mediaman
Are you presenting this as evidence that OpenAI's naming scheme is bad?
If you don't want to spend the time to know the difference between all these different suppliers and their models, then why are you using whatever this program is that gives you so many choices, most of which are bad? Why is this OpenAI's fault?
I see a ton of criticism and not a single person coming up with anything better, unless it's to take away options, like removing effort level.
What about buying Intel CPUs? A car? Electrical components?
If you're not just using this at a consumer level to ask about cooking recipes (and if that's your use case, you really don't care about API pricing, or any of this, and can just use the default), and are using it professionally, then this seems like a low bar of effort to understand what you're buying.
I do grant you the relationship between model and effort level has an unclear relationship, but there are also tradeoffs in most complex technical products that are not always clear; this is exacerbated by the fact that this product is stochastic.
Why are engineers using this as a professional tool expecting this to be bought like a consumer gadget and then blaming the maker?
The lineup of models is still the same.
There are four models: luna, terra, sol, and astra.
Pricing increases for each level. This new 6.1 Terra is 50% cheaper than Sol 6.1.
They are lowering prices for newer releases, so yes if you compare Sol 6.1 to 5.6 Terra it won't look right, but why are you doing that? Look at the same release version for comparable numbers.
Each model improves in successive releases but they maintain pricing and capability coherency within a release version.
Yes, Sol 6.1 is supposed to rival Astra for some work, which is what you'd expect from a later versioned release model against a prior release bigger model.
Assuming they release Astra 6.1 at some point, I expect it'll be better than Sol 6.1, at a higher price.
I don't understand what people are asking for here? No reductions in pricing? Different model names on every version release? Not releasing any models for a version number until they're all ready?
The plan is to let anyone on the subway, with no fares. It turns out that the violent, unstable people who make public spaces unwelcoming are disproportionately represented among those who cannot pay $2.50 for a fare. So they scare off regular transit riders who mode shift to something else. It's vastly more expensive and difficult to react to violence on the subway than to keep violent people out with a small fare, so it's financially and logistically impossible to actually keep it safe. Then your transit ridership tanks and people who hate funding transit can declare the whole thing a giant waste of money and any new funding dies politically because now there's no public ridership constituency left.
Nobody cares about the revenue gained from $2.50 fares. What matters is the tiny amount of resistance that offers to the small number of people who are largely responsible for making public transit unwelcoming.
It doesn't matter if it's committed or not, it still won't be recognized on the formal financial statements.
The OP's question was clearly skeptical though:
---
napkins, and envelopes at least kind of make sense.
Who was ever writing on the back of cigarette packs?
---
Implication being that writing on the back of cig packs does not make sense. I'd agree with your reading if not for the "at least kind of make sense" bit.
Jev's pricing is 4.2 cents per million input tokens, less than half of Gemma 4. Qwen, gemma etc all have terrible non-uniform probabilities; in my testing, Jev's probability scores perform much, much better.
I'm usually seeing round trip classifications from Jev at <200ms for a completed query even with 10k+ input tokens. Gemma 4, depending on the provider, is about 200 ms time to first token - not time to completion. You can likely get better response running it locally, but then you're arguing why you're renting a low utilization 5090 when you could be paying $42 per billion tokens.
Jev can also run dozens of dimensions of classification in one call, without autoregressively generating a sequence of tokens. (i.e., it is much faster and has less order dependency.)
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.
I don't understand how the "just use gemma" crowd has this belief if they've actually built and tested the systems against each other.
However, if the robot is in a contained environment like a home, there is less third party liability than for something like vehicles. Not none, of course, but less. And when you can get your arms around third party liability, first party liability can be better controlled by contractual agreements.
In general I expect on-device motor execution will allow it to safely come to a stop or enter a safe position if it loses the link to the execution planning cloud model.
New trucks optioned out cost $80k+ and plenty of people pay for that without having much justification of any kind for it.
Getting it to write well is really hard because there’s no real way to verify whether it’s good prose or not. You and I can tell, but we can’t write a verifier that codifies our judgment.
Maybe they’ll find a way to improve this, but for now it’s certainly one of the harder problems to solve for LLMs.
Part of it is that I think they also have poor theory of mind, which I imagine is also a hard thing to train it to do.
No-one pays $2k/mo for a fractional CFO.
No-one pays $2k/mo for a fractional sales head.
No-one pays $2k/mo for environmental paperwork consulting for a manufacturer.
They can do it all in house. Why would they?
Right?
Except all of these actually do happen, of course, and each is a big industry.
As soon as this industry lost the inscrutability of the syntax of the work product, it began catastrophizing about the total end of any commercial demand at all.
Yet by this logic there should be almost no service providers at all for any industry that lacks legal certification barriers.
Why is populism doing so well on the right? Is that what the pools of money want?
You could ask the same question about why anyone would rent a VPS. I can just run my own hardware, it's just a computer!
Buy vs rent is not just about what's possible, it's about what's economic.
Dogs are not people. Many in Seattle think their dogs are people, or even prefer dogs over people, but they are not, and they do not enjoy the same protections against discrimination.
If you have never been a landlord, I can see why you would think that "if they damage the property, they pay for it" is a reasonable rule that would work, because if your dog destroyed a door, the carpet, and the blinds, you are likely to be willing (if unhappy) to pay to fix it. If you have been a landlord, you know why this doesn't work, and you would also know why pursuing the damages in court also doesn't work.
(If you are wondering why: people do not pay. Deposits, which they also now capped in Seattle, often don't cover the damage. And suing people who have no money in order to get a judgment that you can't collect on is expensive and pointless.)
This just socializes the cost of pet damage onto everyone, which is why this is populist slop.
(Some of the other junk fees make much more sense: mandatory fees that you have to pay, that are fixed in nature, should just be in the advertised lease price. Think "common area fees" you can't opt out of.)
It does impose significant compliance burdens with massive fines.
You have to report who paid for the ad, where the money came from, what political process it relates to, you have to do this through the entire chain (i.e., ad agency, ad exchanges, publishers), you have to attest to the ad's accuracy, they have to publish all the information through a central EU repository, they are not allowed to use any person-level ad targeting, they have to track whether it pertains to any active EU bill or legislative process within three months of decision-making and, if so, they have to attest that the entity and funds come only from the EU (not just directly, but indirectly).
The penalties for mistakes are enormous. The act allows penalties of up to 6% of global revenue. That's around $10 billion for Meta.
I'm pretty sure I would make the same decision when the EU is threatening to fine me for $10 billion for making any mistake on the rules for the TTPA.
Meta doesn't do this outside the EU, and they didn't do this in the EU prior to the TTPA. Draw your own conclusions about whether this is a "meta problem."
It imposes substantial bureaucratic compliance burden on advertising platforms to accept ads that deal with political issues.
Meta decided that the revenue wasn't worth the compliance burden, so they do not accept any political advertising at all. They filter anything that smells like political or social issues, whether it be a campaign about anti-immigration or a play about feminism, because if they accept it they will get heavily fined by the EU.
This is why Meta does not ban this advertising outside of the EU. It's a EU regulatory problem. The Register does not mention this anywhere in the article.
But I do want them using it to help them research, find interesting contrary views on a subject, gut-check assertions (how many weirdly wrong statements in articles would have been caught by feeding it to an LLM and asking it to question unfounded or incorrect claims?), check technical accuracy of things outside the writer's core domain, track interview transcripts, search/summarize across prior work, do data exploration of public/government reports. Etc. It's a very empowering tool for journalism!
Instead, it's become a religious issue that is more or less at the level of "one drop" theory.
But you have to be willing to break out of the bubble, including getting on a plane, and I see very few people being willing to do that.
Your assumption is that the benchmarks are essentially identical in difficulty, with the only difference being their age and thus whether they could have been trained on.
It is possible to trick pangram - they bias toward a low false positive and a higher false negative - but it is not true that it is essentially random.
DeepInfra sells DS v4-flash at 0.08 in, $0.18 out. Gemma4 they sell for $0.07 in, $0.34 out. OpenAI's price for luna is $0.20 in, $1.20 out.
Why would you assume OpenAI is somehow uniquely incompetent at making small, fast models? And that they're worse at serving it than DeepInfra? Any observer can see they are making money here.
I never understand why people who are convinced there is a big con just don't check market prices and see if there's money to be made.
That doesn't mean their business is great -- they're losing tons of money, but it's because they spend too much on fixed costs, and they can't stop spending money on training next generation models with no end in sight, not because the inference is margin negative, which is a flimsy idea that just clouds the actual business issue.
So we can go round and round on this, each with our made-up objections about how it's temporary or unrealistic or impossible or whatever, or we can just accept the prices as listed and use that to guide our economic decisions.