The wiki

Model Context Protocol gives AI applications a shared interface to external data and actions. It standardizes how applications request capabilities; builders decide how to use the returned context and which actions to allow.

Why Model Context Protocol separates tools, resources, and prompts—and how client support, transport, authorization, and governance shape what developers can build.

A venture capital firm investing across AI models, infrastructure and applications, with an Anthropic partnership and enterprise AI market research.

The company behind Facebook, Instagram and WhatsApp develops AI frameworks, downloadable models and consumer agents, alongside immersive hardware and large-scale computing infrastructure.

Meta’s AI work spans open research, PyTorch, downloadable Llama models, and a consumer assistant that later incorporated models from Meta Superintelligence Labs.

Meta Superintelligence Labs develops Meta’s Muse models, alongside research in visual segmentation and agent coordination.

Meta’s AI organization behind the Muse reasoning and media models, with work spanning model training, agent orchestration, and personal AI products.

An AI evaluation nonprofit measuring autonomous capabilities, real-world productivity and risks from frontier developers’ internal use of agents.

OpenAI research leader whose work spans human feedback, coding-agent evaluation and alignment, with an expanded safety role reported in July 2026.

Artificial Analysis cofounder whose work connects independent AI evaluation with the practical trade-offs of building applications.

Co-founder of Markprompt, whose work combines documentation retrieval, configurable AI behavior, and an interest in interfaces beyond chat.

Co-founder of Phind and founder of Standard Signal, working on AI search, programming assistance and autonomous trading.

Cursor co-founder whose public work explains the shift from AI autocomplete to delegating software development to autonomous agents.

OpenAI API tech lead in 2024 and co-lead of Structured Outputs, combining model training with engineering constraints to make model responses usable in software.

From personal-computer software to cloud infrastructure, Microsoft combines AI research, training systems, developer tools, work applications and model partnerships.

Microsoft’s research organization combines basic and applied computing research, with AI work spanning synthetic training data, compact language models and multi-agent software.

Developer of Marble, an AI tutor for LeetCode, and presenter of OpenCode, a programming tutor centered on critical thinking.

AI researcher and entrepreneur whose work includes Databricks’ Dolly models, AI-assisted financial research and studies of social and economic networks.

Instagram’s co-founder and former CTO moved from leading Anthropic’s product organization into experimental AI products, with a focus on increasingly delegated agent work and its human bottlenecks.

Suno co-founder and CEO whose work combines generative audio models with a focus on making music creation accessible to more people.

Technology executive whose work spans Microsoft AI, Yandex and Shopify, with public views on agent critique, automated experimentation and customer simulation.

AI researcher and technology leader whose work spans semantic video indexing, urban mobility analytics, customer support and applied AI in financial services.

An AI foundation-model company developing language, video, speech and music models, with a growing enterprise and developer platform business.

Lighthouse founder whose work spans immigration automation, AI-focused investing, and public writing about technology and human agency.

Reflection co-founder and CEO whose collaborative reinforcement-learning research spans visual control, in-context learning, and language-model reward modeling.

A large model needs more than a fast inference engine: it needs capacity that scales, predictable latency and workflows that keep the product running.

Mistral AI develops language and speech models, enterprise training tools and private compute infrastructure, combining downloadable models with commercial applications and services.

Voxtral TTS generates speech one frame at a time. The same conversation connects specialized models, enterprise training, formal proofs, and learning from agents that work for hours.

Positron AI’s CEO, formerly Lambda’s COO, advocates specialized inference hardware and simpler access to AI infrastructure.

An AI cloud platform that grew from Python functions into elastic inference, agent sandboxes and training infrastructure.