Hmm. Maybe that would actually work.

The most likely current scenario is that even without all the damage it does, LLM-based coding is significantly more expensive (when you have to pay the real cost) than competent human engineers doing the same job. Of course, if you add all the defects LLM code has into the mix ...

The kernel works totally fine without systemd. And it is still quite possible and will likely remain so to run your servers and clients without systemd too. Doing so has saved me from several frantic security patching session by now. That thing it just total crap, but I have no issue with others doing it to themselves, as long as I can stay away.

Indeed. Linux also has a tendency (which I agree to) to give them enough rope to hang themselves. If the slop becomes to bad, I have full confidence it will get stomped on, hard. That said, contributors that are actually competent, understand careful review and documentation and deliver high quality with some AI support helping them, I have absolutely no problem with. The current, most serious concern with LLM coding is that it costs more and delivers inferior results compared to humans in many application scenarios. And that it is a huge risk to use it without very careful review, which may just take more time than doing it by hand in the first place. If somebody really insists on inflicting that on themselves or if people actually understand where to not use LLMs and use them only were it makes sense, I am totally fine with that.

Complete nonsense, of course. You just gave an nice example of how to NOT do it.

Yes. Up to and including "cognitive surrender". The most idiotic morons even want AI to write code that then gets reviewed by AI and they deeply believe everything will be fine with that approach.

AI can help. It cannot do the full job. For LLM-type AI that will likely remain true forever, because LLMs cannot do reliability.

The Kernel is very rationally maintained. If AI slop becomes too much of a problem (and it likely will), it will get banned. It has its chance now. Maybe it will work out, maybe not. With other abject AI failures to do secure code or to find most security problems in code, I expect that in 2-3 years, this will get resolved. All it takes is some major screw-ups.

And why, oh so wise one, have they not done that blatantly obvious and cheap thing?

All you are proving here is that you have absolutely no clue how things actually work.

Don't be a lazy fuck. Do some research before disgracing yourself.

Well...

Well, the US had sort-of peace and even nuclear inspection, before criminal idiot Trump came along and had to, again, misdirect away from him raping teenagers.

And no, the US cannot reasonably occupy Iran. It would break the US military and economy. And it would have to be done very long term. The US did not even manage that in Afghanistan, and that has about half the population of Iran.

Agreed. This is completely pathetic.

Hence the numbers I gave. Most people questioning authority do indeed just want some other authority in power.

With AI agents not even being remotely secure, anybody running this is asking for it.