Hi Everyone! We've compiled a list of upcoming events and exclusive resources for the @FounderCoHo community. Feel free to check it out and participate! go.foundercoho.com/weekly-pr…

Today, we brought AI researchers and builders together at Stanford for AI Researchers, October 2026 co-hosted by @daytonaio and @FounderCoHo From long-horizon, stateful agents to world models, reinforcement learning environments, and the next generation of AI evaluations, the evening explored what it takes to train, scale, and evaluate increasingly capable AI systems. Huge thanks to our speakers, Dr. Zhenyu Liao (@amazon ), Dr. Xide Xia (@Steadyworks, formerly Meta), Xiangyi Li (@xdotli) (@benchflow_ai ) and Muhammad Annas Hashmi (@muhashmii)(@daytonaio), for sharing their research, insights, and perspectives on where agent RL and AI research are headed. And thanks to everyone who joined us at Stanford for the thoughtful conversations, knowledge sharing, and community building. We're excited to keep bringing researchers, founders, and engineers together to explore the ideas and infrastructure shaping the next wave of AI. #FounderCoHo #Daytona #AIResearch #ReinforcementLearning #AIAgents #Stanford

𝗪𝗵𝗮𝘁 𝗱𝗼𝗲𝘀 𝗶𝘁 𝘁𝗮𝗸𝗲 𝘁𝗼 𝗺𝗼𝘃𝗲 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀 𝗳𝗿𝗼𝗺 𝗱𝗲𝗺𝗼 𝘁𝗼 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻? At FounderCoHo’s 𝗢𝗰𝘁. 𝟵 𝗦𝗙 𝗧𝗲𝗰𝗵 𝗪𝗲𝗲𝗸 panel, builders from 𝗚𝗼𝗼𝗴𝗹𝗲, 𝗢𝗽𝘂𝘀𝗖𝗹𝗶𝗽, 𝗮𝗻𝗱 𝗛𝗮𝗿𝘃𝗲𝘆 shared practical lessons: 𝗔𝗻𝗻𝗶𝗲 𝗪𝗮𝗻𝗴 @anniewangtech : Combine AI reasoning with deterministic logic. Choose the right approach for each step—including evaluation. 𝗝𝗮𝘆 𝗪𝘂 : Define your target customer. @OpusClip's pivot showed the value of solving a clear customer need; creative tools also need users with a story and intent. 𝗦𝗮𝗺𝗲𝗿 𝗠𝗮𝘀𝘁𝗲𝗿𝘀𝗼𝗻 @samertm : Make answers verifiable. Citation matching helps Harvey’s legal users check outputs against source documents. 𝗣𝗵𝗶𝗹𝗶𝗽 𝗟𝗮𝗻: Start with the user’s workflow. Build evaluations around real failure cases, then account for production costs and system constraints. 𝗧𝗮𝗸𝗲𝗮𝘄𝗮𝘆: 𝘜𝘴𝘦𝘧𝘶𝘭 𝘢𝘨𝘦𝘯𝘵𝘴 𝘯𝘦𝘦𝘥 𝘤𝘭𝘦𝘢𝘳 𝘪𝘯𝘵𝘦𝘯𝘵, 𝘮𝘦𝘢𝘴𝘶𝘳𝘢𝘣𝘭𝘦 𝘲𝘶𝘢𝘭𝘪𝘵𝘺, 𝘢𝘯𝘥 𝘳𝘦𝘭𝘪𝘢𝘣𝘭𝘦 𝘦𝘹𝘦𝘤𝘶𝘵𝘪𝘰𝘯. Thanks to our speakers, moderator @jingconan Jing Conan Wang, sponsor @Deepvista, and Byron and the Olive team for hosting! Great to bring together @RuhanDong, @batmany13, @asharoraa, @Danil_KV, @carolinarizo, @freejenko, and everyone who joined. #FounderCoHo #AIAgents #SFTechWeek.

Hi Everyone! We've compiled a list of upcoming events and exclusive resources for the @FounderCoHo community. Feel free to check it out and participate! go.foundercoho.com/weekly-pr…

𝗧𝗵𝗲 𝗻𝗲𝘅𝘁 𝘄𝗮𝘃𝗲 𝗼𝗳 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀 𝘄𝗼𝗻’𝘁 𝗷𝘂𝘀𝘁 𝗻𝗲𝗲𝗱 𝗯𝗲𝘁𝘁𝗲𝗿 𝗺𝗼𝗱𝗲𝗹𝘀. 𝗧𝗵𝗲𝘆’𝗹𝗹 𝗻𝗲𝗲𝗱 𝗯𝗲𝘁𝘁𝗲𝗿 𝗲𝗻𝘃𝗶𝗿𝗼𝗻𝗺𝗲𝗻𝘁𝘀, 𝗯𝗲𝘁𝘁𝗲𝗿 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲, 𝗮𝗻𝗱 𝗯𝗲𝘁𝘁𝗲𝗿 𝘄𝗮𝘆𝘀 𝘁𝗼 𝗲𝘃𝗮𝗹𝘂𝗮𝘁𝗲 𝘄𝗵𝗮𝘁 𝘁𝗵𝗲𝘆 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝗱𝗼. On October 13, @FounderCoHo and @daytonaio are bringing together researchers at Stanford to explore the next frontier of Agent RL. Speakers: • Dr. Zhenyu Liao — Amazon • Dr. Xide Xia — Steadyworks, Ex-Meta • Xiangyi Li (@xdotli) — BenchFlow, SkillsBench • Muhammad Annas Hashmi (@muhashmii) — Daytona We’ll Explore : • Stateful, long-horizon agents • World Models for faster RL training • RL environments for interactive learning • The future of agent evaluation • Infrastructure for sessionful, multi-day rollouts 𝗧𝗵𝗲 𝗡𝗲𝘅𝘁 𝗪𝗮𝘃𝗲 𝗼𝗳 𝗔𝗴𝗲𝗻𝘁 𝗥𝗟: Infrastructure, World Models, and the Future of Evals 📍 𝗦𝘁𝗮𝗻𝗳𝗼𝗿𝗱 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗶𝘁𝘆 📅 𝗢𝗰𝘁𝗼𝗯𝗲𝗿 𝟭𝟯, 𝟮𝟬𝟮𝟲 ⏰ 𝟱:𝟯𝟬–𝟴:𝟯𝟬 𝗣𝗠 𝗣𝗦𝗧 𝗥𝗲𝗾𝘂𝗲𝘀𝘁 𝘁𝗼 𝗷𝗼𝗶𝗻: luma.com/ai-researchers-oct

The demo always works. Production breaks. Join us Fri Oct 9 for #SFTechWeek: what it really takes to run AI agents in production. Speakers from Google, OpusClip & Harvey. Palo Alto, 5–8 PM. 🍸 RSVP 👉 partiful.com/e/tqX7BGeGKHpIg…

Eric Ries's (@ericries) One Regret About The Lean Startup: It Never Said "For the Better." What if "best practices" are why great companies get destroyed? Eric Ries, author of The Lean Startup and Incorruptible, joins us for a conversation on building companies that last — and what he wishes he had said differently. Takeaways: • “Change the World” Needed Three More Words • Companies “Murdered in the Name of Profit” • 5x More Likely to Reach Year 50 • Every Frontier AI Lab Has a Mission Guardian • The Todd Park Rule • Token Maxxing Is the Ultimate Vanity Metric Full interview: youtube.com/watch?v=2NtKUPkg…

Hi Everyone! We've compiled a list of upcoming events and exclusive resources for the @FounderCoHo community. Feel free to check it out and participate! go.foundercoho.com/weekly-pr…

𝗪𝗵𝘆 𝘄𝗼𝗻'𝘁 𝗟𝗟𝗠𝘀 𝗿𝗲𝗮𝗰𝗵 𝗔𝗚𝗜? 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗺𝗶𝗴𝗵𝘁 𝗯𝗲 𝗹𝗲𝘀𝘀 𝘁𝗵𝗮𝗻 𝟭% 𝗼𝗳 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲. Professor James Zou (@james_y_zou) head Stanford's AI for Science Lab (Nature papers, FDA-cleared health algorithms, NYT recognition). 𝗛𝗶𝘀 𝗰𝗹𝗮𝗶𝗺: text captures a fraction of what happens in the brain and body. Real superintelligence needs training on physiological signals, not just words. 𝗪𝗵𝘆 𝗶𝘁 𝗳𝗮𝗶𝗹𝘀: Every word you say follows tens of thousands of biological signals that already fired. LLMs only see what survives translation into text. What his lab is building instead: • 𝗩𝗶𝗿𝘁𝘂𝗮𝗹 𝗟𝗮𝗯: clones his own team into autonomous AI agents (Nature) • 𝗣𝗮𝗽𝗲𝗿𝟮𝗔𝗴𝗲𝗻𝘁: turns static papers into interactive "virtual authors" • 𝗦𝗹𝗲𝗲𝗽𝗙𝗠: one night of sleep data predicts risk for ~130 diseases As Prof. Zou puts it: "𝘛𝘩𝘦 𝘸𝘳𝘪𝘵𝘵𝘦𝘯 𝘭𝘢𝘯𝘨𝘶𝘢𝘨𝘦 𝘪𝘴 𝘢 𝘷𝘦𝘳𝘺 𝘴𝘮𝘢𝘭𝘭 𝘧𝘳𝘢𝘤𝘵𝘪𝘰𝘯 𝘰𝘧 𝘵𝘩𝘦 𝘵𝘩𝘰𝘶𝘨𝘩𝘵𝘴 𝘢𝘯𝘥 𝘵𝘩𝘪𝘯𝘨𝘴 𝘵𝘩𝘢𝘵 𝘩𝘢𝘱𝘱𝘦𝘯... 𝘵𝘩𝘦 𝘤𝘶𝘳𝘳𝘦𝘯𝘵 𝘱𝘢𝘳𝘢𝘥𝘪𝘨𝘮 𝘰𝘧 𝘵𝘳𝘢𝘪𝘯𝘪𝘯𝘨 𝘓𝘓𝘔𝘴 𝘪𝘴 𝘢𝘭𝘭 𝘵𝘳𝘢𝘪𝘯𝘪𝘯𝘨 𝘰𝘯 𝘵𝘩𝘪𝘴 𝘥𝘰𝘸𝘯𝘴𝘵𝘳𝘦𝘢𝘮 𝘢𝘳𝘵𝘪𝘧𝘢𝘤𝘵, 𝘸𝘩𝘪𝘤𝘩 𝘪𝘴 𝘭𝘦𝘴𝘴 𝘵𝘩𝘢𝘯 1% 𝘰𝘧 𝘵𝘩𝘦 𝘢𝘤𝘵𝘶𝘢𝘭 𝘵𝘩𝘰𝘶𝘨𝘩𝘵𝘴." He expects AI to produce Nobel-caliber discoveries within a decade. Full interview: youtu.be/awrU6VSiPDw

🔴 𝗪𝗲’𝗿𝗲 𝗟𝗜𝗩𝗘! 𝗔𝗜-𝗡𝗮𝘁𝗶𝘃𝗲 𝗚𝗿𝗼𝘄𝘁𝗵 & 𝗖𝘂𝘀𝘁𝗼𝗺𝗲𝗿 𝗢𝗽𝗲𝗿𝗮𝘁𝗶𝗼𝗻𝘀: 𝗙𝗿𝗼𝗺 𝗙𝗶𝗿𝘀𝘁 𝗧𝗼𝘂𝗰𝗵 𝘁𝗼 𝗥𝗲𝘁𝗲𝗻𝘁𝗶𝗼𝗻 Join us as we break down how AI-native systems can connect acquisition, onboarding, and retention. 🎙️ 𝗝𝗶𝗻𝗴 𝗖𝗼𝗻𝗮𝗻 𝗪𝗮𝗻𝗴 — Founder & CEO, DeepVista AI 🎙️ 𝗔𝗻𝗻𝗶𝗲 𝗪𝗮𝗻𝗴 — Software Engineer, Google Cloud GCP 🎥 Watch live: go.foundercoho.com/livestrea…

We've compiled a list of upcoming events and exclusive resources for the @FounderCoHo community. Feel free to check it out and participate! go.foundercoho.com/weekly-pr…

𝗣𝗿𝗼𝗺𝗽𝘁 𝗲𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 𝗶𝘀 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗱𝗲𝗮𝗱. 𝗜𝗳 𝘆𝗼𝘂'𝗿𝗲 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗿𝗲𝗮𝗹 𝘀𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝘄𝗶𝘁𝗵 𝗔𝗜, 𝘆𝗼𝘂 𝗻𝗲𝗲𝗱 𝗦𝗽𝗲𝗰𝘀. Lazar Jovanovic (@lakikentaki) (@Lovable's first professional vibe coding engineer, 500+ apps shipped) says it plainly: 𝘗𝘳𝘰𝘮𝘱𝘵𝘴 𝘢𝘳𝘦 𝘧𝘰𝘳 𝘱𝘳𝘰𝘵𝘰𝘵𝘺𝘱𝘦𝘴. 𝘗𝘳𝘰𝘥𝘶𝘤𝘵𝘪𝘰𝘯 𝘳𝘦𝘲𝘶𝘪𝘳𝘦𝘴 𝘸𝘳𝘪𝘵𝘪𝘯𝘨 𝘚𝘱𝘦𝘤𝘴. 𝗪𝗵𝘆 𝘆𝗼𝘂𝗿 𝗮𝗴𝗲𝗻𝘁 𝗱𝗿𝗶𝗳𝘁𝘀: A prompt works for a single UI tweak. But across a multi-step sequence, chat context degrades. The agent loses the thread, starts guessing, and hallucinates dependencies that break existing code. 𝗛𝗼𝘄 𝘁𝗼 𝗽𝗿𝗲𝘃𝗲𝗻𝘁 𝗮𝗴𝗲𝗻𝘁 𝗱𝗿𝗶𝗳𝘁: Stop prompting your way out of a polluted context: • Move requirements out of chat into structured .md Spec files • Define data models, Supabase tables, and boundaries upfront • Never prompt "fix this" on repeat; use Plan mode to trace the root cause • Ruthlessly delete dead code and reset context when complexity creeps in As Lazar puts it: "I'm doing more engineering than vibing these days." 𝗙𝘂𝗹𝗹 𝗰𝗼𝗻𝘃𝗲𝗿𝘀𝗮𝘁𝗶𝗼𝗻: youtu.be/9rzLqBPI4Y8

Hi👋! Looking to connect with people who are actually building, shipping, and marketing cool stuff. 🚀 If you’re a: • Founder • Builder • Indie hacker • Developer • Vibe coder I’ll check it out and give you a follow. 👀 Let’s connect with more people who are building. 👇

Upcoming Event : 𝗔𝗜-𝗡𝗮𝘁𝗶𝘃𝗲 𝗚𝗿𝗼𝘄𝘁𝗵 & 𝗖𝘂𝘀𝘁𝗼𝗺𝗲𝗿 𝗢𝗽𝗲𝗿𝗮𝘁𝗶𝗼𝗻𝘀: 𝗙𝗿𝗼𝗺 𝗙𝗶𝗿𝘀𝘁 𝗧𝗼𝘂𝗰𝗵 𝘁𝗼 𝗥𝗲𝘁𝗲𝗻𝘁𝗶𝗼𝗻 👇 𝙈𝙤𝙨𝙩 𝙂𝙏𝙈 𝘼𝙄 𝙛𝙖𝙞𝙡𝙨 𝙛𝙤𝙧 𝙤𝙣𝙚 𝙨𝙞𝙢𝙥𝙡𝙚 𝙧𝙚𝙖𝙨𝙤𝙣: 𝙞𝙩’𝙨 𝙨𝙩𝙖𝙩𝙚𝙡𝙚𝙨𝙨. Context drops between sales, onboarding, and CS. Handoffs break, promises get lost, and retention turns into constant firefighting. At our next @FounderCoHo livestream, we’re breaking down how to design an AI-native operating layer that preserves persistent customer context from day one. What we're covering: • Fixing broken sales-to-onboarding handoffs • Architecting persistent context across the full lifecycle • Automating workflows without manual admin drag 🎙️ Speaker: Jing Conan Wang @jingconan — Founder & CEO @DeepVistaAI , Ex-Google Brain Research Lead 🎙️ Moderated by: Annie Wang @anniewangtech — Software Engineer @googlecloud GCP 🗓️ Wednesday, Sep 9 | 6:00 – 7:00 PM PST 📍 Youtube Livestream 🔗 RSVP: luma.com/livestream-9-9

We've compiled a list of upcoming events and exclusive resources for the @FounderCoHo community. Feel free to check it out and participate! go.foundercoho.com/weekly-pr…

𝗪𝗵𝗮𝘁 𝗺𝗶𝗴𝗵𝘁 𝗔𝗜 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗸𝗻𝗼𝘄 𝗮𝗯𝗼𝘂𝘁 𝗯𝗶𝗼𝗹𝗼𝗴𝘆 𝘁𝗵𝗮𝘁 𝘄𝗲 𝗱𝗼𝗻'𝘁? Stanford professor Alice Ting explores a fascinating frontier: AI models may have learned patterns about the chemical forces driving proteins that science has yet to uncover. For AI researchers and biologists alike, the opportunity may lie in understanding what these models actually know. Because even at the frontier of AI, innovation still starts with the fundamentals: 𝗰𝗵𝗲𝗺𝗶𝘀𝘁𝗿𝘆 𝗮𝗻𝗱 𝗯𝗶𝗼𝗽𝗵𝘆𝘀𝗶𝗰𝘀. #AI #Biology #AlphaFold #Biotech #ProteinScience