Companies Are Throttling Employees’ AI Use Because It’s Too Expensive
www.404media.co
Sources and leaks from Amazon, Adobe, Atlassian, Citi, and more show what is really happening with AI right now: companies are trying to rein in AI use as costs spiral out of control.

Sources and leaks from Amazon, Adobe, Atlassian, Citi, and more show what is really happening with AI right now: companies are trying to rein in AI use as costs spiral out of control.

  • yesman@lemmy.world
    3·
    2 months ago

    Curious to see how it works out when the AI companies need to grow exponentially to be profitable while their biggest customers are cutting back.

    • Mirror Giraffe@piefed.socialEnglish
      1·
      2 months ago

      Yeah. Unless an efficiency miracle happens or the tooling improves to such extent that the end product is worth the cost this bubble will pop shortly. The fact that the LLM companies are losing money on every token sold makes it hard to believe that they are looking to adjust prices downward anytime soon.

      I presume that if there are use cases that the LLM is extraordinarily capable of that makes it more efficient than the standard procedures that could be what drives LLM usage forward. From what I’ve heard a lot of document normalisation has LLM in it’s pipelines and probably there are other areas that can take a price hike but I doubt it’s enough to justify the data center boom.

  • JeeBaiChow@lemmy.world
    1·
    2 months ago

    The great part is, the people who could most efficiently use the tokens, i.e. the experienced engineers, either got fired or simply wont use AI because they recognise its value proposition better than the Csuites. Youre left with the hapless junior hires trying repeatedly to coax a working solution from the systems, burning though loads of tokens as a result with less than stellar reaults.

    Its beautiful!

    • DaddleDew@lemmy.world
      1·
      2 months ago

      I’m so glad you asked.

      Kitboga, a YouTuber who loves to mess with scam callers, has recently made a video showing what you can do to mess with an AI agent.

      In a nutshell, they are designed to do what you ask them to do and be WAY more accommodating than any human would be. The more crazy things you prompt it to do that pulls the conversation towards something that no longer resembles a normal conversation, the more they will tend to ignore their original prompts and be easily talked into putting themselves in infinite loops. They also start to spectacularly glitch out because there is no training data for what you’re making them do.

      Tons of fun to be had.