[HN Gopher] Why I'm Betting Against the AGI Hype
___________________________________________________________________
Why I'm Betting Against the AGI Hype
Author : flail
Score : 27 points
Date : 2025-12-01 17:09 UTC (5 hours ago)
(HTM) web link (www.notesfromthecircus.com)
(TXT) w3m dump (www.notesfromthecircus.com)
| FrankWilhoit wrote:
| "...a philosophical confusion about the nature of intelligence
| itself...."
|
| That is how it is done today. One asks one's philosophical priors
| what one's experiments _must_ find.
| JuniperMesos wrote:
| Interesting that in an article entitled "Why I'm betting against
| AGI hype", the author doesn't actually say what bet he is making
| - i.e. what specific decisions is he making, based on his
| prediction that AGI is much less likely to arise from LLMs than
| the probability the market is implicitly pricing in suggests.
| What assets is he investing in or shorting? What life decisions
| is he making differently than he otherwise would?
|
| I say this not because I think his prediction as stated here is
| necessarily wrong or unreasonable, but because I myself might
| want to make investment decisions based upon this prediction, and
| translating a prediction about the future into the correct
| executions today is not trivial.
|
| Without addressing his argument about AGI-from-LLMs - because I
| don't have any better information myself than listening to
| Sutskever on Dwarkesh's podcast - I am somewhat skeptical that
| the current market price of AI-related assets is actually pricing
| in a "60-80%" chance of AGI from LLMs specifically, rather than
| all the useful applications of LLMs that are not AGI. But this
| isn't a prediction I'm very confident in myself.
| karmakaze wrote:
| Armchair commentary.
|
| > I've listened to the optimists--the researchers and
| executives claiming [...]
|
| Actually researchers close to the problem are the first ones to
| give farther out target dates. And Yann LeCun is very vocal
| about LLMs being a dead end.
| klysm wrote:
| He is starting a business that depends on them being a dead
| end
| techblueberry wrote:
| Sounds like he's putting his money where his mouth is.
| nomel wrote:
| > farther out target dates
|
| And, that's why there's so much investment. It's more of a
| "when" question, not an "if" question (although I have seen
| people claim that only meat can think).
| arisAlexis wrote:
| Same guy that predicted LLMs couldn't do something in 5000
| years and they did it next year? (Google this, seriously)
| m463 wrote:
| I don't think there's a lot of "AGI hype".
|
| I think all the hype is more about ai replacing human effort in
| more ambiguous tasks than computers helped with before.
|
| A more interesting idea would be - what would the world do with
| AGI anyway?
| fragmede wrote:
| Hire digital employees rather than human ones. When all your
| interaction is digital, replacing the human on the other end
| with a theoretically just as capable AI is one possibility.
| Then, have the AI write docs for your AI employee, spin up
| additional employees like EC2 instances on AWS. Spin up 30 to
| clear out your Trello/Monday.com/Jira board, then spin them
| back down as soon as they've finished, with no remorse, because
| they're just AI robots. That's what you could do with such a
| technology anyway.
|
| That's for regular human-level AGI. The issue becomes more
| start for ASI, artificial super intelligence. If the AI
| employee is smarter than most, if not all humans, why hire
| humans at all?
|
| Of course, this is all theoretical. We don't have the
| technology yet, and have no idea what it would even cost
| if/when we reach that.
| arisAlexis wrote:
| Can't you think what a world with a species smarter than humans
| could be like? Yeah, it's difficult
| drpixie wrote:
| Summary of the current situation...
|
| LLMs have shown us just how easily we are fooled.
|
| AGI has shown us just how little we understand about
| "intelligence".
|
| Standby for more of the same.
| arisAlexis wrote:
| Contrarianism as a mental property of humans
___________________________________________________________________
(page generated 2025-12-01 23:02 UTC)