Maybe simplistic humans. Well, to be fair, that covers most of them.

As LLMs are based on a massive theft of information, everybody should now steal from the LLMs as well. It is just fair.

Yes. Also, I have seen estimates that they probably burned $20M in tokens for that attack. That makes them like 1000x (or more) more expensive than a competent pen-tester.

Yes. At least for sites easy to hack and there are a lot of those. But the lack of effective isolation is surprising enough to ask whether this was intentionally done. At the very least it is gross negligence.

They found that obvious and somewhat obvious lies still get them 95% of the effect. Hence why go for cleverly crafted when simple does the job nicely.

I mean on some level the AI fans are cult-like. A cult mind-set destroys the ability to fact-check.

Yes. Basically the wizard's apprentice all over. Again.

It may be that. Observable history makes that pretty likely. Or they may be facing genuine shortages of money and are trying to actually slow things down. They may also have noticed that they are doing very criminal things if they run models than then hack others.

Exactly. The question is however whether they can keep enough stupid money flowing to survive even if they slow down training. Merely running the models is still quite a bit more expensive than what they can charge for it and they lose money on every prompt.

I think we may be seeing the begin of the crash. Nothing is sure, obviously, as this tech has been massively overvalued for several years now bit they still managed to get mountains of more money.

To be fair, most of that trust was quite irrational to begin with. But exploiting irrational trust is what scammers do and the general consent is that these people still belong in prison.

Indeed. Open Source models would mean you get the training data and training methods, essentially the "build system" for the model. As training effort is typically very high, that would not help a lot though.

Yes. Concord failed because it was more of the same and nobody cared. That is not what Kojima does. My take is Sony got Kojima exactly for this, then got cold feet when they saw he had new ideas.

Sure. But there is a thing called an "decryption oracle". Apple can definitely push an update to this specific device that decrypts all the message that the attackers want decrypted and send them back. It may require the user to decrypt something else and then they just use that unlock. Unless the secure enclave directly communicates with the user and asks the user about every individual message, this type of attack is possible.

Also note that this type of attack is well-known to cryptographers and competent users of cryptography. It is not special or obscure or anything.

So, getting massive fines levelled on Apple UK would do nothing for Apple US, you think? Think again.

What this probably shows is that the guardrail-approach is not working. If they need to massively overblock to keep things out that they think they need to keep out to avoid bad press, then this strongly Impacts, as you say, usefulness. And they would not do that unless they think they have to.

Raw networking and kernel modules are areas that are hard (I have experience with both) and even only having a good lookup tool (which is one of the things LLMs can do, as long as you verify the answers) would be very beneficial. If they block that, they do so because they have not found anything better.

The whole idea of LLM agents and LLMs doing active stuff is not ready for prime-time. Not a surprise, that usually takes 10-30 years (sometimes longer) from the initial demo, and we are now at year 5 or so and this is high-complexity stuff, so probably 10 years will not be enough. And no, throwing a ton of money or effort at the problem does not accelerate things. Sure, you can do some early stunts that way (like going to the moon a few times and then not doing it again for decades, because the tech was not really there and things were done with extreme effort), but getting things to maturity takes time. One thing people generally do not understand, but the history of science and technology and engineering in particular, is clear. There are no real shortcuts and things are moving slowly.

My take is something ins wrong with Sony. They last disaster was due to lack of innovation and new ideas. They probably got Kojima to do something about that, but then they saw innovation and new ideas and got cold feed.