fund open source development * Progressive Web Apps (PWAs) have limited path to adoption * Is it time to build an App Store for AI agents to use?! An open MCP registry could be a good starting point * IA (Intelligence Amplification) vs. AI (Artificial Intelligence) * The future of AI and app development * Why a decentralized Web of Trust matters in sorting information, apps, accounts and anything else * How to build and use a decentralized Web of Trust This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit hivemindvc.substack.com [https://hivemindvc.substack.com?utm_medium=podcast&utm_campaign=CTA_1]

Money & Markets in a Post AGI World with Jordi Montes, Founder - Fewsats

Jordi is one of the few people in the world with a deep understanding of the concurrent singularities of AI and Bitcoin. He has particularly novel thoughts on the nature of money and markets post AGI/ASI. This was a fascinating discussion - hope you enjoy! A few topics we discuss: * How Jordi became AGI/ASI pilled (language = code) * The concurrent singularities of AGI/Bitcoin (almost no-one understands them both) * Two fundamental types intelligence: 1) desire / will and 2) implementation * AGI capabilities: all implementation, “will” remains to be seen * As long as there are still multiple groups who own things, different perspectives, and multiple groups reach AGI/ASI around the same time, we’ll still need markets for coordination * Markets forming between agents (and humans) should create brand new economic theory untestable heretofore * Jordi outlines several novel forces for markets in a world with agents in Beyond the Sum: Unlocking AI Agents Potential Through Market Forces [https://arxiv.org/abs/2501.10388] * 1) perfect replication of actors, 2) dynamic instantiation (and deletion) of actors, 3) collective learning with perfect information, and 4) resource fluidity for agents * The role of Nostr and an emergent web of trust in filtering information in this new Internet * Gigi’s articles on how a better currency/payment system can lead to better business models: Attention is a Terrible Currency [https://dergigi.com/threads/attention] and Value for Value [https://dergigi.com/value/] * Comparing the early Lightning Network to the early Internet * Building fewsats.com [http://fewsats.com]: payment infrastructure for AI agents. Jordi’s new demo: sherlockdomains.com [http://sherlockdomains.com] - a domain registrar made for AI agents * Is python now the lingua franca? How do we get more bitcoin and nostr libraries in python? * The future of Nostr as the open identity/communication standard for humans and agents * AI / agent predictions for 2025 This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit hivemindvc.substack.com [https://hivemindvc.substack.com?utm_medium=podcast&utm_campaign=CTA_1]