<p>Frontier AI is colliding with real-world infrastructure. Eiso Kant (Co-CEO &amp; Co-Founder, Poolside) joins the MAD Podcast to unpack Project Horizon— a multi-gigawatt West Texas build—and why frontier labs must own energy, compute, and intelligence to compete. We map token economics, cloud-style margins, and the staged 250 MW rollout using 2.5 MW modular skids.</p><p><br></p><p>Then we get operational: the CoreWeave anchor partnership, environmental choices (SCR, renewables + gas + batteries), community impact, and how Poolside plans to bring capacity online quickly without renting away margin—plus the enterprise motion (defense to Fortune 500) powered by forward deployed research engineers.</p><p><br></p><p>Finally, we go deep on training. Eiso lays out RL2L (Reinforcement Learning to Learn)— aimed at reverse-engineering the web’s thoughts and actions— why intelligence may commoditize, what that means for agents, and how coding served as a proxy for long-horizon reasoning before expanding to broader knowledge work.</p><p><br></p><p>Poolside</p><p>Website - https://poolside.ai</p><p>X/Twitter - https://x.com/poolsideai</p><p><br></p><p>Eiso Kant</p><p>LinkedIn - https://www.linkedin.com/in/eisokant/</p><p>X/Twitter - https://x.com/eisokant</p><p><br></p><p>FIRSTMARK</p><p>Website - https://firstmark.com</p><p>X/Twitter - https://twitter.com/FirstMarkCap</p><p><br></p><p>Matt Turck (Managing Director)</p><p>Blog - https://www.mattturck.com</p><p>LinkedIn - https://www.linkedin.com/in/turck/</p><p>X/Twitter - https://twitter.com/mattturck</p><p><br></p><p>(00:00) Cold open – “Intelligence becomes a commodity”</p><p>(00:23) Host intro – Project Horizon &amp; RL2L</p><p>(01:19) Why Poolside exists amid frontier labs</p><p>(04:38) Project Horizon: building one of the largest US data center campuses</p><p>(07:20) Why own infra: scale, cost, and avoiding “cosplay”</p><p>(10:06) Economics deep dive: $8B for 250 MW, capex/opex, margins</p><p>(16:47) CoreWeave partnership: anchor tenant + flexible scaling</p><p>(18:24) Hiring the right tail: building a physical infra org</p><p>(30:31) RL today → agentic RL and long-horizon tasks</p><p>(37:23) RL2L revealed: reverse-engineering the web’s thoughts &amp; actions</p><p>(39:32) Continuous learning and the “hot stove” limitation</p><p>(43:30) Agents debate: thin wrappers, differentiation, and model collapse</p><p>(49:10) “Is AI plateauing?”—chip cycles, scale limits, and new axes</p><p>(53:49) Why software was the proxy; expanding to enterprise knowledge work</p><p>(55:17) Model status: Malibu → Laguna (small/medium/large)</p><p>(57:31) Poolside&#39;s Commercial Reality today: defense; Fortune 500; FDRE </p><p>(1:02:43) Global team, avoiding the echo chamber</p><p>(1:04:34) Next 12–18 months: frontier models + infra scale</p><p>(1:05:52) Closing</p>