Why should he stop exactly explaining the situation?

Lenders take on the risk of a default, and when the government lends money, the risk is socialized.

The loan is being made to a private, for-profit corporation, who will be able to keep any profits generated by this scheme (however unlikely that may be).

Whatever activity the loan is for is irrelevant, whether it's for cranking up a crusty old nuclear power plant, or for bailing out a Wall Street firm during a market panic.

It's different from humans in that human opinions, expertise and intelligence are rooted in their experience. Good or bad, and inconsistent as it is, it is far, far more stable than AI. If you've ever tried to work at a long running task with generative AI, the crash in performance as the context rots is very, very noticeable, and it's intrinsic to the technology. Work with a human long enough, and you will see the faults in his reasoning, sure, but it's just as good or bad as it was at the beginning.

Correct. This is why I don't like the term "hallucinate". AIs don't experience hallucinations, because they don't experience anything. The problem they have would more correctly be called, in psychology terms "confabulation" -- they patch up holes in their knowledge by making up plausible sounding facts.

I have experimented with AI assistance for certain tasks, and find that generative AI absolutely passes the Turing test for short sessions -- if anything it's too good; too fast; too well-informed. But the longer the session goes, the more the illusion of intelligence evaporates.

This is because under the hood, what AI is doing is a bunch of linear algebra. The "model" is a set of matrices, and the "context" is a set of vectors representing your session up to the current point, augmented during each prompt response by results from Internet searches. The problem is, the "context" takes up lots of expensive high performance video RAM, and every user only gets so much of that. When you run out of space for your context, the older stuff drops out of the context. This is why credibility drops the longer a session runs. You start with a nice empty context, and you bring in some internet search results and run them through the model and it all makes sense. When you start throwing out parts of the context, the context turns into inconsistent mush.

Ha. I can't access the article to be sure, but I think they mean making the user feel like the computer is incompetent, not making the computer feel bad about its performance like Marvin the Paranoid Android.

Sure, but nothing tangible would be lost - just some notional "Market Cap" reduction which is only really meaningful to Wall Street types, although that would also impact people's pension's etc.

Where nVidia would be exposed would be is companies have bought significant quantities of their chips, but go bust when the bubble pops without settling their invoices from nVidia first. Best case that ties up nVidia's cash while the bankruptcy process happens and they get a significant slice of the outstanding financial pie. Worst case the companies in question all implode with significant debts to multiple parties, their assets (mostly nVidia's chips) get sold off to pay off those debts, but nVidia still gets next to nothing once the assets are divied up. Combined with a heavily reduced market cap, that leaves nVidia with a significant reduction in cash-in-hand and a lack of value with which to help secure loans and investment, which in any case would need to come from a financial market reeling from a multi-trillion dollar bubble pop. The reality is more likely to be somewhere in the middle, but nVidia isn't going to sail through this completely unscathed either.

Doing that would be stupid because we know the government will try to bail them out.

You can't make rational economic decisions when the government will interfere with the market any time it feels like it. That just turns the whole thing into a casino.

Capitalism is an economic system. Plutocracy is a type of government. Voting is not based on wealth. Just because the rich hold extra sway doesn't make it a plutocracy, it just means our politicians are corrupt.

I had to look it up to see how long ago it was called those. Damn Im getting old.

It probably makes more sense given their scale for them to have their own power generation -- solar, wind, and battery storage, maybe gas turbines for extended periods of low renewable availability.

In fact, you could take it further. You could designate town-sized areas for multiple companies' data centers, served by an electricity source (possibly nuclear) and water reclamation and recycling centers providing zero carbon emissions and minimal environmental impact. It would be served by a compact, robust, and completely sepate electrical grid of its own, reducing costs for the data centers and isolating residential customers from the impact of their elecrical use. It would also economically concentrate data centers for businesses providing services they need,reducing costs and increasing profits all around.

Been using Firefox since it was called Firebird. Never once encountered that problem. It's shocks me you found a business that doesn't want money.

Yes.

For "security' I must be forced to log in again every few hours and change my password every few weeks, but downloading random code off the Internet and then shipping it to customers... sure, that's fine.

It's not just about the packages and whether they are malicious or not. These, so far at least, are not - AFAICT they don't even *claim* do anything at all that is functionally useful to a coder so they are never going to get downloaded; their sole purpose is to earn the uploader some of these TEA tokens which, when amalgamated across a few hundred thousand packages, is presumably worth something to them, or why bother? Now that the jig is up, the people that do like to peddle such malware are probably not looking too kindly on whoever pulled this off.

That's the secondary issue here ; like many similar things, whoever came up with this TEA token either didn't consider, or didn't care about, human nature. Anyone with half a clue, or the slightest care about the integrity of such a scheme, should be well aware by now that if you can earn something of value (which need not be monetary) by doing some online clicks, likes, shares, uploads, or whatever then some asshat is going to try and exploit the system so they can get all the benefits without the effort. If your system isn't baking in countermeasures against that kind of abuse, then it's a PoS that should never have left the drawing board but, all too often, human nature rears its head again and says "ship it anyway!" and the enshitification continues.

They hyperscalers are building out (or re-activating) grid-level power supplies for DCs that are not online yet, so are essentially not included in the 2025 figures. All Electrek (a pro-green energy site with a very obvious bias to that effect) are saying is that we collectively built out enough solar and wind to exceed the overall global increase in demand during 2025. Sure, that's a good thing, but it says nothing about how much excess non-green capacity was decomissioned last year (relatively speaking, hardly any), or by just how much that annual green capacity roll out will need to ramp-up to avoid building/re-activating additional non-green power plants to power the new hyperscaler DCs over the next few years (quite a lot, unless the AI bubble pops and most of them never get built).

Frankly, I'd prefer it if they just stopped trying to put a positive spin on everything green and told it like it is for those too dumb/lazy to read between the lines. Yes, we're collectively rolling out greener energy sources at a decent clip, but still far, far, below the rate needed to achieve any meaningful mitigation of mankind's effect on the climate in the timescales that are probably required. According to Electrek nearly everything moving us towards a net zero economy automatically gets an A+ when the reality is probably closer to a B-, or even a C+, must try harder.

It's weird because I watched a Youtube video the other day where some guy who has a long-established UK car channel had collected all the registration data from DVLC for EV sales in the UK and showed that they'd dropped significantly in 2025.

Is this a case of classifying hybrids as EVs or something?

[sarcasm] Of course it can. Also cranial measurements are reliable indicators.[/sarcasm]