# Jev by Typesafe AI 392 Hacker News commenters recommended Jev by Typesafe AI and 170 criticized it, November 2025 to November 2026. It was recommended in 2 of 13 months. **Verdict:** A cheap, low-latency general classifier you redefine with a prompt instead of retraining, though critics call it overhyped and little more than standard LLM tech. ## Why people recommend it - Very cheap and low latency (8 comments): "the idea that you can get a general classifier that is very accurate across any domain and on any modality with minimal latency and a very low price point is wild" ([source](https://news.ycombinator.com/item?id=49859333)) "I've been running jev in some specific use cases and see much better speed and better accuracy than just using structured output from gemma/qwen." ([source](https://news.ycombinator.com/item?id=50111970)) "Jev gets you the zero-shot, zero-infra benefits of LLMs, while being closer to the speed and cost of traditional ML classifiers" ([source](https://news.ycombinator.com/item?id=49807008)) - Change tasks by prompt, no retraining (6 comments): "the real win with jev isnt beating a trained classifier on accuracy, its that you can change what you're classifying by editing a prompt instead of retraining and redeploying a model." ([source](https://news.ycombinator.com/item?id=49808268)) "it allows you to define a new semantic decision at runtime in natural language, without training a model for that task" ([source](https://news.ycombinator.com/item?id=49872751)) "Jev is prompted with natural language, so it is flexible and a good fit to replace subagents for certain tasks" ([source](https://news.ycombinator.com/item?id=49810452)) - Consistent, well-calibrated probabilities (5 comments): "The core feature is good quality out-of-the-box domain and cross-domain calibrated probability you can use as a threshold." ([source](https://news.ycombinator.com/item?id=49842286)) "On actual NLP problems (including ones with uncertainty under human review) it does appear to be well calibrated" ([source](https://news.ycombinator.com/item?id=49937555)) "This one public benchmark (the only I've seen) seems to give the open versions way too much credit." ([source](https://news.ycombinator.com/item?id=49787750)) - Helps mitigate prompt injection (3 comments): "Especially the “zero hallucination” output and the (if it works how the documentation make it sound) prompt injection resistant architecture." ([source](https://news.ycombinator.com/item?id=50026150)) "the huge benefit in real systems for Jev like solutions i see is the cybersecurity / prompt injection mitigation" ([source](https://news.ycombinator.com/item?id=49766920)) "My AI harness uses Jev or a compatible model to approve shell commands that look safe without asking the user." ([source](https://news.ycombinator.com/item?id=49932279)) ## What people warn about - Overhyped with fake demos (5 comments): "What really pissed me off was that y'all got manipulated into thinking it was something revolutionary, while really it was a marketing stunt." ([source](https://news.ycombinator.com/item?id=50111904)) "There really isn't any architectural magic to Jev, it's just a straightforward application of normal LLM tech with some good marketing." ([source](https://news.ycombinator.com/item?id=49886415)) "It is impressive, but all the hype and fake demos are selling it as a model that is as smart as frontier reasoning LLMs" ([source](https://news.ycombinator.com/item?id=49850779)) - Little beyond standard LLM tech (5 comments): "There really isn't any architectural magic to Jev, it's just a straightforward application of normal LLM tech with some good marketing." ([source](https://news.ycombinator.com/item?id=49886415)) "None of that makes sense, if you can get 90% of the performance from an open source model that's been available for years." ([source](https://news.ycombinator.com/item?id=49767918)) "The consistency was nice to see, BUT, as someone who trained NLP models prior to LLMs, it's just BERT with more data." ([source](https://news.ycombinator.com/item?id=49765997)) - Confidence scores add little (3 comments): "Jev's "confidence" output is fake/redundant - it's just a formula applied to probabilities, it conveys no additional information." ([source](https://news.ycombinator.com/item?id=49803546)) "it was 0-for-3 at giving yes/no probabilities that were particularly close to the obviously correct no-information answers" ([source](https://news.ycombinator.com/item?id=49839995)) "my conclusion was that the speed and the confidence score come from reading out a single token, not from Jev itself" ([source](https://news.ycombinator.com/item?id=49981222)) - Closed source, data retention concerns (4 comments): "specifically calling this kind of model a breakthrough, without crediting previous art, and keeping everything closed source" ([source](https://news.ycombinator.com/item?id=49766884)) "Jev is much better than the traditional ML crowd gives it credit for, but my enthusiasm hits a wall when it comes to their data" ([source](https://news.ycombinator.com/item?id=49788645)) "They know it's not. The company actually has or had public statements that they didn't like public benchmarks for comparison." ([source](https://news.ycombinator.com/item?id=49851633)) ## Alternatives hackers recommend - [Laya](https://hackerapproved.com/p/laya.md): 44 recommend - [Cloudflare Clef](https://hackerapproved.com/p/clef.md): 21 recommend - [OpenJev](https://hackerapproved.com/p/openjev.md): 21 recommend - [Typesafe AI](https://hackerapproved.com/p/typesafe-ai.md): 16 recommend - [Strands Decider 2B](https://hackerapproved.com/p/strands-decider-2b.md): 7 recommend ## Cite as Hacker Approved, "Jev by Typesafe AI", https://hackerapproved.com/p/jev/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).