Replies

  1. This is a sample of problematic AI when used in genomic cases that rely on the accuracy of palindromic sequence detection, for example in CRISPR.

    0
  2. Richard Stallman Talks, Liberec, May 2025: About #AI/#LLM : "The most important thing you need to know about them is that you can't trust any of their output."

    0
  3. AI won't be good until:

    1. Accounting system - autonomous agents need accounting to see if they are better or worse.

    2. Clean data: need independent sources of data for purchase like AP, Reuters or encyclopedia company.

    3. Boundaries/fences - don't want 2 different AIs fighting over thermostat.

    0
  4. Is that Gemini?

    If so, I've found that riddled with errors and inaccuracies, when it emphatically insists it is right about something and ignored easily accessible, reliable information.

    0
  5. To verify that Gemini did in fact screw this up I tried it. Scroll down and the 1st few results reinforce the evidence that we are becoming more and more stupid every day.

    "What do it mean"???!!!

    How do we recover from the decline of basic intelligence, nevermind common sense, in society?

    0