The wiki

Engineering leader whose work at Superhuman combines AI email development with bottom-up adoption of AI tools and qualified measures of team throughput.

Part 2 of the January–February 2024 recap: Gemini’s million-token context, prompt transparency, anniversary demos of Lindy, RWKV, Pixee, and Julius, and listeners on evaluation and technical community.

Hugging Face researcher working on open foundation models, educational training data, and the SmolLM family.

Bloop co-founder whose work moved from AI code search to planning and reviewing coding agents, and whose public arguments examine what improving models leave for application builders.

PyTorch contributor and Lightning AI CTO whose work spans inspectable model optimization, open-source developer tools, and medical imaging.

Machine learning researcher focused on language-model data, former Olmo co-lead at Ai2, and co-leader of the Dolma open pretraining corpus.

Multimodal AI researcher and contributor to SigLIP and PaliGemma, whose career spans robotic perception, Google Brain, OpenAI and Meta.

Product security practitioner who argues that reusable AI instructions belong in the software supply chain, with review before distribution and approval enforced at execution.

Andon Labs cofounder and researcher studying how autonomous AI agents manage businesses, reason about physical spaces, and coordinate robots.

Part 3 of 3 of Latent Space’s January–February 2024 recap closes with listener questions about research coverage, AI user experience, and organizing a community around working demos.

Learning to represent sound lets Suno reuse transformer machinery for music. The harder questions are how to generate fast enough for listening, give people useful control, and judge whether a song moves them.

Software engineer and Vercel CTO in 2025–26 whose public work addresses agent reliability, evaluation and security, following a career building web frameworks at Google.

Mosaic cocreator, Netscape cofounder and venture investor advocating AI development, coding agents and technological acceleration.

Interpretability engineer and researcher whose Goodfire work spans concept-based image editing, research agents, and tools for understanding and controlling models.

OpenAI researcher whose work includes Codex, DALL·E 2 and GPT-4 vision leadership, appointed Chief Research Officer in March 2025.

Gradient co-founder and contributor to its long-context Llama 3 models, whose public views emphasize empirical evaluation and useful enterprise AI.

Meta founder whose AI strategy emphasizes developer control, personal agents, and scientific tools built with Priscilla Chan through Biohub.

Infrastructure investor, networking researcher, and developer whose AI work spans product tooling, compute economics, and generative 3D.

AI engineer working on Notion’s agent and creator of tools that combine AI-generated code, direct manipulation and automated verification.

Factory co-founder focused on practical autonomy in software engineering, developer ownership, and organization-wide outcomes.

Databricks cofounder and CTO, Berkeley computer scientist, and contributor to open-source data infrastructure and compound AI systems.

AI and data investor at a16z whose work includes LLM application architecture, practical open-model projects, and support for open source developers.

Gray Swan cofounder and Carnegie Mellon professor working on adversarial attacks, agent security, and enforceable safeguards for AI systems.

AI engineering educator and creator of Skills for Real Engineers, a collection of small, composable tools for planning, testing and designing software with coding agents.

Co-developer of Hex’s failure-funnel evaluation framework, with public work on diffusion-model quantization and relational-database modeling.

Chai Discovery co-founder and research leader whose public technical views emphasize simpler models, scalable learning and controllable molecular design.

Flux cofounder whose public explanations of AI hardware design emphasize circuit context, engineer judgment, and user-controlled workflow memory.

Max Krieger’s AI interface work includes Seemixer’s approach to programming by pointing and delve’s approach to exploring generated content.

Machine-learning researcher whose joint work spans variational autoencoders, symmetry-aware networks and graph convolutions, and whose materials-discovery work emphasizes practical impact and gradual automation.

AI engineer and educator associated with a PDF-chatbot project and coauthor of Learning LangChain with Nuno Campos.