AI6/10
Prajwal Tomar Explains Jev as a Non-Generative Decision Model
Prajwal Tomar describes Jev, a free AI model built by a ChatGPT co-inventor that scores options rather than generating text. He argues this design avoids hallucination and makes decision-making cheap, and links to a post about building a harness with Jev.
Original post · 1 min read
If you're still confused about what Jev actually does, read this.
The guy who co-invented ChatGPT spent two years building an AI that cannot write a single word. On purpose. Then made it FREE.
Every AI you've used talks to you.
Jev doesn't talk at all. It just decides.
You give it an email or a lead and some options, it spits back how likely each one is. No essay, no explanation.
It literally can't hallucinate because it never writes anything.
The talking part of AI is solved. The deciding part just got stupidly cheap.
That's where all the boring money is.
If the launch went over your head, read this one.
The guy who co-invented ChatGPT spent two years building an AI that cannot write a single word. On purpose. Then made it FREE.
Every AI you've used talks to you.
Jev doesn't talk at all. It just decides.
You give it an email or a lead and some options, it spits back how likely each one is. No essay, no explanation.
It literally can't hallucinate because it never writes anything.
The talking part of AI is solved. The deciding part just got stupidly cheap.
That's where all the boring money is.
If the launch went over your head, read this one.
Sydney Runkle @sydneyrunkleBuilding a Harness with Jev — Agents run in a loop: an LLM decides what to do, a tool executes, a model evaluates the results, and then continues in that loop until the task is complete.
Agents and LLMs were initially difficult to
♥ 315 · ⟲ 17 · 👁 81.6KView on X ↗
