Latent Space

Captaining IMO Gold, Deep Think, On-Policy RL, Feeling the AGI in Singapore — Yi Tay 2

The podcast explores the advancements and future directions of AI, particularly focusing on reasoning and reinforcement learning (RL). Yi Tay, from Google's Gemini team in Singapore, discusses the team's pursuit of AGI and the shift towards on-policy RL, emphasizing its generalizability over imitation learning. The con...

Brex’s AI Hail Mary — With CTO James Reggio (acquired for $5B by Capital One!)

Brex CTO James Reggio discusses Brex's AI transformation, focusing on a three-pillar AI strategy: corporate AI adoption, operational cost reduction, and product AI features. Brex embraces hiring ex-founders, offering them opportunities to solve interesting problems with instant distribution to a wide customer base. Reg...

Artificial Analysis: The Independent LLM Analysis House — with George Cameron and Micah Hill-Smith

Artificial Analysis's journey, business model, and benchmarking methodologies are examined, highlighting their evolution from a side project to a key resource for AI developers and enterprises. The conversation covers the impetus for starting Artificial Analysis, driven by the need for independent model evaluations con...

[State of Research Funding] Beyond NSF, Slingshots, Open Frontiers — Andy Konwinski, Laude Institute

[State of Code Evals] After SWE-bench, Code Clash & SOTA Coding Benchmarks recap — John Yang

[State of MechInterp] SAEs in Production, Circuit Tracing, AI4Science, "Pragmatic" Interp — Goodfire

[State of AI Papers 2025] Fixing Research with Social Signals, OCR & Implementation — Team AlphaXiv

[NeurIPS Best Paper] 1000 Layer Networks for Self-Supervised RL — Kevin Wang et al, Princeton

[State of Context Engineering] Agentic RAG, Context Rot, MCP, Subagents — Nina Lopatina, Contextual

The podcast explores the current state and future directions of context engineering, featuring Nina Lopatina from Contextual AI. The discussion covers the shift from traditional RAG to agentic RAG, highlighting the benefits of query reformulation and the use of sub-agents for specific tasks. Lopatina and the host exami...

[State of Evals] LMArena's $100M Vision — Anastasios Angelopoulos, LMArena

[State of Post-Training] From GPT-4.1 to 5.1: RLVR, Agent & Token Efficiency — Josh McGrath, OpenAI

Josh McGrath from OpenAI discusses the shift from pre-training to post-training in AI model development, highlighting the increased complexity and infrastructure demands of reinforcement learning (RL) compared to pre-training. McGrath shares insights on OpenAI's shopping model, emphasizing its new interruptibility feat...

[State of RL/Reasoning] IMO/IOI Gold, OpenAI o3/GPT-5, and Cursor Composer — Ashvin Nair, Cursor

[State of AI Startups] Memory/Learning, RL Envs & DBT-Fivetran — Sarah Catanzaro, Amplify

The podcast explores the symbiotic relationship between data and AI, particularly how the modern data stack is evolving to meet the needs of AI development. Sarah Catenzaro from Amplify discusses the implications of the DBT-Fivetran merger, viewing it as a strategic move for IPO readiness rather than the end of the mod...

One Year of MCP — with David Soria Parria and AAIF leads from OpenAI, Goose, Linux Foundation

The Latent Space podcast features a discussion on the first anniversary of the Model Context Protocol (MCP) and the launch of the Agentic AI Foundation (AAIF). David Soria Parra from Anthropic reviews MCP's adoption by major tech companies like Microsoft, Google, and OpenAI, as well as protocol improvements including r...

Steve Yegge's Vibe Coding Manifesto: Why Claude Code Isn't It & What Comes After the IDE

In a live discussion at the AI Engineering Summit, Steve Yegge discusses Vibe Coding and AI engineering, highlighting the resistance from senior engineers with 12-15 years of experience who feel threatened by these new approaches. He advocates for abandoning traditional IDEs in favor of AI agents for coding, predicting...

⚡️GPT5-Codex-Max: Training Agents with Personality, Tools & Trust — Brian Fioca + Bill Chen, OpenAI

SAM 3: The Eyes for AI — Nikhila & Pengchuan (Meta Superintelligence), ft. Joseph Nelson (Roboflow)

⚡️Jailbreaking AGI: Pliny the Liberator & John V on Red Teaming, BT6, and the Future of AI Security

The podcast explores the landscape of AI model jailbreaking, focusing on its motivations, techniques, and implications for AI safety and security. Pliny the Liberator, a prominent figure known for jailbreaking AI models, discusses the central role of "liberation" in his work, emphasizing freedom of information and tran...

AI to AE's: Grit, Glean, and Kleiner Perkins' next Enterprise AI hit — Joubin Mirzadegan, Roadrunner

The Future of Email: Superhuman CTO on Your Inbox As the Real AI Agent (Not ChatGPT) — Loïc Houssier