The future of AI should be transformative—not incremental. Today, leading labs like OpenAI and Anthropic focus on LLMs optimized for human assistance. Yet, instead of unleashing a sci-fi vision of commoditized intelligence and economic revolution, we've settled for copilots and marginal productivity gains.

Our team—led by Diogo Almeida (OpenAI, Google Brain, co-inventor of InstructGPT, RLHF, ChatGPT, and GPT-4), Erik Gafni (Ravel co-founder, multimodal AI), and Sasha Sheng (Meta AI)—is redefining this paradigm.

Leveraging our co-founder's pioneering work in Reinforcement Learning with Human Feedback (RLHF) and Instruction-Following models at OpenAI, we've fundamentally redesigned LLMs for automation. By re-imagining the entire stack from the ground up, we optimize explicitly for AI-to-computer interactions rather than human-in-the-loop scenarios.

Let's automate the boring stuff: joinus@typesafe.ai