Fascinating. This is even worse than I thought. For this, there should be ample correct data in the training data set. But it still cannot do it. Well.

The reason scams like this are illegal is because there is always plenty of pople who believe this crap against all rationality. If regular people were able to fact-check, scammers would die out.

The home-user does not really matter in this. It is corporate users that provide basically all MS profits.

Then add Azure getting hacked and being massively insecure, several times now. Add that many people are looking into leaving o365 because MS blocked a user for political reasons.

MS is done for. They just will take quite a while dying. But there is no realistic chance they can turn things around anymore.

Yes. That should get this person barred for life from any research position. Had that gone on for longer, it would have done significant damage.

Lets face it: LLMs are somewhat better search, can summarize non-complex texts and can do simple, well-known work with not very good reliability. But that really is it. And that in no way justifies the hype.

There are other factors at work as well. But we will see what happens. Still, being well educated is always a huge advantage, even if it sometimes requires time to manifest. Obviously, if you study a BS subject or do not apply yourself, you will be screwed except in a boom.

Yellowstone needs to be drained of heat, otherwise it will burst the ugliest of bursts.

What nonsense. First, the most important part in teaching is to select the materials and structure them in a way that makes sense. Second is the actual teaching and anybody halfway competent does far more than just reading the slides. It is about demonstrating you know your stuff, the materials are worthwhile working through, you respect the time of the participants and any good lecture will also need a real entertainment factor.

I think you have never designed and then held a lecture. And if you ever have heard lectures, apparently they were not any good.

Indeed. As an example, I currently have a student looking at all the major AIs (including coding ones with paid subscriptions) for code security review. With small, well known samples they are good. With larger samples, they are >50% fail. With CVEs (the things that matter) they are so far almost 100% fail.

Add that using AI coding assistants makes you about 20% slower, and the only thing AI could be called for this application is "completely unsuitable".

We see these ideas that are obviously nonsense all the time. This one has been picked apart by multiple people with industry experience already.

What these things are is essentially the venture capital version of the scam mails you get in your mailbox every day. If you make it big enough and insane enough, someone with more money than brains will think he spotted an opportunity that everyone else missed and will invest.

Why is it, you think, that 99% of these things vanish without a trace after an initial storm of publicity?

Indeed. I mean they have been subjected to slow for a full semester now. Anybody that does not see the limitations after that has a problem.

Hahaha, in 4 years the collapse of the hallucination that LLMs are actually very useful will have concluded. Things are already mightily crumbling. Those that apply themselves and learn stuff will find something worthwhile in 4 years. Obviously, with this mockery of teaching, that will be impossible, but real teaching is still being done. You just need to insist on it.

Tell that to those that place profits over everything else in life. Failures like Elon Musk that, despite all his money, obviously suffers from a deep (and justified) feeling of inadequacy.