40 bit is within reach of practical guessing attacks. 128 bit is not, but just barely (the limit where it becomes reliably unguessable is somewhere in the 80...100 bit range). If you are serious about things you use at the very least 256 bit. And, obviously, you do independent expert review on any major change. This one is so abysmally insightless and bad, it may well have been an intentionally placed backdoor. I wonder whether they can still reliably say who did the change.

As to who certified that, this is the crapto space. Where everything is easy, nothing needs to be certified and stupid is the name of the game.

Hahahaha, no. I am a bit more sophisticated than you are. Stop projecting.

You think pointing out facts costs me credibility? Well, maybe with the stupid. You seem to be a prime example.

Oh, and look. I am posting at +2. That must kill you inside.

Seriously, this abysmally bad idea has no place in the news. Only blithering idiots are still in there.

It is actually worse. LLMs do sort-of have a world-model, but they cannot really use it. What they lack is deductive capabilities with the power needed to do plausibility checking against that world model. And there is no way to create that because statistical "deduction" will always be very shallow and very unreliable.

As a simpler description, LLMs have all the data, but they have no insight that would allow them to use that data competently. They can only make statistical guesses and that is not enough outside of some small and not very important application areas. Hence engineers will not get replaced. But they may get pissed and some people may find nobody wants to work for them anymore.

I do not think it looks risky. I think by now failure is basically assured. No risk in that.

And there are more threats. The EU CRA and also the DSA place requirements on software that are simply outside of what LLMs (or semi-competent coders) can do. Software is about to get a whole lot harder.

Sorry, but AI is not great for coding. Sure, if you were coding on slop-level before, you can do it faster and with even less understanding using an LLM to assist you. But liability and other requirements are finally coming to software. The EU CRA and DSA creates requirements that LLMs (or human-creates slop) cannot fulfill. And they both come with real teeth.

Indeed. It is an R&D project scaled-up to absolute insane dimensions. That cannot work.

No risk in that. I mean more and more enterprises see no revenue growth from using AI, AI code is problematic in several regards including copyright, prompt-injection is unsolved and may be unsolvable, and now it turns out that running these things may make you hack everybody around you.

There really is nothing this tech has going for it except somewhat better search. The rest is either massively risky or does not deliver.

Or, alternatively, that the cretins in power try to steal the election and cause massive problems, up to and including civil war.

Indeed. I am wondering whether they intentionally drag their LLM investments out. They must know it is 95% hot air.

It does only if you ignore that the CRA and DSA in the EU are now starting to become effective. If you are fine with cappy code as before, sure. But then you will lose a rather large market.

On the plus side, all AI hypes so far (8 or 10 or so) have _all_ crashed down to "somewhat minor improvement for some things". This one here will likely do the same. Especially as more and more enterprises find there are no revenue gains manifesting from intense AI use.

Indeed. The risks would be on "insane" level. I mean even if the LLM bubble keeps going (I really hope not), at some point they just replace hardware instead of buying for new installations.