GoKawiilTypeSafe AI, founded by ChatGPT co-creator, launches non-text model Jev
Diogo Almeida, a former OpenAI researcher who helped invent RLHF and build ChatGPT, has released a new transformer-based model called Jev through his startup TypeSafe AI. Unlike large language models, Jev doesn't generate text—it produces calibrated probability outputs, making it faster, cheaper and immune to hallucination. Developers have shown strong demand, with the company briefly unable to keep up with API traffic.
GoKawiil's interpretation of the reporting above, not reported fact.
Almeida's move signals a growing belief among AI insiders that pure language generation isn't the best fit for automation tasks like classification and decision-making in software systems. Early adopters like Vercel report Jev delivering results many times faster and more accurately than OpenAI's models for such tasks, suggesting a shift toward specialized, non-linguistic AI architectures for enterprise automation.
- Jev is a transformer model that outputs probabilities instead of text, avoiding hallucination.
- TypeSafe AI was founded by Diogo Almeida, a former OpenAI researcher behind RLHF and ChatGPT.
- Vercel reported Jev running 5 to 18 times faster than an OpenAI model for safety classification tasks.
Source: techcrunch.com — Tim Fernholz, 2026-09-18
Published there as: “A new kind of AI model from a ChatGPT inventor is thrilling developers”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.