realusername

If foreign competition is closing the gap only by distillation, then the frontier labs can focus on preventing distillation and maintain their lead that way.

We already know it's false because you would have hundreds of competitors if it was that easy.

The reason why these Chinese labs are releasing good models is simpler, they have access to a tremendous pool of talented people.

That doesn't really explain if I can use said website on my phone with a password if I login with a passkey on my computer

Well, also expect the next Pixel bootloader to be locked out if they become too big.

The announcement is very funny in that regard, devices with last update in 2023 like the Galaxy S10 are still ok despite having every root exploit under the sun available but god forbid you want to update it to the last version

All the third party roms I know offer sign builds and remote attestation, GrapheneOS isn't special here, it's just that usually app developers just ignore the signatures

If it become to big, it will go the same way as CyanogenMod, Google will send some people to pressure the manufacturers to drop them off

Indeed, if there's one thing China did well, it's that they heavily invested in education and have a very education focused culture.

And in this field, having an army of well educated PHDs is making all the difference

I also don't believe it, if it was as easy as that, we would have hundreds of competitors.

The truth that Anthropic and OpenAI will not say, is that these Chinese labs have a lot of talented people.

MS Office is the last roadblock for Linux desktop. It's literally the last thing holding back many users from moving to anything else.

I think the last one has DRM shenanigans specifically to prevent it to run on Wine? Not sure of the legality of that regarding antitrust legislations but it seems that's what they've done.

It largely works and it's a massive business success.

I'd argue these tech companies got popular because they have good models, nothing else.

Even OpenAI took two years to fix basic chat scrolling.

That argument is moot as distillation also requires a lot of hardware and software, if copying models was as easy as that, we would have hundreds of competing models.

Apps trying to check for jailbreak is still a thing on iOS and as you would expect, it's jank and can trigger on stock iOS if you are unluky

The real economy - the one that makes food, builds homes, provides medical care, etc. - is so much bigger and more important than the AI industry.

The real economy has seen pretty poor growth under Trump tariffs chaos though and I'm not sure the US economy could survive a crash of the tech companies

By 2010 or so, it had the capability to leverage the GPU on pretty much any device

If your device was running Windows sure. On Linux, Flash absolutely sucked with 100% CPU usage to render the most basic still image and on Android it was kind of similar.

It's very easy to say "Yes this game will run on my console at 60 FPS because its identical to the other consoles where it runs at 60 FPS." Differing builds and drivers are not really a concern in the console world, where-as they are in the PC world.

It used to be a selling point of console indeed, however nowadays console are separated by Pro/Non-pro, different revisions and you aren't really guaranteed on how well your game is going to run unless you watch a Youtube let's play of the game you want.

One reason is that Apple does not permit regional binaries, so you cannot use the same listing in the EU/Japan/Brazil/whatever other place may force Apple to act sociable and the rest of the world.

Well, sounds like the next lawsuit waiting to happen then if the regulations can't be applied because of that

I had the same thing but in reverse, it's impossible to convince my company that the React SPA simply doesn't scale despite all the problems we have with it.

They can see each problems one by one but can't put it in the same picture as it's "proven technology"

It's not well known because it's annoying to install ... and it's annoying to install because the stores are huge cash printing machines so that's not going to change.

Huuh, what does this mean? GPT models frequently do 100% of what I tell them to do, anything less and I'd see no point in using agents for work at all.

I never managed to have this experience even with SOTA models, they routinely make architectural mistakes, wrong assumptions and take shortcuts they should not take. Less for sure but they still do it often. I didn't try Fable yet though so can't comment on it.

So based on that, since I have to watch everything they do anyways, why would I pay extra?

But "Kimi only 70% correct" sounds like it's so bad it's not worth using

If you want an analogy, it's like the numbers of 9s in server availability and since currently I'd rate nothing above 90%, it's zero nines. Since I have to deal with unreliability with every provider, I don't see why it would be worth it to pay more to still deal with it.

I'm using Kimi 2.7 and GPT 5.5 at home, Opus (4.8 I think) at work and I really don't see much difference honestly.

Sure, Claude might be 90% correct and Kimi only 70% correct but does that matter when 90% isn't enough to make it work autonomously anyways?

My workflow is just strict supervision of what's happening, I also edit the agents file with anything I see the model doing that I don't like.

My sessions are also short, after any task which is completed, I just kill the session and start a new one so I don't think I have more than 15 min sessions unless it's tech discovery.

I don't think any AI company meets the bar you are setting here.

If this is really their moat then the Chinese companies will win.

I already switched to a Chinese provider personally, I don't think the difference between both really justifies the wide gap in pricing

It's not like you're reading everything the agent is doing in any case

Personally I do, these tools aren't mature enough to be used without supervision

That's because 50k km is pretty much nothing for a brand new car and if it breaks before that, it was badly designed to start with.

Yeah it's actually better and it's not even close, provided that the cars make more than 50k km.

Petrol engines have terrible efficiency and that's also what makes the petrol numbers so bad.

"Amateurs discuss tactics, professionals discuss logistics"

Napoleon

What's the other possible explanation? He wanted to point out a spider at the ceiling?

I don't think that you can find a definition of malware that those won't fit in.

The calendar malware pushes garbage ad notification in loops on the iphone and is triggered without user intent.

And then it's not just "fake apps" but apps designed to steal your credentials appearing at the top of the appstore.

There has never been a widespread malware attack on the iphone

I had a calendar malware that I had to remove from my sister's iphone just last year.

There's been also (and still exists) fake apps pushed with appstore ads, not sure if you count this or not.

It's somewhat more acceptable for entreprise use cases but remember that attestation doesn't tell you that the machine is secure, just that the software wasn't modified.