Depends ho crappy that "national security" is. From available evidence, it may be shockingly bad and then even an Artificial Idiot may be a "threat"...

Spreadsheets are not racist, so there is no legal case. LLMs are known to have bias, so there may be one. At least that is what I think is the motivation here.

The problem seems to be that he is difficult to get rid of. Business-strategy wise, he has basically only screwed up for a long time now.

Sad as that is, cannot really argue with your point. The thing is, these may be people that do not have much choice.

Yep. Crappy people doing crappy things.

Seems somebody failed the creativity exam.

Fairphone.

Possibly. Many will currently hope the problems pass fast enough, but they might not. So unless they have reserves and are prepared to spend them...

Indeed. And they are going to be called upon to help clean up the mess many people made going all in on LLMs ...

Yes. Pretty clear leadership failure IMO. And now they are paying for it.

I think it is becoming more and more obvious that LLMs are just another, relatively small improvement of things. Like happened with all the other AI hypes before. Useful? Yes. Complete game-changer? No. And it is not only that running these models is far too expensive and that is several years in.

Hahahaha, no. QCs (if they ever work) are exceptionally slow and computations needs lots of set-up and repetitions. Completely unusable for anything even remotely real-time. Also, QCs only can do tiny computations (if they ever even can do those) that need to have special mathematical properties to make them worthwhile.

Some people just have not clue how things actually work and are loud about it. What else is new...

Seriously? There is "running a model" and then there is "running a model". For most things you either need the large ones or a specialist model.

For that to happen, they would probably have to keep it going for another 20 years or more, and mostly free or cheap. Somehow I do not see that.