Article
CommunityJev Guides & Articles
BIG CHANGE: Jev and AI decisions inside everyday software
A BIG CHANGE analysis of the Diogo Almeida Latent Space interview: Jev's most consequential proposition is the unit of automation, a bounded judgment inside an existing process, and the launch claim that Jev can't hallucinate needs a narrow reading because schema matching does not establish truth.
- Format
- Article
- Reading time
- 3 min
- Published by
- Community
- Author
- BIG CHANGE Editorial
- Last verified
- Oct 2026
Highlights
- The piece reviews the September 21 Latent Space interview with Almeida and the docs, and says it did not independently benchmark Jev.
- Its core reading: a company may automate a bounded judgment inside an existing process before it can responsibly hand over the whole process.
- On can't hallucinate: returning damaged is valid even when the package was merely late, because schema matching does not establish the truth of an answer.
- Buyers should run three separate tests: a schema test, a factual test and a policy test; passing one does not substitute for the others.
- Almeida rejects perfect calibration and acknowledges model errors; the confidence field is a statistic from the answer distribution, not a measured probability.
Watch out
Copy reviewed against the published analysis on 2026-10-17; the digest text was truncated near the close, so the final buyer-comparison passage was not fully readable. The post analyzes an interview and public documentation rather than independently benchmarking Jev, and its customer-service examples are labeled hypothetical rather than deployment results.
Questions about this entry
Short answers drawn from the linked project or page.
That Jev's most consequential proposition concerns the unit of automation: a company may be able to automate a bounded judgment inside an existing process before it can responsibly hand over the whole process.
BIG CHANGE says it reviewed the complete English captions of Latent Space's two-hour, 22-minute September 21 interview with TypeSafe CEO Diogo Almeida and checked the product documentation, without independently benchmarking Jev.
TypeSafe's guarantee covers matching the permitted output schema, not the truth of the answer. An application offering damaged, late and other can get the valid value damaged even when the package was merely late.
A schema test for whether software can consume the answer, a factual test for whether it agrees with the evidence, and a policy test for whether the resulting action is allowed. Passing one does not substitute for passing the others.
At 1:09 in the interview he rejects a suggestion of perfect calibration and acknowledges model errors. The post adds that the API's confidence field is a statistic derived from the answer distribution, not an independently measured probability that the selected answer is correct.
Choice, which selects among specified options; Score, which evaluates against a rubric; and Noul, which expresses a statement's probability on a zero-to-one scale. Choice and Score return distributions and a confidence field; Noul does not have that separate field.
More components mean more interfaces to maintain, questions can accidentally omit context, and two apparently independent judgments may depend on the same misleading evidence. A workflow assembled from individually acceptable parts can still produce an unacceptable result.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
Hacker News: Simon Willison's overview of Noul, Choice, and Score, and where a decision model fits.
benwerd53 pointsJev introduces a new shape of LLMRead the thread on Hacker NewsHacker News: swyx posts the Latent Space interview with Diogo Almeida.
swyx4 pointsJev: System One Models for Prod, Not God – With Diogo Almeida, CEO, TypeSafe AIRead the thread on Hacker NewsX: Diogo Almeida in five minutes
X: LLMs vs. Jev, clearly explained
More Jev resources like this
Ian Maurer's ThinkThen analysis argues Jev's faster and cheaper are measurable today while its better is a bet on RLCD: unlike RLHF, which pleases humans, and RLVR, which optimizes benchmarks, Jev is trained to be reliable for programmatic use, in Diogo Almeida's own framing.
Jev Guides & ArticlesArticle#community#press#explainer
Community
Rogue Security's post uses Jev and System One models to frame agentic security as a decision problem: the same send_file action should be allowed, blocked or reviewed depending on who requested it and where the data goes, illustrated with policy case files rather than product results.
Jev Guides & ArticlesArticle#community#press#security
Community
A podcast episode on Jev as a decision model rather than a chat model: hosts Luis Calderon and Carina discuss roughly 20x faster Coach testing, why that preliminary figure differs from TypeSafe's vendor benchmarks, and why type safety does not guarantee a correct decision.
Jev Guides & ArticlesArticle#community#press#podcast
Community
VentureBeat's Sam Witteveen argues a large share of enterprise AI only needs a label, not generated text. Per TypeSafe, Jev returns choices, scores and yes/no answers in 70 to 500 ms at $0.042 per million input tokens, with output tokens free.
Jev Guides & ArticlesArticle#community#press#explainer
Community
TWIML episode 779 features TypeSafe co-founder and CEO Diogo Almeida on Jev, the company's model for bringing fast, reliable intelligence directly into software, and why text-generating models are poorly suited to many real-world automation decisions.
Jev Guides & ArticlesArticle#community#press#podcast
Community
On The MAD Podcast, Matt Turck talks with TypeSafe CEO Diogo Almeida about why intelligence per dollar matters more than enterprise adoption, and why RLVR benchmark gains still leave models 95% Einstein, 5% Mr. Bean.
Jev Guides & ArticlesArticle#community#press#podcast
Community
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.