Recent indicators in the AI industry point to a growing crisis of profitability, with OpenAI caught right in the middle.
  • hello_hello [comrade/them]@hexbear.netEnglish
    32·
    4 days ago

    China seems long overdue to overtake the collective West in LLMs. They have societal support, industrial base and government planning to do so.

    Nvidia corporation and government sanctions are really what’s preventing a full takeover (as well as capitalist mutual aid)

    • footfaults@lemmygrad.mlEnglish
      33·
      4 days ago

      Just like Bitcoin mining in China, being able to waste resources more efficiently than the West is still waste.

      • hello_hello [comrade/them]@hexbear.netEnglish
        11·
        4 days ago

        The product itself is questionable, but the tech soft power and innovation drive for domestic self sufficiency to cope with this new demand will benefit China in the long run.

        But I hope the Chinese central govt isn’t banking on AI to solve their domestic economic model, we aren’t getting computer automated socialism anytime soon from this.

  • fox [comrade/them]@hexbear.netEnglish
    21·
    4 days ago

    Last year David Cahn wrote an article for VC firm Sequoia Capital titled AI’s $600 Billion Question.. The gist being that AI needs to make that amount of money merely to break even. There’s no shot in hell that with the massive investment in data centers and no changes to pricing that the hole isn’t deepening.

    • Lyudmila [she/her, comrade/them]@hexbear.netEnglish
      9·
      4 days ago

      Oh, the pricing is changing, and it’s just gonna deepen the hole even more because most people are gonna go “$20 a month? For what?” and suddenly the user base that is currently the unprofitable product will largely conglomerate onto the remaining free platforms bogging them down so hard they also have to change their pricing model.

      There simply is no way to make the current stuff cost effective at the scale they’re currently using, even with the bleeding edge most efficient models. Models need to become multiple orders of magnitude more efficient to become even remotely viable. Running distills on local machines is the only option that makes sense financially for actual use because you’re putting all of the energy and compute costs onto the user directly.

      The big push right now is to get NPUs into the hardware into every single person’s hands because… reasons. Microsoft Copilot+, Chromebook Plus, Gemini Built-in, Apple Intelligence, etc. are all examples of this which are still highly dependent on cloud computing, but the effort is to make as much as possible local. What they’re actually able to do with that hardware kinda remains to be seen because most of it is barely doing anything at this point.

      “Why waste time on optimization so you could run software locally when instead you could spend 1000x as much on hardware to run inefficient code?” 10/10 dentists software engineers.

    • aanes_appreciator [he/him, comrade/them]@hexbear.netEnglish
      4·
      3 days ago

      Thing is, how does this fit in with the neoliberal tech trend of just… losing money?

      The entire gig economy was subsidised by massive loans and investments, with only a few companies just recently reaching any kind of profitability! Uber only turned +$9mil this year after accruing nearly $250mil in debt!!!

      These businesses are run like states: massive debt pools held aloft by the debt and supplementary income that pays off other debts. You might as well be able to buy bonds for these things.

      My question is though, how does this fit into Marxist economic theory? I’m aware of the “reducing rate of profit” and such, but how does this fit for swathes of the economy operating on negative profit as its ideal model? I suppose these firms are “micro” relative to the “macro” of the whole economy.

      Are these gargantuan loss making businesses dragging the whole system down with them?

      • burlemarx@lemmygrad.mlEnglish
        3·
        3 days ago

        The fall of the rate of profit is a tendency (not a prophecy) related to competition between industrial companies that overcompete themselves. This is one of the crisis theories of Marx.

        However, capitalism didn’t have the level of financialization it reached in the last 30 years. The whole financial bubbles we see are of the type of overproduction, or production vs realization.

        There’s a big trend in finance of overestimating the future capacity of the capital of a firm to be able to reproduce itself by becoming profitable in the future. The problem is that the expectations spiral out of control, especially in case of tech companies that favor massive growth instead of financial sustainability.

        The problem of this approach is that economists and financiers forget that exchange is a social relation. The money spent on constant capital does not circulate back into the economy unless the aggregate demand can keep up with it. With AI requiring massive investments and promising to destroy many jobs or bringing salaries down makes the scenario of exchange value that will never become use-values, or over production, or in more modern language, financial insolubility.

        When new programmers are hired, they generate new demand for products, housing, services and more. AI products require an infrastructure that is most of the time completely automated, with automated chip production, automated energy generation and automated datacenter management. These costs will never circulate back by generating new demand, and are instead just sinks of wasted resources.

        And there’s one more thing, which is the sunk cost fallacy. Financiers invested a lot of money into something expecting it to grow. When it’s not solvent, they tend to add even more money and to hype that thing so more people and firms put even more money into it. Then there’s this solvency gap into their investment. Do you know what they are going to do? Pour even more money in it with the expectation of it becoming solvent in the future. At this point, there must be a lot of derivatives circulating around based on AI stocks. So it is possible that we are seeing another bubble popping. We just don’t know when.

        Edit:

        There is even the scenario where AI companies hike their prices, but then it compromises existing demand. Then the AI companies can become insolvent if at any time they need to raise prices while at the same time reduce the demand for the technology, and aren’t able to become profitable. Then once again, the companies will become insolvent and the only way they can continue is if investors pour more money into it.

  • Des [she/her, they/them]@hexbear.netEnglish
    18·
    4 days ago

    don’t forget to repeatedly thank chat gpt in the prompt. oh have the AI voice read everything out too

    yes short term energy suck but you are basically contributing the bankrupting this company

    • LanyrdSkynrd [comrade/them, any]@hexbear.netEnglish
      3·
      3 days ago

      Microsoft owns most of the upside of OpenAI already. Everything they’ve invested so far is recoupable as 20% of future profits, and much of what was invested was in the form of Azure credits. They have a perpetual license to any OpenAI technology developed to date and in the future until they achieve AGI.

      I don’t see why they’d want OpenAI. What they really want is a bunch of other suckers to come in and keep financing it so they can profit from the previous deal.