The wiki

Amplify Partners investor, AI engineering survey organizer and Barrchives host whose writing examines agent controls, evaluations and AI product decisions.

Baseten builds infrastructure for serving, optimizing, and training AI models, and is expanding into agent execution through its acquisition of Blaxel.

A wearable personal AI company building memory and assistance from everyday conversations, later incorporated into Amazon.

Wearable audio gives Bee context without constant explanation. Turning that context into useful help requires editable memory, selective attention, and careful choices about power and permission.

Co-founder of the AI Engineer conferences, with work spanning vendor-neutral technical events and the announcement of LLM-powered attendee matching.

Replicate co-founder and Cog co-creator whose work makes machine learning models usable by software developers.

Software developer and educator building AI-assisted tools for development, conversation, and personal knowledge.

Machine-learning researcher whose collaborative work spans diffusion models, text-to-3D generation, dynamic scene reconstruction and neural-network learning.

A benchmark tests particular capabilities under particular conditions. Understanding those choices helps explain both model progress and the gap between a leaderboard score and a useful product.

Clémentine Fourrier on reproducible comparisons, the limits of preference rankings, and the engineering choices behind model scores, from the July 2024 conversation.

Part 3 follows an agent’s failed travel plan into the science of benchmarks, then examines how Self-RAG and process supervision use intermediate judgments to select better outputs.

Datology’s case for optimizing filtering, rewriting and training order together—and measuring their value in learning efficiency, model capability and inference cost.

Sourcegraph co-founder and Amp co-founder whose work on AI coding tools emphasizes code context, transparency, and reliable automation.

How evolutionary clues, generative models and inference-time search support molecular design—and why useful predictions still have to survive experiments.

Part 4 of Latent Space’s NeurIPS 2023 recap closes with the common ingredients of attention-free models, their remaining bottlenecks, and the different demands of interactive responses and bulk processing.

Part 2 of Latent Space’s Q2 2024 recap examines the infrastructure agents need, the work customers will pay for, and the possibilities and limits of ChatGPT’s early advanced voice mode.

Brushes, database columns, and inspectable programs show how interface design can turn language-model capabilities into useful, controllable work.

AI engineer and RiskAngle cofounder whose documented work includes coding-agent integrations and guidance on preserving reasoning during tool use.

Coauthor of Diffusion Transformers and former head of Sora, with research spanning image generation, learned optimization and visual alignment.

A biomedical research organization combining experimental biology, biological data and AI to build tools for understanding and preventing disease.

A biomedical research nonprofit combining AI models, experimental biology and open datasets to help scientists understand cells and develop approaches to disease.

The developer of FLUX evolved from text-to-image models into reference-based editing and unified image, video, audio, and action modeling, combining downloadable weights with commercial APIs.

Biomedical AI researcher associated with scGPT and Xaira’s X-Cell, whose work spans single-cell analysis, perturbation prediction and medical AI education.

Stronger code generation made Bolt possible. A browser runtime, an observable repair loop, and integrated deployment made it useful. Enterprise agents face the same challenge under very different constraints.

A public-benefit biomolecular AI company that grew from MIT research into open models, molecular-design pipelines and hosted scientific software.

Creator of Claude Code, known for a composable terminal agent and an engineering approach that favors simple mechanisms around capable models.

An enterprise content platform extending file collaboration, permissions, and governance to AI agents.

An AI evaluation and observability company that expanded from testing prompts and models into production-log infrastructure and agent investigation tools.

Machine-learning scientist working on RNA drug discovery at Atomic AI and cohost of Latent Space’s AI for Science podcast.

Nomic cofounder and AI researcher whose collaborative work spans open language models, reproducible embeddings, and computational approaches to how meaning changes.