I can’t believe I’m nitpicking this on HN of all places, but the Taurus-Miltank thing is just fan headcanon based on nothing more than the analogy with real-life animals. While it’s true that Tauros is a male-only species and Miltank is female-only, they have separate Pokédex numbers and were introduced in different generations.

They can breed (because they’re in the same egg group) but the offspring can only ever be a Miltank. The only way to breed a Tauros is via ditto, same as with any other male-only species.

> The only way to breed a Tauros is via ditto, same as with any other male-only species.

You're half right. I was wrong about Tauros and Miltank.

But Volbeat and Nidorino are male-only species that don't need to be bred via Ditto. You can breed Volbeat with Ditto and get an Illumise. You can breed anything with Illumise and get a Volbeat. You can breed anything with Nidoran-female and get a Nidoran-male. You can breed a Nidorino with Ditto and get a Nidoran-female.

Outside-of-universe, this is obviously because the technology of representing Pokemon within the game has changed over time. Later male-female pair species are given the same name and the same Pokedex number. But within-universe, that explanation isn't available.

There are real examples of males and females being initially mistaken for different species, as well as for adults and juvenile forms.

e.g. https://en.wikipedia.org/wiki/Cetomimidae

  In early 2009, the Royal Society published an article detailing the discovery "that three families with greatly differing morphologies, Mirapinnidae (tapetails), Megalomycteridae (bignose fishes), and Cetomimidae (whalefishes), are larvae, males, and females, respectively, of a single-family, Cetomimidae."

In this case the males and females "actually are" different species by an objective definition, though not by every objective definition.

Is Nidorino/Nidorina headcanon too? Each can also be one gender and have different entries in the Pokédex. Are they considered each other's male/female version?

For days? Someone spent days trying to convince Claude to do something?

If you asked yesterday, and asked again today, then you asked for days. OP might be trying to express that it wasn’t just a temporary fluke.

I've long maintained that the real indicator that AGI is imminent is that public availability stops being a thing. If you truly believed you had a superhuman, godlike mind in your thrall, renting it out for $20/month would be the last thing you would choose to do with it.

Simpler explanation : they don't have enough GPUs to release this much larger model.

Yep, I'm skeptical about their inference efficiency, given how much they're scrambling to reduce compute when they're already the most expensive by far (and in my experience not the best quality either).

However we cannot observe these things directly and it could be simply that OpenAI are willing to burn cash harder for now.

This is actual reason. So any investors reading our system card.... write us another check and watch the $$$$$$$$ roll in. It's so dangerous we can't even release it!

And/or it isn’t cost effective to run.

Thats what he said

Quite, given Claude is down this morning...

but simpler explanation work until they don't all of a sudden

That logic makes sense, but them hyping up the model is a sign that this is just another marketing stunt. Otherwise, we wouldn't even be hearing about it rather than a media blitz designed to stoke demand for their dangerous and exclusive world changing super model.

This is the same scheme that OpenAI has used since GPT 2. "Oh no, it's so dangerous we have to limit public access." Great for raising money from investors, but nothing more than a marketing blitz campaign. Additionally, the competitors are probably about to release their models, while Anthropic is still lagging on the necessary infrastructure to serve their old models. So they have to announce their model before the others to stay at least somewhat relevant in the news cycle.

Anthropic needs money like the 112B OpenAI got. They could be hyping and this is good hype. Who knows how benchmaxxed they are.

If they provide access to 3rd party benchmarking (not just one) than maybe I'll believe it. Until then...

You don't need to believe it. The real story will be if companies allowed to use it, stick with it.

You have to recoup your training costs though? But I’m sure you would have better option than renting it to the general public if you indeed have a perfected AI

If you truly have an artificial superhuman mind, you don't need to rent it out to profit from it. You can skip to the chase and just have it run businesses itself, instead of renting it to human entrepreneur middlemen.

Running businesses and dealing with customers can be a major pain. There’s a lot of soft work in any business on top of the technical work.

Why bother with all that when you can simply charge an extortionate rate and customers will pay it anyway because it’s still profitable?

Public APIs get distilled, this is why Deepseek and Qwen are so competitive.

I am very confident that frontier models won’t be public at strong AGI levels, and certainly not at superhuman levels.

Because other than SWEs, very few other segments extract significant value from cutting edge AI at present. I suspect that for the average Joe conversing with their chat, GPT-4o was more than adequate (and really, when OpenAI tried to phase that out, the public revolted and they brought it back in).

So companies might pay good money for these models for programming but elsewhere, I don't see where they capture particular interest yet.

It could be both? But renting to a few for a really large amount of money would be very low effort for massive revenue, compared to starting new businesses

Another option is to become a holding company with equity stakes in both suppliers (e.g. AMD) and vertical market customers.

I'm curious if any models are being trained explicitly on business management.

I'm also wondering how performance would be tested, and how much results would depend on specific surrounding contexts (law, regulations, and so on) and what happens legally if a model breaks applicable laws.

I mean actual going-concern businesses with customers, marketing, deliverables of some kind, and support. Not toy activities like share trading.

It only makes sense to rent out tokens if you aren't able to get more value from them yourself.

I would go a step further and posit that when things appear close Nvidia will stop selling chips (while appearing to continue by selling a trickle). And Google will similarly stop renting out TPUs. Both signals may be muddled by private chip production numbers.

You would if there was one other company with a just as capable god like AI. You’d undercut them by 500 which would make them undercut you. Do that a couple of times and boom. 20 dollars.

That's still assuming that they're competing as consumer tools, rather than competing to discover the next miracle drug or trading algorithm or whatever. The idea is that there'd more profitable uses for a super-intelligent computer, even if there were more than one.

But would miracle drugs and trading algorithms be as profitable as AI research/chip design/energy research? Probably if AI is by far the biggest growth in the economy majority of the AI's usage internally should (as incentivized by economics) in some way work towards making itself better.

I think they'll just increase the price to $1k/month. I don't think they will gate it as long as they can make sure it doesn't design a nuke for you, etc.

in this case it's far from it, hacking stuff is a small dimension of AGI

That's the thing, when that level comes we will never know it's here. The only thing we'll have as evidence is the company who has it will always have a "public" model that is just slightly ahead of all competitors to keep market share while takeoff happens internally until they make big bang moves to lock in monopoly level/too big to fail/government protection to ensure utter victory.

I don't think OP was making a value judgment or anything. It's just weird to say you won't consider Codeberg because you need reliability when Codeberg's uptime is at 100% and Github's is at 90%.

Windows is not public infrastructure. If the government's reliance on it has reached the level of "national importance", then that's the problem that needs to be addressed, not Windows' ownership.

Public infrastructure should be built on open-source, period.

Why is Windows not public infrastructure? Because it's privately owned, or because it's not relevant to enough of the public? I argue that it is public in function. My thinking matches yours as regards OSS.

The government haven't yet mandated you use windows. Yet. It will be soon, like with androids and iphones, for user identification so the government knows who sends every network packet.

Compared to getting them nothing, yes. But the OP's point is that this doesn't prevent the child from mentally comparing themselves to peers that have a smartphone, and viewing their Tin Can as a "restriction" imposed by their parents.

Which it is. I don't understand the need to wink-wink-nudge-nudge pretend it's anything else by the others in this thread. Just own it, restrictions aren't bad by default.

Well then clearly you haven't taken a look at https://status.claude.com.

At that 3rd party side GH is currently noticeable worse then claude ...

Like they are down to one 9 availability and very very close to losing that to (90.2x%).

This also fit more closely to my personal experience, then the 99.900-99.989 range the article indicates...

Through honestly 99.9% means 8.76h downtime a year, if we say no more then 20min down time per 3 hours (sliding window), and no more then 1h a day, and >50% downtime being (localized) off-working hours (e.g. night, Sat,Sun) then 99.9% is something you can work with. Sure it would sometimes be slightly annoying. But should not cause any real issues.

On the other hand 90.21%... That is 35.73h outage a year. Probably still fine if for each location the working hour availability is 99.95% and the previous constraints are there. But uh, wtf. that just isn't right for a company of that size.

You may have fumbled the calculator at one point. 20Min per 3 hours is 88.8% uptime. 99.9% uptime is 11 seconds down.

what I meant is having both at the same time

- at most 20min per 3 hour

- and 99.9% uptime on a yearly basis

as in your yearly budged of outage is ~8.76h but that budged shouldn't happen all at once and if there is an outage it at most delays works by 20min at a time, and not directly again after you had a downtime

but I did fumble the 90.21% part, which is ~35.73 days i.e. over 857 hours....

> On the other hand 90.21%... That is 35.73h outage a year.

Days, not hours.

This is ... surprisingly honest? The one above is "missing" status page; and most status pages would legally have to be filed in the "fiction" section of the library.

I get the email notifications from Anthropic’s status monitor, and I think they might be my most frequent emailer these days.

If this was an Apple TV app, I think it would be the default mode in my household.

Improves outputs relative to what? Compared to previous contexts of 1M, it improves outputs by allowing them to exist (because previously you couldn't exceed 200K). Compared to contexts of <200K, it degrades outputs rather than improves them, but that's what you'd expect from longer contexts. It's still better than compaction, which was previously the alternative.

I don't think they're claiming "no degradation at scale", are they? They still report a 91.9->78.3 drop. That's just a better drop than everyone else (is the claim).