Episodes (Page 2)
✨
Yi Tay shipped Gemini Deep Think and IMO Gold, launched Reasoning and AGI team in Singapore expanding from 12 to 300+ researchers pivoting from architecture research to RL-driven reasoning
Yi Tay
✨
James Reggio led Brex's three-pillar AI strategy (corporate, operational, product AI) with SOP-driven agents outperforming overengineered RL in operations at scale
CTO James Reggio
✨
Artificial Analysis evolved from 2024 Sydney basement side project to independent gold standard for AI benchmarking trusted by developers, enterprises, and major labs
✨
LMArena raised $150M Series A at $1.7B valuation with $30M annualized consumption revenue ($2.5M MRR) after September 2025 evals product launch expansion
Anastasios Angelopoulos
✨
Kevin Wang et al. won NeurIPS 2025 Best Paper by scaling RL networks to 1,000 layers deep, defying decade-long conventional wisdom that depth fails in RL
✨
John Yang recaps SWE-bench's evolution from ignored (Oct 2023) to industry standard after Devin's launch, expanding from Django-heavy to 9 languages across 40 repos
John Yang
✨
Josh McGrath at OpenAI describes post-training evolution from 2023 PPO vs DPO debates to current RLVR era where data quality and signal trust matter more than optimization method
Josh McGrath
✨
Ashvin Nair from Cursor shipped RL breakthroughs on GPT-4o/o1/o3; reasoning team scaled from 12 to 300+ people; IOI Gold felt reachable in 2022 but only materialized when o1 shipped
Ashvin Nair
✨
Sarah Catanzaro transitions from modern data stack investing (DBT, Fivetran) to AI infrastructure frontier, warning that $100M+ seed rounds with no near-term roadmap are now standard
Sarah Catanzaro
✨
MCP evolved from local-only experiment (Thanksgiving launch) to de facto protocol adopted by OpenAI, Microsoft, Google, Block with hundreds of enterprise agents at scale
✨
Steve Yegge argues Claude Code, Cursor, and 2024 stack are obsolete, advocating for 'factory farming' era where developers manage fleets of AI agents instead of writing code
✨
Brian Fioca and Bill Chen from OpenAI's Codex and GPT-5 teams explain training agents to work 24+ hours, manage own context, and spawn sub-agents for parallel codebase work
✨
SAM 3 introduces concept segmentation
Joseph Nelson
✨
Pliny the Liberator and John V lead BT6
AI to AE's: Grit, Glean, and Kleiner Perkins' next Enterprise AI hit — Joubin Mirzadegan, Roadrunner
✨
Joubin Mirzadegan, now at Kleiner Perkins, is incubating Roadrunner, an AI-native rethink of CPQ and quoting workflows built for consumption-based pricing models
Joubin Mirzadegan
The Future of Email: Superhuman CTO on Your Inbox As the Real AI Agent (Not ChatGPT) — Loïc Houssier
✨
Loïc Houssier transitioned from applied cryptography and nuclear submarine optimization to CTO of Superhuman Mail, recently acquired by Grammarly
✨
Pim from General Intuition spun out Medal (12M-user game clipping platform) to raise $134M seed from Khosla for training world models on gameplay footage
✨
Fei-Fei Li and Justin Johnson's World Labs launched Marble, a generative world model creating editable 3D environments from text, images, and spatial inputs for games, film, VR, and robotics
✨
Tenex co-founders reveal model paying AI engineers for output rather than hours, enabling some to earn $1M+ annually with 10x productivity gains
✨
Deedy Das discusses his journey from Glean (enterprise search scaled to $7B AI-native company) to Menlo Ventures and early investment in Anthropic before revenue