My AI/LLM predictions for the next 1, 3 and 6 years
22 points by azhenley 7 hours ago | 6 comments
  • stevage 6 hours ago |
    On the one hand, yay for being brave enough to share predictions.

    On the other, I share the author's complete lack of confidence in being able to make any accurate predictions here. Many will make predictions. Some, by chance, will be correct.

  • tedtimbrell 6 hours ago |
    > So my dystopian prediction for 2031 is that if that form of AGI has come to pass it will be accompanied by extraordinarily bad economic outcomes and mass civil unrest.

    I guess six years is the longest time horizon in the question but that "if" around AGI and its impact does a lot of work. This maybe assumes it'll happen sooner? Or, is it similar to the AI art prediction where the 6 year horizon is just a long enough period of time such that we'll see if we're headed towards AGI?

  • jll29 6 hours ago |
    The OP is anxious to predict the future because he has been wrong or late in the past, but that should simply been seen as a first attempt.

    Because there is enormous value in predicting somewhat well, it's only fair that a first round doesn't go very well. In fact, there are studies that professional economists routinely perform worse than "the ordinary person in the street" regarding their (economic) forecasts!

    It's a great idea to publish predictions and then to look back and reflect, because it fine-tunes (to use a fashionable AI term) one's ability to predict better.

    Also, check out the research by Philip E. Tetlock and co-workers for academics that have studied people's ability to predict, and the book "Superforecasters" for the story of some people who are particularly great at it (I met one MIT physics Ph.D. who works for the European Commission part of that group once).

    • wenc 5 hours ago |
      Tetlock's research tended to focus on very specific, falsifiable True/False questions rather than open-ended questions with predicted quantities in them.

      A good Tetlock question might be: "Will ChatGPT-5 be released in 2025?"

      But not: "Will we see AGI in 2025?" (too non-specific on what AGI is)

      Even this one is iffy: "Will OpenAI's valuation increase by at least 50% in 2025?" (because there's a number in the statement)

      Once you have a binary question, superforecasters can then start doing Fermi estimation to find bounds, and then initializing and updating Bayesian priors based on new information and continuous research. The answers are typically of the form, "Yes, with a 76.2% probability."

      So "superforecasters" are prediction specialists on a narrowly defined scope, but they are not prophets. The kinds of predictions that they make don't make for fun, entertaining reading.

  • jcgrillo 6 hours ago |
    If we ~randomly become aware on a 1, 3, or 6yr timescale that the billionaires are the smallest minority, and then collectively decide to redistribute their wealth equitably and their intestines uniformly, will that change your predictions?
  • csallen 6 hours ago |
    > Generative art tools are a lot of fun to try out but the lack of fine-grained control over the output greatly limits its utility outside of personal amusement or generating slop.

    This take surprises.

    Generative AI images and music are already amazing for a variety of practical and commercial use-cases. Video isn't quite there yet, but I think it's just a matter of time.

    With images specifically, I use Midjourney all the time for illustrations for articles. I'm using it to create assets for a game I'm making, too. In both cases it's saving me tons of time, and improving my creations, because there's no way in hell I would've created this art (nor paid anyone to do it for me) beforehand. The AI features in Canva are great for creating illustrations for presentations, cards, infographics, etc., that are perfect for what you're trying to go for. I'm also using ElevenLabs to create sound effects for my game, which is incredibly fun. And I've heard some AI music that's truly impressed me.