What actually matters in vector databases in 2025, why “modern search for AI” is different, and how to ship systems that don’t rot as context grows.| www.latent.space
Can GPT-5 build better dev tools for itself? Does it improve its coding performance?| Latent.Space
Our friends at Roboflow ran the numbers today on GPT-5's underrated (but also improved) vision capabilities.| Latent.Space
The big reveal of GPT-5 was entirely unexpected but is welcome nonetheless - there's a router!| Latent.Space
We're excited to publish our hands-on review from the developer beta.| Latent.Space
On the heels of their $32m Series A: Why fast apply models got bitter lesson'd, pioneering the plan + act paradigm for coding, and why people are use coding agents for non-coding tasks| www.latent.space
What we learned from surveying the top Tiny Teams at the World's Fair| www.latent.space
It is ever easier to make wishes into reality, therefore we should be careful what we wish for...| www.latent.space
The Bitter Lesson vs Agent Harnesses & World Model, Solving Poker and Diplomacy, Debating RL+Reasoning with Ilya, what's *wrong* with the System 1/2 analogy, and the challenges of Test-Time Scaling| Latent.Space
Annotated notes on Andrej's talk at YC AI Startup School 2025| Latent.Space
The Modular/Mojo creator returns to the pod to discuss the CUDA monopoly, matching NVIDIA performance with AMD, and building a company of "elite nerds",| Latent.Space
AI Engineer World's Fair 2025 Recap — Recapping as much as we can, with exclusive infographic summaries from our friends at Thoth.ai!| Latent.Space
Emmanuel Amiesen is lead author of “Circuit Tracing: Revealing Computational Graphs in Language Models”, which is part of a duo of MechInterp papers that Anthropic published in March (alongside On the| Latent.Space
A WHOLE bonus track free for you to enjoy and Pro Tips for those joining us live.| Latent.Space
The cofounders of Sequoia-backed Factory.ai stop by the studio to dive into the story behind one of the most successful enterprise code agent platforms. Try not to make the Star Wars joke...| Latent.Space
Cheap SWE Agents are enabling teams with more millions in ARR than employees. Reflections after the OpenAI Codex and Google Jules Launches, Cognition DeepWiki, and LMArena's $100m raise.| Latent.Space
The World’s Fair is now TEN days away!| Latent.Space
ChatGPT Codex is here - the first cloud hosted Autonomous Software Engineer (A-SWE) from OpenAI. Josh Ma and Alexander Embiricos tell us how to WHAM every codebase like a power user.| Latent.Space
Cat Wu and Boris Cherny from the Claude Code team stop by to tell all!| Latent.Space
Selling GPUs to avoid bankruptcy, empowering researchers with short term clusters, and why CoreWeave is maybe a real estate business| www.latent.space
OpenAI dropped o3 pricing 80% today and launched o3-pro. Ben Hylak of Raindrop.ai returns with Alexis Gauba for the world's first early review.| www.latent.space
MCP's coauthors on the origin, challenges and future of the protocol.| www.latent.space
Learnings from Anthropic's extraordinarily successful Launch and Workshop| www.latent.space
2025's biggest surprise so far: Reasoning is less of a moat than anyone thought.| www.latent.space
How Ben Hylak turned from ol pro skeptic to fan by overcoming his skill issue.| www.latent.space
Announcing the theme of the second ever AI Engineer Summit. Apply now!| www.latent.space
Llama 2 lead and Llama 3 post-training lead Thomas Scialom of Meta/FAIR, on the Chinchilla trap, why Synthetic Data and RLHF works, and how Llama4's focus on Agents will lead us to Open Source AGI.| www.latent.space
Mar-Jun 2024 Recap: People are raising doubts about AI Summer. Here's why AI Engineers are the solution.| www.latent.space
Clémentine Fourier of HuggingFace on why you should stop using LLMs as Judges, what comes after MMLU, how prompts formatting sways benchmark results, and why leaderboards are GPU poor| www.latent.space
How we can use AI for as a "partner in thought", losing faith in long context windows for improved reasoning, and why we should stop anthropomorphizing LLMs| www.latent.space
1 longform interview, 12 more papers and 3 talks from ICLR 2024, covering Coding Agents like OpenDevin, the Science of Benchmarks, Reasoning and Post-Training, and Agent Systems!| www.latent.space
Scaling Llama3 beyond 1M context window with ~perfect utilization, the difference between ALiBi and RoPE, how to use GPT-4 to create synthetic data for your context extension finetunes, and more!| www.latent.space
Listen now | Why AI UX should let you supervise the process, not just the output of AI Agents, and how they reinvented the Notebook to help revolutionize Systematic, Transparent, Unbounded Academic Research| www.latent.space
5 recent Latent Space appearances for you to get your LS fix on all AI Engineering topics imaginable.| www.latent.space
The Deputy CTO of Microsoft joins the Latent Space crew & AIE co-founder Ben Dunphy to talk about Microsoft's embrace of AI Engineering and to present the AI Engineer World's Fair, coming to SF soon!| www.latent.space
Peak ChatGPT? Also: our usual highest-signal recap of top items for the AI Engineer from Feb 2024!| www.latent.space
Why Google failed to make GPT-3, how Adept is the "most misunderstood company" in AI, why multimodal knowledge work models like Fuyu are the future of AGI, and why Adept is NOT a research lab| www.latent.space
Listen now | swyx & Alessio discuss the research trends of January and the industry chaos of February 2024. Also: we celebrate the 1 year anniversary of Latent Space!| www.latent.space
Listen now | The PyTorch creator riffs on geohot's Tinygrad, Chris Lattner's Mojo, Apple's MLX, the PyTorch Mafia, the upcoming Llama 3 and MTIA ASIC, AI robotics, and what it takes for open source AI to win!| www.latent.space
Back to foundations! The origins of RLHF, sociology's influence on it, the tension between human vs synthetic data, and emerging research in the field| www.latent.space
The Data Wars, The War of the GPU Rich/Poor, The Multimodality War, The RAG/Ops War. Also: our usual highest-signal recap of top items for the AI Engineer from Dec 2023!| www.latent.space
Our selection for AI Engineers: Word2Vec (with Jeff Dean), State Space Models (with Chris Ré), Emergence Mirage, DPO, Datablations, QLora, LlaVA, DataComp, Tree of Thought, CogEval, Voyager| www.latent.space
What are our LLMs actually trained on, and are we actually running out of data?| www.latent.space
Listen now (52 mins) | How Hex is putting Magic into notebooks, why LLMOps is an "iron mine" and not a "gold rush", and how RAG is RecSys for LLMs| www.latent.space
Notion’s RenAIssance, why Chat is NOT all you need, why knowledge work is more than generating text, and the AI-augmented workspace of the future. Plus: AI×UX NYC meetup recap from Paul Butler!| www.latent.space
Bridging the Capability Overhang from Generative AI to Generative UI| www.latent.space
Listen now | On learning AI fast and how AI's learn fast, the mission of doing more deep learning with less, inventing ULMFiT and why it's now wrong, and how to play the AI Discords game| www.latent.space
Emergent capabilities are creating an emerging job title beyond the Prompt Engineer.| www.latent.space
