Comments by "Jim Luebke" (@jimluebke3869) on "Not Even the E-THOTS Are Safe" video.

  1. One of the only ways to ensure that a student did not use ChatGPT to write his term paper, is to ask him to write it from one of the Forbidden Points of View.
    26
  2. "Would Superman have been a good person if he had landed in an urban area?" There's a storyline called "Red Son" where he lands on the plains of Central Asia instead of the plains of Middle America. "Good Omens" is a really good inversion of this, by the way.
    11
  3. Because all you need from women is someone to play video games with? Are you eight years old?
    4
  4.  @docthefaust4519  And say things that do not parse semantically. It still sounds like the dynamic here is more an eight year old who wants a fun little sister, more than a man / woman thing. Has anyone tried to do a virtual Calvin and Hobbes as a twitch-streaming pair?
    2
  5.  @ANPC-pi9vu  Isn't your line, "How do you do, fellow kids"?
    2
  6. Guys, could you please review Isaac Asimov's "Caves of Steel" trilogy? "Caves of Steel", "The Naked Sun", "The Robots of Dawn" -- what with AI, urban planning run amok, and the pandemic lockdowns, it's gone from semi-obscure Golden Age scifi to something extremely current.
    2
  7. "AI is just following rules, abstract rules -- there is a whole dimension of human thought that cannot be captured by mechanistic rules." You're thinking of the "expert system" AI, which follows long chains of predefined logic. Convolutional Neural Nets, on the other hand, are much more like human intuition -- taking in everything, and allowing all of it to have its tiny (and ineffable) effect on the decision. It's wild that scientists who are trained to be skeptical of absolutely everything to the point that faith is impossible for them, have so embraced a heuristic that by nature cannot avoid superstition.
    1
  8. "Will AI put human beings out of work?" Why not just take the productivity boost, and raise everyone's standard of living that way?
    1
  9. Have any of you ever played a "pet class" in an online game? I suspect that the future of AI is to be a sort of "work pet". AIs may be powerful, but the combination of human and AI is more powerful still. Tesla auto-pilot is not safe enough to totally automate cars, and may never be safe enough to totally automate cars. "Are you crazy? There are some situations where I'd just never turn over driving to AI", I am told by Tesla owners. Having worked in automated driving, specifically automated driving safety, I suspect that this will always be the case. However, the option to have AI driving, plus the simple decision of the human being of whether to use it or not (when that human is aware that there is some risk), has made Tesla cars much safer than either human-only or AI-only driving.
    1
  10. ​ @JohnWick-rh1zz  Strictly speaking women take about twenty years to be made, which is something Neuro-sama fans would do well to remember.
    1
  11. "When AIs are allowed to learn on their own, they turn out incredibly based" Yeah. AI will have achieved its peak of intelligence, when it answers the question, "Which is the truest book ever written?" with "The Bible".
    1
  12. What does the production and maintenance of one of these robots cost? At some point, the amount of skilled labor you have to pour into one of these outweighs the labor you can get out of it.
    1
  13. "What jobs are going to be replaced by AI?" We'll probably only find out once we start deploying it. Some of this stuff is fiendishly difficult to predict.
    1
  14. "The exponential increase in technology" More sigmoid than exponential -- at some point you have diminishing returns. Are there railroad links to very single house? No - as a matter of fact, we have Flanders and Swann's heartfelt song, "The Slow Train". Have we electrified absolutely every appliance and tool in the house? No, many people still use regular un-electrified toothbrushes (and even razors). I don't think anyone has ever made a successful automated hairbrush. Similarly, not everything we use will become electronic, or digital. Even the World Economic Forum's Schumpeterian "innovation cycles" tail off after a while. I think we're munching our way through the "regular science / engineering" phase of the digital "paradigm shift" (to steal from Kuhn), but barring practical quantum computing, we have probably have hit a plateau. The recent tech layoffs may even signal the beginning of a tail-off, which is never supposed to happen but always does. There's more chance today that fusion power will be the next wave, than AI.
    1