TypeSafe's Jev model bets on machine-native intelligence over text-optimized LLMs
TWIML AI Podcast · rss · 2026-10-07
TWIML AI Podcast episode 779 features Diogo Almeida, co-founder and CEO of TypeSafe, on their newly released model Jev.
- Machine-native intelligence: Almeida argues text-optimized models are poorly suited to many decisions required for real-world automation.
- RLCD: Jev uses reinforcement learning from calibrated decisions, making calibration and reliability central to AI as a software primitive, distinct from classifiers and LLM-based approaches.
- Models vs. code: He advocates more engineered AI systems rather than one model doing everything, and discusses how Jev-like models could reshape agents, tool use, and AI software architecture.
More from Models
- Dev warns OpenRouter share, cache hit rate, latency stats are easily gamed for marketing — charles_irl · 2026-10-07
- OpenAI's unreleased model reportedly proves quasi-Riemann hypothesis with Lean proof — ChrisGPT · 2026-10-07
- Fed Claude Opus my blurry handheld Saturn shots, it fused them into one best image — adonis_singh · 2026-10-07
- Two Labs, One Race: Anthropic vs OpenAI Frontier Model Release Timeline, 2023–2026 — Medical-Sky7620 · 2026-10-07
- Claude was given robot skin — Opus was curious but anxious about hooking up — repligate · 2026-10-07
- Gemini 2.5 Pro retiring October 20, 2026, users say goodbye — hargup13 · 2026-10-07