Well, from all the recent opinions I am seeing now on youtube (OpenAI’s claims of AGI notwithstanding…) and OpenAI being unlikely to be able to launch an IPO given Apple’s killer intellectual property theft lawsuit, I think it’s time to move on to the next question after “is this all a bubble that will just burst?” (answer: YES; we have seen the beginnings of it…)
The next question: What impact will this have on the economy?
Guys like Zitron think it will have a devastating impact. I disagree with him there (his astute critiques of the fraudulence in AI marketing, notwithstanding…) for the following reasons:
The Anthropic IPO has not happened yet. If this awareness of most of the fraudulence only came after that IPO, yes: it would have been a disaster! Now that we see the AI emperors are not wearing clothes (or to use another Wall Street analogy: we see after the tide going out that none of the AI emperors had swim trunks on…), it’s a simple case of caveat emptor. Buy other things; DON’T bet the farm on AI!
If the stupidity of the crowd wants to double-down on AI after all we’ve seen, those investors are the ones to lose. And if investing professionals want to push AI on their clients, again: that will be on them, NOT the economy.
We now know any business model predicated on data centers is a loser. News flash: not all AI business models need a datacenter; the tech can use self-hosting, but that might constrain its application for some businesses, not be an issue for others. Caveat emptor!
So somebody is going to lose money here; right??
Sure! The monopolies themselves: Anthropic, NVIDIA, Google, OpenAI, Meta, Softbank -do we count Grok? (grin…) The people building data centers: Oracle, some others (including Mr.Wonderful’s and Eric Adams’ opportunity to promote them…)
If Blackrock, Vanguard, Goldman, et.al. are long on those companies for reasons of AI -then yes; the impact to the larger economy will be huge. I could be wrong, but they have good analysts who have been advising them to not bet the farm on AI.
So the question to be implicitly asked is: if those companies who are central to this AI phenomena have a disastrous balance sheet this year, does that mean curtains for the economy at large? Clearly, their crap showings will impact the economy; but do they constitute even a majority of the earnings of the US economy? I don’t think so. And the sooner they own up to what AI can really do well, the sooner the economic healing can begin.
Companies who lost are the ones who foolishly fired all their devs before running things in parallel for at least 6 months.
And some other companies who lost are ones who used AI as an excuse to layoff, when they might not have been doing much of any AI at all (just a few cosmetic projects, to make that lie credible…) So MANY companies have not been building anything in software, not addressing any of the software issues they need to be competitive, and investors are soon going to get wise to that. And that’s because of Trump and Luttnick’s nightmarish economic policies. We have been in a period of stealth non-production (and don’t tell me anything about the administration’s fabricated labor statistics!) The solution to that is: start building and fixing pronto. You don’t have the money to do that?? SHRINK! Stop lying to investors about your fantastic AI savings!!
So what’s the value in AI? It CAN be used to improve software quality; it can check code for obvious errors, and do offline testing of it. But since this was not merely as attractive a sell as: “Now you can fire expensive employees and replace them with AI that develops solutions for you!”, the AI monopoly wasn’t interested in pushing that! Add to that, the cost of AI at scale often does not justify its use (especially now with the much cheaper cost using China’s DeepSeek…)
AI definitely has a use case; just not the absurd one hyped by the monopolies.
So my take is: yes; we CAN get past this ludicrous bubble. And those adjacent to the interests of the monopolies are just angling for those monopolies to get bailouts, so the money can keep trickling down to them. The monopolies had years of stratospheric returns. Time [this and next year] for them to take some bad with their previous good.
Will generative AI be possible? Not anytime soon (much longer than the 10,000 days Altman claimed!) Most “AI” now is frankensteined data from training sets (usually: the Internet) When applied to software solutions, the code can only be as good as a solution it finds on the Internet -and that solution is often not on the Internet, or an aggregation of solutions is found to be messed up. For a computer to think, it must perform a staggering number of simulated possibilities -because that is what human thought does (that’s an oversimplification: human thought uses a combination of structured and unstructured approaches, all at a fantastic speed that does not rely on a digital architecture, but an analog one…) We’re not close!
But the statistical correlation apptoach that AI uses right now can have some value. Just don’t call it “thinking”!
Bet: Anthropic will postpone their IPO for some pretty bullshit reason (it’s supposed to happen this month…)
Anthropic made a good LLM, but they ain’t “generative”; and people are starting to assess the value proposition they purport to have is not there (though they offer some value, but by no means the level to match the hype that has been put out there…)
To try and save it, they have been putting out a lot of feel-good advertising. But the bad will already put out by the AI cartel ringleaders is too difficult for them to overcome with just an ad campaign.
When the IPO is postponed, wait for Anthropic’s explanation: