The wiki

An AI search company that evolved from answering programmers’ questions to generating interactive answers and custom mini-apps.

Better context and stronger models turned a search product into a coding assistant. Specialized training promised more control over cost and capability; consistent correctness remained the harder problem.

AI workshop presenter whose teaching spans generative music, creative control, and local exploration of training datasets.

Applied Intuition connects autonomy models to the engineering that makes them usable: sensor supervision, reliable operating systems, validated simulation and inference that fits the machine.

An application-security company combining AI agents and deterministic code transformations to triage scanner findings and deliver fixes through developer workflows.

Better images are only part of the problem. Playground’s ambition connects model quality, precise editing, visual control, and feedback from people making things.

Pseudonymous AI jailbreak researcher who publishes prompt techniques and argues for transparency, capability exploration, and harm reduction outside the model.

An AI foundation-model company built around coding, execution feedback and an integrated Model Factory, with a consequential NVIDIA licensing deal reported in August 2026.

A foundation-model company pursuing software intelligence through code execution feedback, an integrated Model Factory, and the Laguna family of agentic models.

An AI gateway company that expanded from LLM operations into governance and agent infrastructure, acquired by Palo Alto Networks in May 2026.

Josh McGrath on the engineering behind reinforcement learning, efficient agents, and useful long context, in a conversation at NeurIPS 2025.

Linear attention, state-space models and RWKV reduce the cost of processing sequences. Their harder challenge is preserving the information a task needs—and making that computation fast on real hardware.

An AI podcast hosted by Daniel Whitenack and Chris Benson, launched with Changelog in 2018 to make AI practical, productive, and accessible.

From incompatible datasets to efficient training: how Axolotl turns model-building experience into reusable configuration.

Conviction partner and former Neeva engineer whose writing examines AI business durability, search infrastructure and enterprise automation.

Creator of Galileo’s Agent Leaderboard, focused on evaluating whether AI agents complete tasks and turning failures into verifiable improvements.

Machine-learning researcher and DatologyAI founding member studying how training data shapes foundation-model capabilities, memorization, and safety.

Prime Intellect builds open models and infrastructure for improving them, spanning distributed compute, reinforcement learning, executable environments, and agent tooling.

Princeton University's AI work spans software-engineering benchmarks, deep reinforcement learning, agent reliability, and an expanding institutional structure for AI and data science.

Pediatrician and philanthropic co-founder whose work connects AI-powered biological research with the goal of improving patients’ lives.

Braintrust connects production logs, comparative evaluation, and interactive debugging so teams can improve AI behavior—and decide which work belongs in ordinary code.

An open-source AI evaluation and security company that expanded from application testing into model-file scanning and agreed to be acquired by OpenAI in March 2026.

Atlan co-founder who advocates treating organizational context as governed infrastructure for enterprise AI.

Google DeepMind researcher applying machine learning to scientific problems, including protein structure prediction.

The organization behind Python’s Pydantic validation library expanded into AI agents, observability, evaluation and controlled execution of agent-written code.

Applied Intuition’s co-founder and CEO focuses on putting AI into physical machines, with an emphasis on production constraints and sustainable commercialization.

AI systems engineer whose work spans distributed model training, practical training guidance, and hardware-aware model design.

Amp CEO and Sourcegraph co-founder whose work centers on code understanding, autonomous coding agents, and verification of generated software.

AI researcher whose collaborative work spans unsupervised feature learning, sequence-to-sequence models, language-model pretraining, neural architecture search, and chain-of-thought prompting; cofounder of Discovery Loop.

Alibaba’s foundation-model team develops open-weight language and multimodal models, agent tools and interpretability research alongside hosted AI services.