Looking at my Nissan Leaf it is entirely possible to build a reliable car with real buttons and knobs everywhere, and thats the problem for the industry. The used market is flooded with perfectly good Leafs, dirt cheap, the lowest yearly maintenance vehicle on CarEdge.
Even ICE cars have gotten a little too good over the last couple of decades, I see plenty of early 2000s models driving around today looking perfectly fine.
How are these corporations supposed to maximise profit if they keep churning out perfectly sensible vehicles?
Lmao if that were true, everyone would be buying them. We aren’t, because they suck. It’s a decent car if you just need something cheap, but cheap doesn’t mean good. A raspberry pi isn’t better than a M5 just because it’s cheap.
Yeah, a lot of people on Hacker News are asking for a bare-bones EV, but that market isn't big enough for US or Western carmakers to justify building one. Still, I'd bet that over the next few years, as batteries keep getting cheaper, we'll see plenty of these bare-bones cars in Africa and Asia. They're already building electric tuk-tuks, and as solar power electrifies more of the world, we'll get bare-bones EVs.
Toyota's EVs share most of the same components as their non-EVs, like door handles and screens. They have become the fourth best selling EVs behind the Teslas and the Hyundai Ioniq 5 in recent months.
I drive an EV. It sucks. Sucks. Sucks. Worst driving experience of all cars I've owned. The lack of buttons and crappy HVAC controls are things I notice daily.
Outside of Tesla owners, I don't know a single EV owner who praises these elements. The EV market does not want this.
What you're suggesting essentially amounts to forced displacement, since it presumably will be accompanied by reduced funding, abandoned infrastructure, and whatever conversion to a national park would entail. You're talking about someone's home.
This is the equivalent of saying "No! They are entitled to remain in their burning building that no one is coming to put out. How could you even consider rescuing them?" The culture is erasing itself through a total fertility rate below 1. I offer an option to assist, because there is no other option. Leave or remain while decline continues. No one is coming to save them at the federal or policy level.
Who will pay for it? Someone will have to. When? And how long will it take? While the population continues to decline forever.
It is unfortunate, I am not unsympathetic, but unless tens of billions of dollars drops out of the sky for these folks, trajectories hold as they are. With US debt exceeding $40T, the deficit being $1T-$2T/year, and no will to increase taxes for the foreseeable future, “it is what it is.” Plane tickets and moving expense reimbursements are comparatively cheap.
Is it better if we do nothing while also not offering an option to leave for a potentially better life? If these folks prefer to remain with inadequate, intermittent infrastructure for the remainder of their lives, that is a choice. These people deserve better, but help at the levels needed isn’t coming for the foreseeable future, if ever.
> Rodríguez is one of the hundreds of thousands of Puerto Ricans who have left the island in the last decade, a migration propelled by an economic crisis and a string of natural disasters that have devastated Puerto Rico in recent years.
> “We love Puerto Rico. It’s our country. However, objectively speaking, we need to recognize that it’s a country in deterioration and at a political crossroads. If it doesn’t come out of it, things could get much worse,” she said.
> The island’s population declined 11.8 percent between 2010 and 2020, according to the U.S. Census Bureau, whose results are used to determine billions of dollars in public funding. Demographers say that steady migration, coupled with a declining birth rate that’s lower than the death rate, is creating a “demographic crisis” that will fundamentally alter Puerto Rico’s society, economy and culture, as family structures change and there are fewer workers to pay for government services.
> “It will literally be another world, another society. It’s not only that there will be fewer people, it’s that the structure completely changes. It’s a complete transformation,” said Hernando Mattei, a demography professor at the Medical Sciences Campus of the University of Puerto Rico.
Regardless the acquisition took some time to bring down the quality of the products. It doesn't matter if their first few products were sufficiently far ahead that they weren't enshitified immediately when Facebook bought Oculus.
Correct me if I'm wrong, but a zero-shot classifier like Jev is fundamentally different to a classifier with a fixed task (e.g. for safeguards), unless they trained a general purpose system to complete the safeguard task, which seems unlikely.
This whole thing reminds me of DeepMind’s Variational Bayesian Last Layers[0], which never gained much traction in the broader “AI” world, but is a remarkably useful tool. And a relatively obvious one that anyone with experience in SVI and with transformer pretraining, seems to independently rediscover (including me) before finding this paper.
But zero-shot classifiers with this level of intelligence, world knowledge, ergonomics, cost profile, and ease of use are new.
I feel like good engineering doesn't just ignore those things, or at least it didn't before recently. Now I guess social media has added a pressure to reduce everything to a hot take.
> Ergonomics, cost profile, and ease of use are new.
Following AI from the academic papers side; jev really feels silly. They one-pass the LLM tranformer stack and tune the output network for a probability value.
(some clever pararellization optimisations to make it viable to offer as an api, since the normal kv cashing no longer works if you oneshot the tranformer)
The largest change is the packaging; An api with a tolken based pricing, and a schema to define the output structure for quick setup.
Previous projects would probably involve installing pytorch, running a converter script on Qwen, and write a fair bit of matrix math to change the output shape.
An LLM is a zero-shot classifier with a large number of classes. All you need to do is establish what the output means and you can fine-tune an LLM final layer for this task if you like (and others have done). A student of mine did this as an exercise two years ago, and it was cool, but not publishable.
I agree with you on the "ease of use" business though. No one thought to make this sort of thing commercially available.
But there is no hot take here. Jev is not some new paradigm; engineering-wise, it is a trivial modification to the existing pipeline. That doesn't mean it isn't commercially viable.
No it wasn't. Those were models trained from scratch, required large scale data, and the nontrivial parts involved training at scale and the autoregressive task which no one expected to work as well as it does.
It is the difference between developing a foundation model, and using one. I believe Jev falls in the latter category, because the task itself is no different, only the output.
No, all they had to do was come up with a quality post-training recipe, production inference stack that wouldn't fall over, GTM, documentation, schemas, etc. etc.
(also most signs point to this being LLaDA 2.0-adjacent so throw in solving some substantial mid-training)
I think it's 100% a hot take to call what they built trivial. Or at least it used to be.
There was a time when that kind of stuff was something between sour grapes and cluelessness about the gap between an idea and an actual commercial product deployed at scale, but now that's just weirdly normalized.
In fact, if anything I'm the weirdo for repeatedly taking issue with the way people are trivializing it ¯\_(ツ)_/¯
Jev actually isn’t anywhere near the top. It even loses to open weight clones. This tells me that whatever their “calibration” dataset is, it doesn’t seem to be anything special.
You linked to some weird subtable that labeled: " Not the default — not the JevBench Score", that can only be reached after you see what I just linked... lmao are you really this hard up about things?
Jeeze, this is getting sad. I guess after all the mass-psychoses where people thought pointless things are going to change the world, we were due for a mass-psychosis where something interesting just has to be pointless?
Because I was specifically responding to your claim that Jev’s training recipe would give it better accuracy than others. It doesn’t have better accuracy than others. You could do as well or better by distilling qwen for example.
Jev is ranked higher than others on the overall benchmark due to speed and/or cost, not accuracy.
Architecturally, it is trivial. That's something the community would have consensus on, so not a hot take.
I see your point, but Jev doesn't exist in a vacuum. When one (like me) says "trivial", they mean it relative to other attempts and developments in the field, all of which require everything you've mentioned at minimum. Commercialising any product, and doing it well, is hard. But the R&D factor here is substantially more straightforward than almost any other product in its category, because there is no architectural breakthrough here.
"all of which require everything you've mentioned at minimum"
Sorry who else did everything I mentioned? I think the guy behind Laya tried after noticing Jev's traction... but the site's auth went down and has stayed down for a day now.
"substantially more straightforward than almost any other product in its category"
More straightforward than the spite projects based on constrained decoding? Or Laya with it's couple of days post-training ModernBERT?
-
I have no doubt other teams can build models like this and I've love for a frontier lab to give us an even smarter model with these ergonomics... but in the rush to show Jev what's up, we're mostly getting slop.
PS: I don't know anyone who's done anything of note who uses trivial like that. The commentariat do, and the "I could have done that" crowd do, but I don't pay much attention to them until they actually do the thing.
By category, I meant other language models in general. The point of others putting something up to beat Jev is to show that, to date, no one has bothered to produce something like Jev, because anyone with decent LLM experience can roll their own for purpose with little effort and have been doing so for years. And can beat it on any metric you choose.
Let me put it this way. OpenAI and Anthropic have a slight moat over the Chinese labs because they have strong training data and the most advanced RL strategies. It will take the Chinese labs significant R&D effort to bridge that, especially in math (and there is a good chance they will, provided they want to).
Jev has no moat other than the fact that no one else has bothered to package a model in this way. Another lab could build a strong competitor very quickly if they want to put the effort in. That's the point of this post. There is no uncertainty about what they have done, nothing to figure out. Someone just needs to do it. I'm not sure what to say if you can't see the difference between the two. Jev is worth celebrating because of the idea to package it in this way. But it is not a paradigm shift and that is likely a problem for them.
They almost certainly would perform worse than more specialized classifiers trained with less data. It’s kind of a paradox of generalization. I think there’s an interesting space where you use generalized models to generate ad hoc specialized classifiers.
Depends what you man by "more specialised". You wont train very good language understanding without alot of data. It probably uses the core tranformer stack from an LLM.
Classic classifiers are regularly just tuned general models; Training a CCN on ImageNet and tune it for cats and dogs gives better results than just training it on cats and dogs.
There is likley a small network used to tranform model output vector to probabilities, but that wouldn't be massive. Retraining that small network for specific task may beat jev; but that's bairly considered training by modern standards.
Expecting a strong zero-shot performer to perform worse in a low data regime?
That only makes sense if you try to rope in data previously used to establish the model's priors, but that wouldn't make sense in this context. That same additional data is what enables things like...
> use generalized models to generate ad hoc specialized classifiers.
Fixed guard today is not very fixed. For ex, the safeguard qwen released is a full 4b llm model. It has no different to normal llm model arch except tuned for this specific purpose,
So it is tuned specifically to classify content for safeguarding? I'm not familiar with this particular model, but it most likely has a specific classifier head that is tuned for the safeguard task. This is completely different to zero-shot classification.
There's a big difference between reintroducing this unoriginal point in the context of this thread and presenting an entire piece of LLM writing as your own. That phrase from this comment isn't pretending to be an original piece of writing.
I'm not sure I agree. The market pressure for this innovation might not have been there with Nvidia chips still available. In this case it is coming from necessity.
People drive Corollas because they're cheap and reliable. Give me that with an electric motor.
reply