TypeSafe AI announced System One Models, a new class of frontier models optimized for structured decision-making and automation rather than natural language generation. Their first model 'Jev' uses a novel training method called RLCD and claims to be 100x faster than existing LLMs while eliminating hallucinations.
Background
TypeSafe AI was founded by Diogo Almeida, an OpenAI veteran who contributed to ChatGPT research. After two years in stealth, they're pivoting from chat-focused LLMs to automation-oriented models.
- Source
- Lobsters
- Published
- Sep 18, 2026 at 08:05 AM
- Score
- 8.0 / 10