Humans& (stylized as humans&) is a human-centric frontier AI lab that develops models designed to empower collaboration between people rather than replace them. Founded by alumni from Anthropic, xAI, Google, Stanford, and other labs (including Eric Zelikman, Andi Peng, Georges Harik, Noah Goodman, and Yuchen He), it raised a large seed round of approximately $480 million. The company focuses on innovations in long-horizon multi-agent reinforcement learning, memory, user understanding, and social intelligence so AI can act as connective tissue that strengthens organizations and communities. It emphasizes training paradigms that help people work together more effectively. Website: humansand ( dot ) ai
TypeSafe AI is a frontier AI lab focused on building a new class of large language models optimized for true no-human-in-the-loop automation rather than conversational assistance. Founded in mid-2024 by Diogo Almeida (ex-OpenAI, co-inventor of RLHF and InstructGPT), Erik Gafni, and Sasha Sheng, the company rethinks post-training objectives to prioritize calibrated decision-making, reliability, low latency, and cost efficiency. It aims to embed composable, deterministic intelligence directly into software systems so AI can handle complex workflows without constant human supervision or the sycophancy and hallucination issues common in preference-optimized models. The team comes from OpenAI, Google Brain, Meta/FAIR, and other places; the company remains largely in stealth while hiring and preparing releases. Website: typesafe ( dot ) ai