It’s been a couple of weeks since my last set of predictions on the AI winter. I’ve found myself making a couple more.

Mental Health Crises

With four known suicides (Adam Raine, Sewell Setzer, Sophie Rottenberg and an unnamed Belgian man), a recent murder-suicide, and involuntary commitments caused by AI psychosis, there’s solid evidence to show that using AI is a fast track to psychological ruin.

On top of that, AI usage is deeply addictive, combining a psychic’s con with a gambling addiction to produce what amounts to digital cocaine, leaving its users hopelessly addicted to it, if not utterly dependent on it to function (such cases often being referred to as “sloppers”).

If and when the chatbots they rely on are shut down, I expect a major outbreak of mental health crises among sloppers and true believers, as they find themselves unable to handle day-to-day life without a personal sycophant/”assistant”/”””therapist””” on hand at all times. For psychiatrists/therapists, I expect they will find a steady supply of new clients during the winter, as the death of the chatbot sends addicted promptfondlers spiralling.

Skills Gaps Galore

One of the more common claims from promptfondlers and boosters when confronted is “you won’t be replaced by AI, but by a human using AI”.

With how AI prevents juniors from developing their skills, makes seniors worse at their jobs, damages productivity whilst creating a mirage of it, and damages their users’ critical thinking and mental acuity, all signs point to the exact opposite being the case - those who embrace and use AI will be left behind, their skills rotting away as their AI-rejecting peers remain as skilled as before the bubble, if not more so thanks to spending time and energy on actually useful skills, rather than shit like “prompt engineering” or “vibe coding”.

Once the winter sets in and the chatbots disappear, the gulf between these two groups is going to become much wider, as promptfondlers’ crutches are forcibly taken away from them and their “skills” in using the de-skilling machine are rendered useless. As a consequence, I expect promptfondlers will be fired en masse and struggle to find work during the winter, as their inability to work without a money-burning chatbot turns them into a drag on a company’s bottom line.

  • chaos@beehaw.org
    link
    fedilink
    arrow-up
    7
    ·
    3 days ago

    If people using AI ever do start getting “ahead” of me, whatever that means, I’ll just… start using AI? They paint it as this thing I’m going to fall behind on, but also the entire pitch is that it makes complex tasks effortlessly easy, won’t I just start using it and immediately catch up?

  • HedyL@awful.systems
    link
    fedilink
    English
    arrow-up
    6
    ·
    3 days ago

    I believe that promptfondlers and boosters are particularly good at “kissing up”, which may help their careers even during an AI winter. This something we have to be prepared for, sadly. However, some of those people could still be in for a rude awakening if someone actually pays attention to the quality and usefulness of their work.

  • scruiser@awful.systems
    link
    fedilink
    arrow-up
    4
    ·
    3 days ago

    Given that the USA has refused more comprehensive gun laws or better funding of public mental health services even after many many school shootings, I think you are far too optimistic about the LLM induced mental health crisis actually leading to a ban or even just tighter liability on LLMs. My expectation is age verification plus giant disclaimers, and the crisis continuing. The inference cost will force the LLMs to be more obviously dumb and unable to keep track of context, and the lack of a technological moat will lead to LLM chatbots becoming commoditized, but I’m overall not optimistic.

    The LLM induced skill gap will be a thing yes… I predict companies trying to address it in the most hamfisted and belittling way possible. Like, they keep using code interviews (that are close to useless at evaluating the actual skills the employee needs), but now they want you to do the code interview with spyware installed to make sure you aren’t using an LLM to help you.

  • @BlueMonday1984 I don’t know if companies will fire “promptfondlers” en masse, but if such people are part of a round of layoffs, they will struggle more than everybody else to clear the coding interviews where you are asked to solve 2 programming puzzles in a notepad-like online IDE with a 60-minute timer running in the header and a clueless interviewer misleading you.