Jev's Decision Model + NylonME: The 'Disassembly Era' of AI Is Here
A Model That Can't Write Just Broke the Internet On September 15, TypeSafe AI released a model called Jev. The company was founded by former OpenAI researcher Diogo Almeida — a core author of RLHF and InstructGPT, the research that turned ChatGPT into ChatGPT. After two years in stealth and $40M led by DCVC, what they shipped was surprisingly restrained : Jev can't write a single word. No text…
On September 15, TypeSafe AI introduced Jev, an AI model that cannot produce text, code, or chat responses. Instead, it delivers structured decisions with calibrated probabilities for yes/no, multiple-choice, and scale score questions. The model, developed by former OpenAI researcher Diogo Almeida, is 20-200 times faster and 40-400 times cheaper than cutting-edge LLMs for decision-making tasks.
With a pricing of $0.042 per million input tokens and free outputs, Jev quickly gained popularity among developers. Within hours of its release, users began utilizing Jev for tasks like real-time flight lookups and compressing large context windows. Jev's purpose lies in addressing a significant issue in agent systems: a substantial portion of LLM calls are not necessitating language generation.
Researchers and practitioners are now breaking down intelligence into distinct components, such as System 2 reasoning (slow, generative thinking) and System 1 decision-making (fast, intuitive judgments). TypeSafe AI's Jev model exemplifies this transition, turning complex judgment calls into swift function invocations with near-zero costs, thereby driving demand for decision-making components.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
