Article
CommunityJev Guides & Articles
ThursdAI: Jev's System 1 launch, with Allie Laabs
ThursdAI's September 17 episode covers TypeSafe's Jev launch with founding DevRel Allie Laabs: choice, score and null primitives, $42-per-billion input pricing, and a live X timeline classifier running at 12 tweets a second.
- Format
- Article
- Reading time
- 3 min
- Published by
- Community
- Author
- Alex Volkov
- Last verified
- Oct 2026
Highlights
- Founding DevRel Allie Laabs joins at 1:38:48, two days after launch, to explain choice, score and null, and why output tokens are free.
- TypeSafe's own comparison claims 133x faster and 444x cheaper than competitive-level LLMs on the decision tasks tested - company numbers.
- Host Alex Volkov wired Jev into his X timeline in 25 minutes: 12 tweets scored a second, roughly 4-5x faster and 20x cheaper than Qwen 3.8 27B on Cerebras.
- The stats cite Cua's computer-use test: Jev handled 71 of 80 decisions at about 0.1 s versus 4 to 5 seconds for Astra, a roughly 400x price difference.
- The wrap-up caveat: can't hallucinate means the JSON output is always valid, not that answers are right, and Jev benchmarks below Astra on raw intelligence.
Watch out
Copy reviewed against the published episode page on 2026-10-17. The episode aired two days after launch; pricing, speed and comparison figures are TypeSafe's own numbers repeated by the show, and the in-show demos are the host's own builds rather than independent benchmarks.
Questions about this entry
Short answers drawn from the linked project or page.
Founding DevRel Allie Laabs of TypeSafe AI joins host Alex Volkov at 1:38:48, two days after launch. The broader episode also features David Pawlan, Francesco Bonacci and panelists including Wolfram Ravenwolf, Peter Gostev and Nisten Tahiraj.
Choice picks from discrete, mutually exclusive options and returns the full probability map; Score rates along a rubric described in natural language; Null answers a yes/no question with the probability it is true.
$42 per billion input tokens with output tokens free - 'too cheap to meter,' as Allie Laabs puts it. TypeSafe's own comparison claims 133x faster and 444x cheaper than competitive-level LLMs on the decision tasks they tested.
He got access the night before and spent about 25 minutes wiring Jev into his X timeline extension: every tweet scored in real time, 12 tweets a second while scroll-spamming, roughly 4-5x faster and 20x cheaper than Qwen 3.8 27B on Cerebras for under a cent of usage.
Jev handled 71 of 80 computer-use decisions at roughly 0.1 s median latency, against about 4 to 5 seconds for Astra - a roughly 400x price difference, as cited in the episode's stats.
Wolfram Ravenwolf noted that 'can't hallucinate' only guarantees valid JSON - the answers can still be wrong, and Jev benchmarks below Astra on raw intelligence. Classifiers are not new either; changing the question in natural language instead of retraining is.
No. The wrap-up notes it is text-only for now - no images, no hot-dog-or-not - even as the same episode's stats cover computer-use decisions and a live timeline classifier.
More Jev resources like this
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
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
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
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.
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
Blackwell Systems' adversarial technical due diligence on Jev, built only from public material and live API calls. The author reproduces TypeSafe's published addressee numbers within a point and finds a real but noise-type-specific robustness moat.
Jev Guides & ArticlesArticle#community#press#competition
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.