The link rgbrgb posted is a good overview. The best open ones are close to Jev now, but they're big models. And I agree, if you have an eval set for a fixed task, a trained classifier is the better choice.
Their marketing language is misleading. They must still use some transformer language model backbone to encode the text input (BERT or decoder-only LLM). The biggest difference is the output, instead of auto-regressively generating tokens, they produce probabilities over a bounded set of decisions (more flexible classification).
that laya is even a thing is further evidence, people took that author at face value, the paper contents are incomplete and describe something that does not sound like Jev at all
this was the period of arxiv history that led to the new vouching system, laya author contributed to that imo
My understanding is that Laya (or whatever it was called in 2025) was yet another fine-tuned classifier, not a general purpose one.
That said, Typesafe false marketing caused Laya to fit perfectly into pretty much every advantage that they are claiming: "system one decision model", cheap, fast, no hallucinations, structured, confidence output, parallel, calibrated. Their BS is their own demise.
I think Laya's author genuinely bought their BS and thinks he built the same thing. Unlike Typesafe, I don't think he's intentionally misleading people.
The only unique thing about Jev is that it's a general purpose classifier. Funny enough, they were so busy spreading marketing bullshit that they forgot to mention the only real thing that makes Jev unique.
Laya author is spitting more BS than Typesafe, the (incomplete) papers are nothing like Jev, they use RAG and azure hosted services for calculating embeddings, with an orchestrator. Jev is just a model, Laya was put together after Jev, almost certainly based on what the author learned from Typesafe, and then backported "his" idea
I suspect most people only read the blog post, and thought it was great how a VC company "stole" an idea and was "outdone" by a rando... without actually checking the facts. Confirmational reading bias, we live in a post-truth world with dysfunction media ecosystem
To be fair, he's just asking how to get customers. And the post is 2 days before Jev's launch date? I don't think he's trying to sell Laya there (though he probably will at this point).
I am fairly confident if Jev-style decision models are seen as prominent (which, they seem to be), Ollama will support them. Surprised the team hasn't implemented this already.
Depends on the model. The small ones I support today are well below Jev on harder queries, but fine for simple, well-defined questions. The open models that get close to Jev are bigger, and I'm adding support for those next.
Has anyone actually seen better or the same results with Laya compared to Jev? From my experience, Laya performs significantly worse. It's less confident and often makes wrong decisions with more complex queries.
Developer here. You're right, Laya is a lot weaker than Jev, especially on harder queries. It's a small model, so it's fast, but that's the trade-off. The open models that get close to Jev are much bigger, and running those is what I'm working on next.
Probably Kev and/or the decider models. Kev is trained on one of the 4B qwen models, similar for decider but it ranges from 0.8B through to the 35B-A3B model so far I believe.
Yes. JEV generalizes better because they probably have an enormous corpus and trained on it for a long time. Laya's out of the box model is much weaker. However, in the age of LLM's it's incredibly easy and cheap to generate large datasets to fine tune laya for your task, and the training loop is pretty quick and cheap too.
It's so easy that I question why I would ever pay for JEV when eventually I'll have done enough random things that I will also have a large corpus and likely a general model as well.
Isn't the point of Jev that it generalises better?
It's a fast classifier you can use out-the-box, ~1.5bn tokens is about $40 (I've been hammering it)
It just works ... a whole bunch of low-level/low-importance workflow stuff that was getting farmed out to small/fast LLM models now has a competitive alternative ... and bits that hadn't even been considered to go into some external descision/classifier service can be tested/deployed at ~$0.00003/req
I don't get this wall of negativity on it, it's genuinely innovative/useful tech ... would expect HN to be more positive, regardless of whether it's the absolute best execution
It really does just work. And it works so well I already integrated it into my product. Saves me about 75% of costs for the section its working in, which isn't a small amount. I see a lot of negativity and I don't really get it either. Its so cheap and so fast, why not give it a try?
I didn't see any negativity in the post you replied to.
I think the point being made is that Jev is great but it has no competitive moat, and open source versions will very soon catch up if their secret sauce is just synthetic data.
I think your point is valid but many are annoyed that it is presented as groundbreaking, revolutionary, novel frontier tech when it is a known classification system. It’s the hype that feels undeserved. Honestly it was one of the best marketing campaigns I’ve seen.
You should try building something with it, the hype is what it is, but the model is crazy useful.
I'm building a woodworking app and I've managed to create an autopilot that can take a simple instruction ("get me 5 2x4s", "cut the middle 2x4 into 4 equal pieces", "move the 2x4 3 feet left") and the action instantly happens with next to no lag. There is already an llm but now it can share an intent, and the geometry system shows jev the various actions and jev chooses the action that gets it closer to the goal until it has found a state that matches the intent or gives up. The result is the llm can think "higher level" and let the cheap fast model grind out the options in a relative blink of the eye, without the 30s of reasoning the llm would have done about the various operations it could try.
Nothing yet. Unfortunately it sometimes feels like our industry has been overrun by grifters and chancers.
I’m sure this has been a gradual and long decline. Maybe it even started with the dot com boom and accelerated with crypto. With AI it seems to have got worse.
one day, perhaps people will click through to the laya author's arxiv paper content and the why may become clearer, you won't have to read it, a skim will suffice
This is just anecdotal and I might be doing it wrong but I made jev and laya versions of a simple semantic grep tool (https://github.com/lgastako/jevplay) and played with them a bit, and at first it seemed like laya was comparable (eg on queries like "this is a mans name" or "this is a womans name" on names.txt) but the more I played with it, eg. "this is a vegetable" on foods.txt the further the gap widened in favor of jev. Then I started trying variations of the query eg simply "mans name" and for the most part laya just fell apart and didn't return anything useful for a lot of stuff. I was hoping to find that laya was competitive because it's much faster to have the model running locally but it's just not, yet.
Calibrated probability across multi task with zero shot I guess. A reranker is single task and tuning it make it even more narrow. And I guess some piping to make multiclass efficient since you cannot mask logprob for independent questions in the same output space without throwing calibration away.
difference between an instruct based re-ranker and laya/jev I just don't see it
Main difference is that laya/jev/et-al give you a zero-shot classifier that requires no training. You can prompt engineer your way to a quick fairly reliable cheap enough decision engine that you can use to iterate quickly (by prompt engineering).
Right now a lot of people are doing this with LLMs and it's too slow and expensive.
Imo the right iterative approach to productionizing these systems is something like:
1. Build it with an LLM. Iterate on the prompt
2. Start building a real-world dataset
3. When the prompt works, turn it into a clear rubric for Jev or similar
4. Keep iterating until desired accuracy achieved
5. Use the real-world evals you've built to train a custom classifier fine-tuned to your needs
You now have a system that has produced useful results in production from the very beginning and by the end it's a reliable super cheap classifier that can make thousands of decisions per second.
I don’t think that’s it. I sincerely doubt most developers are doing side by side comparisons of calibration quality.
OpenAI has a section on their embeddings model api page for zero shot classification. Of course you can choose an open weights embedding too if you’d like.
I think Jev wins on marketing and convenience. Most SWEs don’t want to talk about embeddings, cosine similarity, or precision/recall tradeoffs. They want something which plausibly works and is easy to use.
Yes it turns all that work of building a classifier into an api call. This is hugely valuable for prototyping and while you iterate on what the product should even do.
I'm not sure what this means for AI startups if their innovations can be copied by OSS so quickly (what, like 2 weeks?). There's "consumer surplus" for everyone, to borrow an economic concept. But we do ideally want some of the surplus to flow to the innovator, too. I know there were precursors, but that's fine - it's hard to have a totally novel idea in such a popular field. I don't know what the end game is for TypeSafe - they'd need to demonstrate perpetually better results, or compete in another axis: UX, support, custom solutions, etc. So much of the time, someone proving a concept, or it simply getting enough publicity, is enough for a "Cambrian explosion" of follow-ups and copies. Famously, that was true for "Attention is All You Need", and the general idea of "next-token prediction" being so powerful.
We've stumbled into general differentiable models..
Presumably the training recipe and training dataset itself cannot be easily copied in a week or two. So if they want to shut down these competitor models they need to make it obvious how they are better than them.
It a paradox when the article is claiming the prior art is absurd, but then goes on to analyse the one side and compare it to another for which most of the values (except scaling the concept) are unknown. And even for scaling, it uses the first, pre-laya instance to judge the limited schema, while overlooking that Laya is just doing this scaling. Important to note that prior art is not having built the exact same thing.
Because what they did is kinda trivial. Its basically like the Dropbox comment really[0], except here you don't need petabytes of storage and infinite VC pockets.
After chatgpt everything in AI mostly became LLMs and building wrappers around them. It's like people forgot how to do ML.
To those of us who actually trained models back in the day, its kind of cute to see people wowed by a classifier. Yes, this is 0 shot and doesn't need training (most people wanting this would've used structured output, this is cool because it's cheaper and faster). But anyone with basic ML knowledge could've built this in a few hours.
The question is mostly why wasn't this productized. And it's interesting indeed that it took this long to become a finished product.
I think GP has a point, though. BERT models have existed for a while but OpenAI made classification via LLM convenient and accessible for regular developers.
People didn’t know they wanted classifiers until OpenAI gave them a taste.
The answer for why it wasn't productized might just be pretty straightforward.
LLMs still are better than Jev at the task, just across the board slower.
Anyone who had a reason to try this already tried it (ads/recommendations) - back in 2023/2024 during the first fine tuning wave and it was accurately determined that it was not worth the effort, the results were more bogus than just using CoT, so frankly parallelism meant nothing if bogus * parallel = bogus.
So thrown into the dumpster and nobody really cared to revisit because it was already tried.
Pretty much sometime between then and now it somehow became the state where the tradeoff makes sense now.
The moat is the RL synthetic data pipeline they set up to train jev. Open sourcing that would be the coup, not the model architecture and training scripts, which are trivial.
Today we might need to evaluate “innovation” in a new standard, and have a different expectation for what innovator would be awarded. Getting public attention in such an era where innovation happens every a few days could’ve already been something precious. And that attention would allow TypeSafe to be heard easily next time. Like OpenAI, Anthropic, or any others, they launch frequently but still each time they launch something new, that would hit headlines. I think that’s the “surplus” flown to innovators today.
I think the idea of a "feature startup" is dead. What used to be a niche subscription business is now an individual Epic level of work. The smallest viable business becomes what two or three years ago was a mid tier enterprise. It is no longer "look at this tool I maintain", but "we take this specific approach using these hundreds of tools merged together to solve a problem in a specific way that nobody is going to compete with. Not because they can't compete if they wanted to, but that the competitions approach diverges in fifty different chosen ways that they are targeting a different market segment essentially."
I adhere to the idea that this is software's "Tower of Babel" moment where everyone just fundamentally ships things in completely diverging architectures, because creating a ground up architecture is no longer something that needs to be avoided for an economically viable business mode that in the past two decades would have otherwise incentivized people into industry standards. In a world where "taste" is the focus, single ingredients in the recipe aren't enough.
Laya is pretty easy to set up on its own without ollaya. I just did that and replaced my current jev API usage to laya running on a GTX 970 with 4GB of vram.
Very small context window, but for some existing small llm work I was doing, it was a drop-in replacement and it makes me happy I can get use out of old hardware I have running.
It would be good to list 1) zero-shot accuracy and 2) latency on the models page . The LLM-based models' latency is probably much higher than the BERT approaches I would assume.
Also curious, it seems from looking at the accuracy scores you gave that it seems to be NLI > Gliclass > Laya (for Bert types)? Why do you seem to feature/recommend Laya more - is Laya better in some way?
I installed it, I tried the examples, it works.... But forgive my lack of imagination... what is this useful for?
Like, their example is of classification for a support interface.... `refund_requested`. Pretty convenient bool given the example is about a refund- what if 99% of submissions don't ask about a refund? Also, is that user not a `churn_risk`? What could possibly qualify as a churn risk if not a user asking for a refund?
https://ollaya.dev/library/laya The examples suffer the same problem of why I'd prefer to use a string column vs an enum. Changing an enum means you need to update the db, using a string you can do whatever.
I'm not trying to be negative, I genuinely want to know about some practical examples (that don't require tons of backwards maintenance).
I have a lot of semi-practical examples of how you can use this model wrapped in unix-ish tools - https://github.com/aurorainfra/grev (readme links to docs of each tool with some more or less practical examples)
Really I think "smart grep" is a pretty good one ('look for an error looking vaguely like this'). Also I think sql-based shell history + decision model is quite good to make the last 'which one of those choices is best fit given users past few commands' etc.
But with Jev you're just paying for input (prefill) which is really fast, and in case of Jev specifically costs 50% of Deepseek V4.1 Flash (which has famously really cheap input token pricing).
I put 250MB / 1M lines of logs through Grev and it cost ~$10USD, DSv4.1 would be at least 10x that and much, much, much slower. With Jev/Grev that 1M requests took 10 mins
Edit: completely misread your question - yeah you could finetune specialized models to do that, probably based on some decent pretrained llm base, that is true for roughly any Jev-shaped problem. Do you want to bother doing that, also having to deal with having to host a zoo of specialized models?
Ok, those are pretty decent examples, and clears up the utility a bit: speed and tokens. Some of it's still a bit iffy (e.g. `cutv 'email address' 'phone number' < examples/users.csv`, csv is already in columns), but I can see using it for some niche queries. Neat tool.
I very much appreciate your to-the-point, non-vibed README as well, ty for that.
Yeah, speed is the one, I believe the default TypeSafe API quota is 1.5-2k queries per second (batched in bigger requests).
On the readme I'm so sorry to tell you that, but it's 100% written by Opus 5.5 with zero "pretty please don't write slop" prompting, it's just how slop is going to look like from now on. I've been writing code for 15 years or sth like that and the code is also what I'd call pretty reasonable..
It has some jargon hallmarks, which I noticed, but vibed or not, it's a massive improvement on other repos. Maybe it's because it's only a few commits so far... perhaps if you were to vibe 100 more commits it would devolve. Or maybe 5.5 really did improve (doubt it, still sounds like an asshole for me). But idk.
I still think that `churn_risk` above is incorrect and unacceptable (perhaps there are sensible fixes, but saying "no churn risk" about a refund, in a leading example on their homepage, flabbergasting).
But if that were solved, I could see giving ollaya/grev to LLMs themselves, giving LLMs their own massive token-saver.
Is for when you want an AI to make a decision. If you have been using gpt or claude or open source models for that, than it’s a way cheaper alternative.
And if you have not been, it’s for when you have to extract the context from text. When you have numbers or fixed options, it’s just a matter of code.
So if you find yourself having to decide if a given user comment is a refund_request, that’s for that.
It’s not perfect, you still have to fine-tune (or calibrate) using examples you have (and keep those examples updated over time). But it’s way better than trying to parse text with regexes.
If you don't want to spend a lot and want low latency, e.g. home automation. "It's cold and dark in here, do something about it", it will then turn on the lights and heater nearly instantly.
For everyone dismissing Jev's innovation as being trivial, no it's not.
It is definitely not the MNIST classifier you had trained in 2019.
The difference is that you only train it once and the modern LLM machinery sort of takes care of that with large contexts.
It's great that Jev proved this is a viable product. I'd expect a great many research innovations coming from making this work better/faster/cheaper, and around interfacing modern agents with it.
Hah you’re probably dating yourself. Keras came out in 2015 and that was one of the early examples with Theano backend. You could train MNIST since 2015 pretty straight forward. But comparing Jev to image classification is an unfaithful argument. Comparing it to ELmo or BERT is analogously better.
Are there many models that are comparable to Jev for generic decision making?
Smarter move if you have an eval set is to just train a classifier and call it a day.
there's this thing with a bunch of similar models https://huggingface.co/spaces/multimodalart/jev-decision-ind...
top open one is trained by perplexity cto for $3k, kinda cool https://x.com/denisyarats/status/2102252088067850507
<<<"i was curious to see if i could train a competitive Jev-like model completely autonomously with a swarm of agents using our internal system."
Bro is writing off the H200 lol
On a sidenote I really can't stand the term "swarm" and definately plays into AI doomerism.
The link rgbrgb posted is a good overview. The best open ones are close to Jev now, but they're big models. And I agree, if you have an eval set for a fixed task, a trained classifier is the better choice.
Cool... but this does seem undermined by the fact that Ollama can add support for decision models at any time.
Fair, and I'd be happy if they did. Ollaya uses the same API as Jev, so your code isn't tied to it either way
and that ollama is go-llama and not rust, so it's not really the ollama of anything
great project for empowering open-source alternatives.
open-source is the only way for safe AI development. whoever doesn’t share the weights/code will lag behind.
Thanks!
Fair point, that example is basically classification. I'll change it to something that looks more like a real decision.
text classification is equivalente to decision. This is exactly the same thing Jev does.
Jev does it more efficiently because it doesn't use an LLM https://typesafe.ai/blog/introducing-system-one-models-and-j...
Their marketing language is misleading. They must still use some transformer language model backbone to encode the text input (BERT or decoder-only LLM). The biggest difference is the output, instead of auto-regressively generating tokens, they produce probabilities over a bounded set of decisions (more flexible classification).
It is not. In a benchmark with actual decisions - navigation, traffic, waypoints - laya does only slightly better than a small classifier.
If it has four legs, a tail and barks why not call it a dog?
Because this specific dog only barks in structured text
This dog only barks when given biscuits
"Decision model" is just marketing jargon.
decision model = classifier
system one model = small non-reasoning LLM
noul = boolean
confidence = f(probabilities)
It's sad to see how gullible engineers are today.
that laya is even a thing is further evidence, people took that author at face value, the paper contents are incomplete and describe something that does not sound like Jev at all
this was the period of arxiv history that led to the new vouching system, laya author contributed to that imo
My understanding is that Laya (or whatever it was called in 2025) was yet another fine-tuned classifier, not a general purpose one.
That said, Typesafe false marketing caused Laya to fit perfectly into pretty much every advantage that they are claiming: "system one decision model", cheap, fast, no hallucinations, structured, confidence output, parallel, calibrated. Their BS is their own demise.
I think Laya's author genuinely bought their BS and thinks he built the same thing. Unlike Typesafe, I don't think he's intentionally misleading people.
The only unique thing about Jev is that it's a general purpose classifier. Funny enough, they were so busy spreading marketing bullshit that they forgot to mention the only real thing that makes Jev unique.
Laya author is spitting more BS than Typesafe, the (incomplete) papers are nothing like Jev, they use RAG and azure hosted services for calculating embeddings, with an orchestrator. Jev is just a model, Laya was put together after Jev, almost certainly based on what the author learned from Typesafe, and then backported "his" idea
I suspect most people only read the blog post, and thought it was great how a VC company "stole" an idea and was "outdone" by a rando... without actually checking the facts. Confirmational reading bias, we live in a post-truth world with dysfunction media ecosystem
I think you're right about Laya (and confirmation bias).
But like you said, at the end of the day he's just a rando.
He's not asking for $40m, not saying "I made ChatGPT, but i hate it, so I built the next big thing". Not claiming to co-invent RLHF.
Laya is just noise. Jev's bullshit affects me today - I see people injecting it into the codebases where it has no place.
Just how to "make fkn $500k ARR fast?"
https://news.ycombinator.com/item?id=49674396
too much LI/Xitter influencer consumption
That's really funny, nice find.
To be fair, he's just asking how to get customers. And the post is 2 days before Jev's launch date? I don't think he's trying to sell Laya there (though he probably will at this point).
I am fairly confident if Jev-style decision models are seen as prominent (which, they seem to be), Ollama will support them. Surprised the team hasn't implemented this already.
Sounds good on latency but how is its actual decision quality vs. Jev?
Depends on the model. The small ones I support today are well below Jev on harder queries, but fine for simple, well-defined questions. The open models that get close to Jev are bigger, and I'm adding support for those next.
Has anyone actually seen better or the same results with Laya compared to Jev? From my experience, Laya performs significantly worse. It's less confident and often makes wrong decisions with more complex queries.
Developer here. You're right, Laya is a lot weaker than Jev, especially on harder queries. It's a small model, so it's fast, but that's the trade-off. The open models that get close to Jev are much bigger, and running those is what I'm working on next.
What are the models? I am super curious in these as well
Probably Kev and/or the decider models. Kev is trained on one of the 4B qwen models, similar for decider but it ranges from 0.8B through to the 35B-A3B model so far I believe.
It doesn’t to be a ton bigger, 16k and reliable 8k would be a godsend. (I run at 2k)
Yes. JEV generalizes better because they probably have an enormous corpus and trained on it for a long time. Laya's out of the box model is much weaker. However, in the age of LLM's it's incredibly easy and cheap to generate large datasets to fine tune laya for your task, and the training loop is pretty quick and cheap too.
It's so easy that I question why I would ever pay for JEV when eventually I'll have done enough random things that I will also have a large corpus and likely a general model as well.
Isn't the point of Jev that it generalises better?
It's a fast classifier you can use out-the-box, ~1.5bn tokens is about $40 (I've been hammering it)
It just works ... a whole bunch of low-level/low-importance workflow stuff that was getting farmed out to small/fast LLM models now has a competitive alternative ... and bits that hadn't even been considered to go into some external descision/classifier service can be tested/deployed at ~$0.00003/req
I don't get this wall of negativity on it, it's genuinely innovative/useful tech ... would expect HN to be more positive, regardless of whether it's the absolute best execution
It really does just work. And it works so well I already integrated it into my product. Saves me about 75% of costs for the section its working in, which isn't a small amount. I see a lot of negativity and I don't really get it either. Its so cheap and so fast, why not give it a try?
I think it’s the infamous Dropbox reaction - anyone can wrap an FTP server, where the innovation?
Starting from a business POV one should inflate terminology, hack together an MVP, and see if the market demands it before doing hardcore R&D.
But starting from technical/craftsman POV all you see is a hack and a lot of big words, so it’s easy to become jaded.
I didn't see any negativity in the post you replied to.
I think the point being made is that Jev is great but it has no competitive moat, and open source versions will very soon catch up if their secret sauce is just synthetic data.
(Whether or not that is true, I don't know.)
It's probably true because even for Frontier LLM models there are many competitors now.
I think your point is valid but many are annoyed that it is presented as groundbreaking, revolutionary, novel frontier tech when it is a known classification system. It’s the hype that feels undeserved. Honestly it was one of the best marketing campaigns I’ve seen.
If you don't mind me asking, what are you using it for?
I've been unable to find a good use case for now.
You should try building something with it, the hype is what it is, but the model is crazy useful.
I'm building a woodworking app and I've managed to create an autopilot that can take a simple instruction ("get me 5 2x4s", "cut the middle 2x4 into 4 equal pieces", "move the 2x4 3 feet left") and the action instantly happens with next to no lag. There is already an llm but now it can share an intent, and the geometry system shows jev the various actions and jev chooses the action that gets it closer to the goal until it has found a state that matches the intent or gives up. The result is the llm can think "higher level" and let the cheap fast model grind out the options in a relative blink of the eye, without the 30s of reasoning the llm would have done about the various operations it could try.
If you're so inclined it would be easy, fast and cheap to distill Jev for your task.
Nothing yet. Unfortunately it sometimes feels like our industry has been overrun by grifters and chancers.
I’m sure this has been a gradual and long decline. Maybe it even started with the dot com boom and accelerated with crypto. With AI it seems to have got worse.
I've been following jevbench twice a day for the past week and that's been a lot of fun. Latest update:
Rank System Score Public / sealed accuracy Evidence
1 decider-4b v2 64.13 83.5% / 34.7% Evaluator-run, offline
2 Jev 1.13 63.29 86.6% / 36.7% Evaluator-run API
3 JevK5 v0.2 62.04 85.3% / 33.1% Evaluator-run
4 Cygnet 12B 61.76 87.9% / 33.8% Evaluator-run, offline
5 Hopper 59.43 82.3% / 34.1% Evaluator-run
28 Kev 4B 36.14 66.2% / 22.4% Evaluator-run
41 Laya 421M 30.25 58.4% / 30.8% Evaluator-run
https://benchmarkheaven.com/jev-models
What the best way to see how a homegrown version compares?
Amazing - was looking for some benchmarks around this earlier
one day, perhaps people will click through to the laya author's arxiv paper content and the why may become clearer, you won't have to read it, a skim will suffice
In my experience it's not close and the benchmarks I've seen don't reflect my experience at all.
But I'm guessing people will find the right training regime and data mix soon to close the gap.
But big things I see are instability and inaccuracy - like pick a random problem.
This is just anecdotal and I might be doing it wrong but I made jev and laya versions of a simple semantic grep tool (https://github.com/lgastako/jevplay) and played with them a bit, and at first it seemed like laya was comparable (eg on queries like "this is a mans name" or "this is a womans name" on names.txt) but the more I played with it, eg. "this is a vegetable" on foods.txt the further the gap widened in favor of jev. Then I started trying variations of the query eg simply "mans name" and for the most part laya just fell apart and didn't return anything useful for a lot of stuff. I was hoping to find that laya was competitive because it's much faster to have the model running locally but it's just not, yet.
It would be really cool to have LLMs and System One in a single tool - in this case, if Ollama implemented it.
next vLLM release will have this
if you use gateways, GoModel support the S1 endpoints, my favorite feature is the virtual models, stable name, I can swap out the backing model(s)
https://gomodel.enterpilot.io/docs/getting-started/quickstar...
(the "kev" in the docs is my fault, I should have said Jev / System1 in my feature request)
I have also tried this and its really awesome
Hey Claude, make ollama for Jev like models. Make no mistakes /s
hey Claude, download and run vllm nightly for me
(already merged)
GoModel (gateway) already supports Jev like endpoints too
https://gomodel.enterpilot.io/docs/providers/jev
Guys I have a real q, what is the difference between an instruct based re-ranker and laya/jev I just don't see it.
Edit: One is that jev/laya are tuned to have better probabilities, but a reranker can be fine tuned to do that as well. And jev/laya use RLCD?
Calibrated probability across multi task with zero shot I guess. A reranker is single task and tuning it make it even more narrow. And I guess some piping to make multiclass efficient since you cannot mask logprob for independent questions in the same output space without throwing calibration away.
Main difference is that laya/jev/et-al give you a zero-shot classifier that requires no training. You can prompt engineer your way to a quick fairly reliable cheap enough decision engine that you can use to iterate quickly (by prompt engineering).
Right now a lot of people are doing this with LLMs and it's too slow and expensive.
Imo the right iterative approach to productionizing these systems is something like:
You now have a system that has produced useful results in production from the very beginning and by the end it's a reliable super cheap classifier that can make thousands of decisions per second.
I don’t think that’s it. I sincerely doubt most developers are doing side by side comparisons of calibration quality.
OpenAI has a section on their embeddings model api page for zero shot classification. Of course you can choose an open weights embedding too if you’d like.
https://developers.openai.com/cookbook/examples/zero-shot_cl...
I think Jev wins on marketing and convenience. Most SWEs don’t want to talk about embeddings, cosine similarity, or precision/recall tradeoffs. They want something which plausibly works and is easy to use.
Yes it turns all that work of building a classifier into an api call. This is hugely valuable for prototyping and while you iterate on what the product should even do.
Jev's value becomes more apparent when the task is a moving target. eg an auto-mode classifier.
Does anyone know what laya multi lang is faster than laya en? I would have thought focusing on a single language would be faster.
This inference engine is soooo much faster btw: https://github.com/tamnd/kime
FAQ[1] says:
[1]: https://ollaya.dev/docs/faq
Would be great if you supported CUDA 12; I don't feel like paying $15K to upgrade my GPU right now
wait another week or so for vLLM's next release
I'm not sure what this means for AI startups if their innovations can be copied by OSS so quickly (what, like 2 weeks?). There's "consumer surplus" for everyone, to borrow an economic concept. But we do ideally want some of the surplus to flow to the innovator, too. I know there were precursors, but that's fine - it's hard to have a totally novel idea in such a popular field. I don't know what the end game is for TypeSafe - they'd need to demonstrate perpetually better results, or compete in another axis: UX, support, custom solutions, etc. So much of the time, someone proving a concept, or it simply getting enough publicity, is enough for a "Cambrian explosion" of follow-ups and copies. Famously, that was true for "Attention is All You Need", and the general idea of "next-token prediction" being so powerful.
We've stumbled into general differentiable models..
Presumably the training recipe and training dataset itself cannot be easily copied in a week or two. So if they want to shut down these competitor models they need to make it obvious how they are better than them.
Are you saying laya copied from jev, and released in two weeks? If so I don’t thinks it’s quite as simple a story as that. https://xtxinversexty.com/layas-prior-art-claim-is-absurd/
It a paradox when the article is claiming the prior art is absurd, but then goes on to analyse the one side and compare it to another for which most of the values (except scaling the concept) are unknown. And even for scaling, it uses the first, pre-laya instance to judge the limited schema, while overlooking that Laya is just doing this scaling. Important to note that prior art is not having built the exact same thing.
Because what they did is kinda trivial. Its basically like the Dropbox comment really[0], except here you don't need petabytes of storage and infinite VC pockets.
After chatgpt everything in AI mostly became LLMs and building wrappers around them. It's like people forgot how to do ML.
To those of us who actually trained models back in the day, its kind of cute to see people wowed by a classifier. Yes, this is 0 shot and doesn't need training (most people wanting this would've used structured output, this is cool because it's cheaper and faster). But anyone with basic ML knowledge could've built this in a few hours.
The question is mostly why wasn't this productized. And it's interesting indeed that it took this long to become a finished product.
[0] https://news.ycombinator.com/item?id=9224
Probably because doing it wrong (using an llm in place of a classifier) is more profitable? (For the people selling inference.)
It wasn't until recently that demand for classifiers at this scale existed. Jev exists because LLM's exist. Without them it wouldn't be (as) useful
That's not at all true.
There have been tons of applications for this. People were using earlier LLMs like BERT for classifiers long before LLMs became viable chatbots.
I think GP has a point, though. BERT models have existed for a while but OpenAI made classification via LLM convenient and accessible for regular developers.
People didn’t know they wanted classifiers until OpenAI gave them a taste.
The answer for why it wasn't productized might just be pretty straightforward.
LLMs still are better than Jev at the task, just across the board slower.
Anyone who had a reason to try this already tried it (ads/recommendations) - back in 2023/2024 during the first fine tuning wave and it was accurately determined that it was not worth the effort, the results were more bogus than just using CoT, so frankly parallelism meant nothing if bogus * parallel = bogus.
So thrown into the dumpster and nobody really cared to revisit because it was already tried.
Pretty much sometime between then and now it somehow became the state where the tradeoff makes sense now.
The moat is the RL synthetic data pipeline they set up to train jev. Open sourcing that would be the coup, not the model architecture and training scripts, which are trivial.
Today we might need to evaluate “innovation” in a new standard, and have a different expectation for what innovator would be awarded. Getting public attention in such an era where innovation happens every a few days could’ve already been something precious. And that attention would allow TypeSafe to be heard easily next time. Like OpenAI, Anthropic, or any others, they launch frequently but still each time they launch something new, that would hit headlines. I think that’s the “surplus” flown to innovators today.
I think the idea of a "feature startup" is dead. What used to be a niche subscription business is now an individual Epic level of work. The smallest viable business becomes what two or three years ago was a mid tier enterprise. It is no longer "look at this tool I maintain", but "we take this specific approach using these hundreds of tools merged together to solve a problem in a specific way that nobody is going to compete with. Not because they can't compete if they wanted to, but that the competitions approach diverges in fifty different chosen ways that they are targeting a different market segment essentially."
I adhere to the idea that this is software's "Tower of Babel" moment where everyone just fundamentally ships things in completely diverging architectures, because creating a ground up architecture is no longer something that needs to be avoided for an economically viable business mode that in the past two decades would have otherwise incentivized people into industry standards. In a world where "taste" is the focus, single ingredients in the recipe aren't enough.
This is precisely why we need strict government regulation of open weight models. :wink:
Laya is pretty easy to set up on its own without ollaya. I just did that and replaced my current jev API usage to laya running on a GTX 970 with 4GB of vram.
Very small context window, but for some existing small llm work I was doing, it was a drop-in replacement and it makes me happy I can get use out of old hardware I have running.
It would be good to list 1) zero-shot accuracy and 2) latency on the models page . The LLM-based models' latency is probably much higher than the BERT approaches I would assume.
Also curious, it seems from looking at the accuracy scores you gave that it seems to be NLI > Gliclass > Laya (for Bert types)? Why do you seem to feature/recommend Laya more - is Laya better in some way?
Nice work! Making open models easier to run locally is valuable on its own. Keeping the API compatible with Jev is a thoughtful touch, too.
I installed it, I tried the examples, it works.... But forgive my lack of imagination... what is this useful for?
Like, their example is of classification for a support interface.... `refund_requested`. Pretty convenient bool given the example is about a refund- what if 99% of submissions don't ask about a refund? Also, is that user not a `churn_risk`? What could possibly qualify as a churn risk if not a user asking for a refund?
https://ollaya.dev/library/laya The examples suffer the same problem of why I'd prefer to use a string column vs an enum. Changing an enum means you need to update the db, using a string you can do whatever.
I'm not trying to be negative, I genuinely want to know about some practical examples (that don't require tons of backwards maintenance).
I have a lot of semi-practical examples of how you can use this model wrapped in unix-ish tools - https://github.com/aurorainfra/grev (readme links to docs of each tool with some more or less practical examples)
Really I think "smart grep" is a pretty good one ('look for an error looking vaguely like this'). Also I think sql-based shell history + decision model is quite good to make the last 'which one of those choices is best fit given users past few commands' etc.
Isn't it better to use an LLM to train modernbert or xgboost et al?
It is /possible/ to use an LLM.
But with Jev you're just paying for input (prefill) which is really fast, and in case of Jev specifically costs 50% of Deepseek V4.1 Flash (which has famously really cheap input token pricing).
I put 250MB / 1M lines of logs through Grev and it cost ~$10USD, DSv4.1 would be at least 10x that and much, much, much slower. With Jev/Grev that 1M requests took 10 mins
Edit: completely misread your question - yeah you could finetune specialized models to do that, probably based on some decent pretrained llm base, that is true for roughly any Jev-shaped problem. Do you want to bother doing that, also having to deal with having to host a zoo of specialized models?
Ok, those are pretty decent examples, and clears up the utility a bit: speed and tokens. Some of it's still a bit iffy (e.g. `cutv 'email address' 'phone number' < examples/users.csv`, csv is already in columns), but I can see using it for some niche queries. Neat tool.
I very much appreciate your to-the-point, non-vibed README as well, ty for that.
Yeah, speed is the one, I believe the default TypeSafe API quota is 1.5-2k queries per second (batched in bigger requests).
On the readme I'm so sorry to tell you that, but it's 100% written by Opus 5.5 with zero "pretty please don't write slop" prompting, it's just how slop is going to look like from now on. I've been writing code for 15 years or sth like that and the code is also what I'd call pretty reasonable..
It has some jargon hallmarks, which I noticed, but vibed or not, it's a massive improvement on other repos. Maybe it's because it's only a few commits so far... perhaps if you were to vibe 100 more commits it would devolve. Or maybe 5.5 really did improve (doubt it, still sounds like an asshole for me). But idk.
I still think that `churn_risk` above is incorrect and unacceptable (perhaps there are sensible fixes, but saying "no churn risk" about a refund, in a leading example on their homepage, flabbergasting).
But if that were solved, I could see giving ollaya/grev to LLMs themselves, giving LLMs their own massive token-saver.
Is for when you want an AI to make a decision. If you have been using gpt or claude or open source models for that, than it’s a way cheaper alternative.
And if you have not been, it’s for when you have to extract the context from text. When you have numbers or fixed options, it’s just a matter of code.
So if you find yourself having to decide if a given user comment is a refund_request, that’s for that.
It’s not perfect, you still have to fine-tune (or calibrate) using examples you have (and keep those examples updated over time). But it’s way better than trying to parse text with regexes.
If you don't want to spend a lot and want low latency, e.g. home automation. "It's cold and dark in here, do something about it", it will then turn on the lights and heater nearly instantly.
sounds reasonable but doesn’t feel natural in a way that I can’t explain easily
Why do you need another model-type specific Ollama? Can't Ollama be made to support these models?
Ollama is for large language models, so this is for large language... yodels?
For everyone dismissing Jev's innovation as being trivial, no it's not.
It is definitely not the MNIST classifier you had trained in 2019.
The difference is that you only train it once and the modern LLM machinery sort of takes care of that with large contexts.
It's great that Jev proved this is a viable product. I'd expect a great many research innovations coming from making this work better/faster/cheaper, and around interfacing modern agents with it.
Hah you’re probably dating yourself. Keras came out in 2015 and that was one of the early examples with Theano backend. You could train MNIST since 2015 pretty straight forward. But comparing Jev to image classification is an unfaithful argument. Comparing it to ELmo or BERT is analogously better.
You missed the point - you had to train BERT or anything similar to get useful results out of it.
Now all you need is to give it more context along with your query.