What people are shipping with Jev.

JEV C’EST UNE DINGUERIE MERCI @typesafeai

@lgandecki @klaudworks @pedronauck @typesafeai read through the convo here and seeing a generally disturbing trend of people claiming to have fixed context compression with jev wit

@CryZe107 @joshbuildings @theo @kwuchu @typesafeai @maubaron Also FYI Astra already thoroughly defeated Jev in chess. But then again, it's not meant to be the smartest (yet).

Google has their own better Gemini Jev in-house with whatever the name. When there’s a demand they will release one separately. Meanwhile so many useful ideas with @typesafeai Jev

@typesafeai @OpenRouter “Typed decisions, not text to parse” captures the value perfectly. Removing the JSON prompting and validation layer makes Jev feel like infrastructure, not

@thekitze BOZO SCORE 8.7 / 100 NOT A BOZO · Shipping Rabbit grift █░░░░░░░░░ 6 cope ░░░░░░░░░░ 2 ego █░░░░░░░░░ 5 hollow █░░░░░░░░░ 4 form ██░░░░░░░░ 14 "wrote the r

You know why so many hard core engineers are excited about @typesafeai Jev? Because it’s SMALL and SCOPED. Reducing features FTW! https://t.co/T9ANz6Bymk

Finally got access to Jev from @typesafeai ! https://t.co/Fo3VFkV6k9

@atomic_chat_hq @typesafeai Damn, I was invested in this clip. Next scroll material: see jev or llm evade missiles on a stream. Le jev: https://t.co/lZhd5sapI6

@wobsoriano @agent_device @typesafeai ui testing is a great use case for jev.

@rafalwilinski @kwuchu @typesafeai How does jev slot in here ?. I’ve been doing similar things using @owickstrom bombadil, but it uses properties that I’ve defined. From my experi

Concrete real world use case with JEV from @typesafeai Filling PDF forms: ~1.5s to fill, ~500ms to classify - insane numbers Especially useful when the input data is dirty (bad

Jev is nasty at coaster design I'm #loving jevin @typesafeai https://t.co/C9kfkOvskW

Who is doing @typesafeai Jev ultrafast but for local computer use, not the browser? https://t.co/dYC6UtIhJ9

@mvanhorn @typesafeai Holy smokes. Been using this for 3 hours wow. Context, memory, latency… pretty much all solved. Yesterday Claw was 80% of the value prop of leading Perplexit

@_pi0_ BOZO SCORE 18.6 / 100 NOT A BOZO · Has Receipts grift █░░░░░░░░░ 9 cope █░░░░░░░░░ 6 ego █░░░░░░░░░ 11 hollow █░░░░░░░░░ 8 form ███░░░░░░░ 27 "annoyingly, thi

@KinasRemek @typesafeai Ten model to początek dużych zmian. Pomijając generowanie tekstu/kodu większość tokenów agentycznych idzie na tool use i klasyfikację stanów. Jev zdaje

Hey @typesafeai ! We’ve just applied for JEV early access from Ante. Excited to try it and explore how it could fit into what we’re building around AI agents and evaluation. 👀 h

@Shasko @rafalwilinski @typesafeai Yea everyone doing something great with Jev model its great

@mikegee @typesafeai BOZO SCORE 63.9 / 100 CERTIFIED BOZO · API-Cost Accountant grift ██████░░░░ 61 cope █████░░░░░ 48 ego █████░░░░░ 45 hollow ███░░░░░░░ 31 form ███