AGI Summit

The people closest to the WORK are often the ones closest to the ANSWER! I was asked during the AGI Summit about why some technology deployments succeed while others never make it beyond the pilot stage. My answer was specific to robotics and physical AI, but I think the lesson applies far beyond AI. Too often, organizations make decisions from the top down. Leadership chooses a new capability and then asks employees to embrace it. I’ve found that the better approach is often the opposite. Start with the people closest to the work. • Ask what slows them down. • Ask what they’d automate. • Ask what would genuinely make their day easier. When the people doing the work see value, adoption becomes much more natural. When they don’t, even the best technology can struggle to gain momentum. Whether you’re deploying AI, redesigning a business process, or leading organizational change, the principle is remarkably consistent: The people closest to the problem often have the clearest view of the solution. Thank you to Lisheng Wang and my fellow panelists, C. Tsai; Saman Farid, Kevin Chen, PhD, for an insightful and intellectually fun discussion at the AGI Summit. It was a pleasure exchanging ideas with leaders who are helping shape the future of AI from so many different perspectives.

Writing software has become asynchronous. You are the main thread. Agents are parallel workers. This weekend at AGI Summit, I shared lessons from building Ads and scaling it to a $100M+ ARR product in six weeks at OpenAI. Traditionally, software engineering has been sequential: take a task, implement it, debug it, then move to the next one. With agents, you can spin off multiple streams of work in parallel and await the results while continuing to move. The new measure of leverage isn’t how quickly you write code. It’s how much high-quality work can progress without requiring your constant attention. A huge thank-you to Jijun Cai Melly Liu Sophie R. for bringing such a thoughtful community together, and to all the volunteers whose work made the weekend possible. #AGISummit #AIEngineering #Codex

Honored to serve as an organizer and the main AGI Stage MC at AGI Summit 2026, and to open the afternoon program with Anthony Bonvino, VP of Investor Relations at AC Future. Anthony kicked off the afternoon with an industry spotlight on “Reimagining Mobile Living with AI,” sharing how AI-enabled mobility can transform vehicles into smarter spaces for living, working, and everyday life. It was a pleasure introducing his session and hearing AC Future’s vision firsthand. Grateful to Anthony for joining us, and excited to stay connected as he returns to California. Looking forward to future panels, founder events, and hackathons together through BayForge. #AGISummit2026 #ACFuture #ArtificialIntelligence #FutureOfMobility #AICommunity #BayForge

Technology changes quickly. Human ambition does not. The past few days gave me two very different views of innovation. At B&H Creators in San Francisco, the Hiller Aviation Museum brought together creators, technology, and more than a century of aviation history. Seeing early aircraft and experimental machines was a strong reminder that many ideas once considered unrealistic became industries because someone kept building after others stopped believing. I then spent the weekend at AGI Summit 2026 at the Palace of Fine Arts in San Francisco. The summit brought together founders, investors, researchers, global technology companies, AI startups, and robotics teams from around the world. My main takeaway: Breakthrough technology is rarely limited by imagination. It is limited by execution, distribution, and the ability to remain focused long enough for the market to catch up. A century ago, aviation pioneers were building machines without knowing how large the industry would become. Today, AI and robotics founders are working through a similar transition. For founders, the opportunity is not simply to add AI to an existing product. It is to identify which workflows, industries, and human capabilities will be fundamentally rebuilt. For investors, the strongest signal is not the loudest demo. It is a team with deep market understanding, clear distribution, and the discipline to execute after the excitement moves elsewhere. Glad to reconnect with friends and meet many impressive people for the first time: Jenny Li, Benjamin Shlemis, Yulia Stepanova, Kevin Mei, Chjango U., TatianaSFcom 🌉🟧, Michael Noffsinger, Robert Lent, Viktoriia Otkydach, Frazer Galway, Will Sumner ▶️, Jason Reyes, Danny Kim, Oscarly Sanchez, Yetty Williams, MBA., Shailendra Lande, Christopher Arnold, Nick Zhong, Hailing Ge, Justin Zelaya, Irka Pawlowski, Alisha Cheng, Minchen Luo, Joel H., Nathaniel Powell, Lingfei (Teddy) Wu, Michael Li, Neeke Gao, Nazar Zinenko, Yuqi Liu, Veronika Loginova, Ali K., Vladimir Hohlov, Hemalatha Jinnagara, PMP, Илхомбек Holiddinov, Kseniya Aleksandrova, Deepti Bahel, June Zhang Thanks to the photographer (Kevin Mei Sanscritic.ai) for the photos. Wishing everyone continued growth, disciplined execution, and the opportunity to build a company capable of becoming a unicorn—and valuable enough to make the world better.

What does aviation have in common with AI development? A lot more than you might think.... That's what Xiaodi Hou talked about in his presentation at the AGI Summit this weekend in San Francisco. Aviation went from an industry that seemed fantastical to one that quite literally transformed the way humans access the world. Physical AI will do the same thing. Keep an eye on our channels for his whole talk, plus a deeper dive into the lessons aviation can teach AI companies about scaling successfully in new technology frontiers.

What should AI startups do when code is free and your competitor can copy you in a few days? SF is booming, even on weekends! The well-attended AGI Summit hosted thousands at the Palace of Fine Arts this past Saturday and Sunday, featuring packed sessions, workshops, demos, and pitch competitions. I spoke on how AI Startups can win, when code is free/disposable. The secret is startups need to develop business strategies into a flywheel, I outlined: 1) Propriatary data 2) Community/Ecosystem 3) Product led growth 4) Distribution 5) Network effects These align to the Blitzscaling book written by Chris Yeh and Reid Hoffman and how I'm investing at Blitzscaling Ventures. There is no other place in the world where several AI events, hackathons, and conferences are all happening on a Saturday weekend, the world continues to gravitate to SF for the thriving AI ecosystem.