I still fear for what AI training will cost (financially and ecologically). The outputs also seem like a force multiplier that's more likely to be used for bad than good, at least without better guardrails. And it doesn't seem to make people any better, aside from a narrow view of productivity.
Hopefully Ed is wrong. Or at least there are more articulate and methodical skeptics who can keep us grounded.
>And it doesn't seem to make people any better, aside from a narrow view of productivity.
This could be said about almost any new technology. Spreadsheets, word processors, nearly any tech startup.
People who use LLMs daily generally feel their lives are better because of them. Yes, including the non-"4o cultist psychosis" types.
As for harms: thoughtful AI worriers and doomers have been trying to sound those alarms for decades, but AI skeptics generally shoot it all down because it would require accepting what "hype" and "boosters" say about likely future capabilities, or something like that.
Thanks for the link. You make some good points.
I still fear for what AI training will cost (financially and ecologically). The outputs also seem like a force multiplier that's more likely to be used for bad than good, at least without better guardrails. And it doesn't seem to make people any better, aside from a narrow view of productivity.
Hopefully Ed is wrong. Or at least there are more articulate and methodical skeptics who can keep us grounded.
I think those are all reasonable worries, and many critics do a better job than Zitron of articulating them.
>And it doesn't seem to make people any better, aside from a narrow view of productivity.
This could be said about almost any new technology. Spreadsheets, word processors, nearly any tech startup.
People who use LLMs daily generally feel their lives are better because of them. Yes, including the non-"4o cultist psychosis" types.
As for harms: thoughtful AI worriers and doomers have been trying to sound those alarms for decades, but AI skeptics generally shoot it all down because it would require accepting what "hype" and "boosters" say about likely future capabilities, or something like that.