SnapDragon
10hr ago
Eventually, most of the "real" challenges that humanity faces will be, at least in my opinion, rendered obsolete. That leaves just about only games to pass the time. They can be complicated games, they might be of relevance to the real world (status games, proof of work or competence), but they're still games we play because we've run out of options. I think this isn't a thing to complain about, once we get there. Our ancestors struggled to survive so that we wouldn't have to.
Forget "eventually"; I think we often fail to appreciate that we're already there, in the first world. Almost none of the "challenges" that our primitive ancestors faced are in any way familiar to us. They worried about whether they would starve next winter; I wonder whether I can justify being lazy and ordering Door Dash today. They might have been permanently crippled from an uncleaned surface cut; I would slap a band-aid on it and take a Tylenol. They banded together and learned to fight so the next tribe over wouldn't kill them all and take their stuff; I put my money into a stock brokerage.
Aging is IMO the one major challenge that hasn't been conquered yet (although we're still living twice as long as evolution intended). In almost every other way we're living the lives of Gods.
SnapDragon
15d ago
If you're going to lean so heavily on your credentials in robotics, then I agree with @rae or @SnapDragon that it's shameful to come in and be wrong, confidently and blatantly wrong, about such elementary things such as the reasons behind LLMs struggling with arithmetic. I lack any formal qualifications in ML, but even a dummy like me can see that. The fact that you can't, let's just say it raises eyebrows.
False humility. :) I have ML-related credentials (and I could tell that @rae does too), but I think you know more than me about the practicalities of LLMs, from all your eager experimentation and perusing the literature. And after all, argument from authority is generally unwelcome on this forum, but this topic is one where it's particularly ill-suited.
What "expertise" can anybody really claim on questions like:
- What is intelligence? (Or "general intelligence", if you prefer.)
- How does intelligence emerge from a clump of neurons?
- Why do humans have it and animals (mostly) don't?
- Are LLMs "minds that don't fit the pattern", or are we just anthropomorphizing and getting fooled by ELIZA 2.0?
- If yes, how does intelligence emerge from a bunch of floating-point computations?
- If no, what practical limitations does that put on their capability?
- What will the future upper limit of LLM capability be?
- Can AI experience qualia? (Do you experience qualia?)
- Does AI have moral worth? (Can it suffer?)
With a decent layman's understanding of the topic, non-programmers can debate these things just as well as I can. Modern AI has caused philosophical and technical questions to collide in a wholly unprecedented way. Exciting!
SnapDragon
15d ago
Thanks for calling OP out on his flagrant errors. It's one thing to make a technical mistake on a non-technical forum; it's another thing entirely to flex, claim industry expertise, and then face-plant by confusing word embedding models with LLMs. I hope people aren't being misled by his, well, "hallucinations". (Honestly, that's an appropriate word for it! Incorrect facts being stated with complete confidence, just like an LLM.)
SnapDragon
18d ago
I'm reminded of the Obamacare debacle, which still fills me with rage. People (correctly) pointed out that women pay more for health insurance, and (incorrectly) said that this was an unfair "woman tax". It was politically brilliant, reframing the fact that women live longer as a societal injustice - against women! And it was 100% successful; Obamacare made gender-based pricing illegal, and now every man in the country is subsidizing the health care of every woman in the country. Forever.
Yup, just came here to mention XKCD. Gotta love the emdash in the disclaimer, too!
More options
Context Copy link