Introducing System One Models and Jev
Introducing System One Models and Jev view on hacker news Posted by albelfio with a score of 1394
Introducing System One Models and Jev view on hacker news Posted by albelfio with a score of 1394
No BS check run yet — press ⚖ to extract this story's claims and verify them against independent sources.
Introducing System One Models and Jev view on hacker news Posted by albelfio with a score of 1394
Introducing System One Models & Jev - TypeSafe AI Blog Designed for software automation, the new architecture replaces text generation with parallel, type-safe probabilistic decisions. Its first release, Jev, claims to eliminate hallucinations while operating up to 200 times faster and drastically cheaper than conventional LLMs. https:// typesafe.ai/blog/introducing-s ystem-one-models-and-jev
expect models like this to help control the 'motor skills' of SDEV, helping make them far safer than human drivers, possibly relatively soon. typesafe.ai/blog/introdu... TypeSafe AI is an AI lab building machine-native intelligence infrastructure for automation, designed to make decisions within software. Try our first System One Model, Jev, in early access.
# jev looks really interesting. I have mixed feelings about it. On one hand, this kind of optimization is a great step toward stopping the awful waste of # LLMs getting shoved into anything and everything. It's actual # ai # ml innovation. On the other, fast, cheap decision-making? Some ghoul is gonna stick it in drones and turn them loose on people. Scarier than OpenAI and Anthropic in many ways.