jevs
13 stories and discussions about jevs, aggregated from every source we track.
How Jev's typed probabilistic decisions, LangGraph workflows, and Tenuo task-scoped warrants combine in a dependency upgrade agent.
TypeSafe's Jev is pitched at exactly the decisions that sit around a phone call. We put it on three real jobs against 17 models: it is 1.7 to 3.3 times faster everywhere, level with the best on judging and extraction,…
Yet another experiment that checks Jev's talking skills. - zie1ony/jev-talks
Deciphering what the Jev architecture is and how I would scale it
Someone on Hacker News suggested fingerprinting Jev’s tokenizer. The API returns how many input tokens it billed you for, so I searched for strings that separate known tokenizers and compared their counts to Jev’s.
Examining Jev's probability calibration, structured outputs, latency, and cost against open source LLMs, with experiments and reliability diagrams.
TypeSafe AI's Jev turns unstructured input into typed decisions and probabilities. We rebuilt the API with an open model, reached 113ms median latency on
Jev’s schema guarantee held in every call. It does not cover which valid answer comes back—so I tested whether the text being judged could choose for it.
What Jev's pricing on 'judgment' reveals, and why I'm building a model that does only five things.
I probed Jev with 10,000 API calls to work out roughly how it’s built, and why most of the grifter takes on X are completely wrong.
A short note on Jev’s generalization, possible encoder-style architecture and training, and Choice and Noul API examples.
Fast thinking models make you think slow again.
Developers test what they can build with TypeSafe's fast, typed decision model