Discussion
How will the "AI" bubble burst?
It'll take a short while for it to sink in, but with the election results and, more importantly, the now documented opposition against datacenters (46:1), the AI boom could end up in a situation with heavy regulation and far less compute power than planned.
@daveburb Yes, I also think that they are happy to find excuses like "not enough electricity", "no permits" and "evil radicals stifle progress" to get out of their dilemma.
In my personal opinion, the rightsizing has begun. LLMs won't go away. But they will become smaller and more focused (domain expert LLMs), reducing training and operation costs significantly. This will have a direct impact on demand for compute power. The bubble won't burst, it will fizzle out.
@jwildeboer I don't think that's right. Demand for general purpose LLMs will keep growing, not shrink. If the US can't cover it, China will, and certainly not for Germany, so for us this only speeds up our decline.
@andre The problem is that they have to make money with those oversized LLMs. Specialised (and thus much smaller and cheaper to train) models are far better positioned to reach a return on investment. The business world is recognising this and are switching from "renting" big models to running their own "good enough" open-weight models that they specialise for their needs.
@jwildeboer Maybe ROI is a problem for the US labs, but China doesn't care about quick returns. They're investing huge amounts in this for strategic reasons. If the US drops out, China takes over, and we won't be the ones getting their good models.
@andre Another very important reason for that movement is compliance. Many companies, especially here in the EU, fear the legal risks of having to share their confidential information with a handful of "big" LLM providers that are mostly in the US.
@jwildeboer I think that's an illusion. In theory specialised models are better, sure. In practice you'd have to keep retraining them every time the requirements change, and they change all the time. A big general model just handles the new task, and that's why people keep paying for it.
@jwildeboer I'll believe the regulation part when I see it. "But China!" has worked just fine on Democrats for a long time, and once their investment advisers tell them how much of their own portfolio hinges on AI they'll be even less inclined to tank their economy, fictitious as it may be. My money is still on popping through overinflation.
@jwildeboer funny enough the UI of the page has common AI-UI design patterns.
@grauhausen Does that change the numbers?
![SCORECARD - 47 MEASURES
RESTRICT AHEAD ALLOW AHEAD
44 won - 2 leading 1 won - 0 leading
Too close to call 0
No results posted [\]
Precincts in 229 of 229
Leading means ahead by 2 points or more; won means a complete count (every
precinct in, or the county's final count in states that count mail ballots for
days). The measures decided earlier in 2026 are not in the Nov 3 counts.
Nothing here is certified.](https://cdn.masto.host/socialwildeboernet/media_attachments/files/117/568/096/156/013/173/original/0a8c7dbff8fb4650.png)
@jwildeboer
I am starting to wonder if big-tech is starting to actually want the building of data centres to be blocked by communities / governments. They have borrowed too heavily to build them, and they will likely never recover their costs. Regulation gives them a way out of the loan contracts that makes good PR. "The nasty government will not let us build it".