Settings

Theme

$10M says we won't see human-superior AGI by the end of 2025

garymarcus.substack.com

14 points by g-w1 · 13 comments

Reader

6 threads
esafak

The bar is this:

"...whatever smart intellectual labor any human working on their own, even the best human in a given domain, can do will be beaten by Al by end of 2025. That is what I am challenging."

I don't see ML models doing superhuman science in a year and half so I would take that bet. They can't even write better prose than the best humans, and language models are their strong suit.

TillE

Generously, Musk has never been serious about making predictions. When he says stuff like this he just means he's excited about it.

  • threeseed

    To be more cynical.

    He tends to do it whenever there is bad news in other parts of his business.

  • Calvin02

    > When he says stuff like this he just means he's excited about it.

    That's an overly charitable interpretation of what he's doing.

    The pattern that I've observed is: 1) he's a narcissist who always wants to be in the news and/or 2) he's trying to distract from something.

    He can't stand it when the press moves on from him. His MO is to wait for a week or two and then throw out something outlandish to suck up the media oxygen.

    He and Trump are very similar people.

    • kiba

      Having read the biography, he's a man who likes and is attracted to drama, regardless how much it burns him.

    • pennybanks

      ima guess most billionaires are narcissists. and this media oxygen stuff i mean it works right. to do able to get attention of the press and that many people is a only reserved for a very small amount of people on this earth.

      • Calvin02

        I guess one way to get data on that would be to ask people “Outside of tech and Buffet crew, how many billionaires can you name?”

      • kiba

        You mean the most well known billionaire are more likely to be narcissists.

jjcon

To be fair, I've never met a human as knowledgeable as ChatGPT before but I've simultaneously never met a human as dumb as ChatGPT before.

jrflowers

I too would like to get paid for having to read whatever stuff the Twitter guy posts online that somehow ends up being displayed as news

morninglight

There will never be a human-superior AGI until we agree on what that means.

For at least 80 years that has changed weekly and continues to do so.

.

aftbit

Yeah but what about by 2035? What's the saying? Something like "people overestimate how much will change in 5 years but underestimate how much will change in 10"

Keyboard Shortcuts

j
Next item
k
Previous item
o / Enter
Open selected item
?
Show this help
Esc
Close modal / clear selection