Re:Yann LeCun and Andrew Ng: Why the 6-Month AI Pause Is a B...
by Karrot_Kream on Friday April 07, 2023 @01:53PM
(#35484641)
While I don't agree about wanting a winter, I feel your pain. I'm personally not that annoyed by people hyping up what they're building, because building stuff takes at least some effort. I'm more annoyed at the high temperature discussion surrounding AI which makes it hard to talk concretely about what new ML models can actually get done. I wonder if there's any solution to the problem of the hype cycle.
Re:Yann LeCun and Andrew Ng: Why the 6-Month AI Pause Is a B...
by akiselev on Friday April 07, 2023 @01:58PM
(#35484704)
> high temperature
The discussions around here are a lot more interesting with a higher GPT temperature.
Re:Yann LeCun and Andrew Ng: Why the 6-Month AI Pause Is a B...
by verdverm on Friday April 07, 2023 @02:20PM
(#35484988)
if it gets too high, it turns into gibberish, maybe we just need an e-drug for AI... They had them in Nier Automata, which by the way is a fantastic story and commentary (even if written prior) given the GPT conversation
When I'm reading on AI news and articles the good stuff drowns in the hype stuff peddling wildly extrapolated capabilities that supposedly will change the world overnight.
And watching some GPT application headlines on HN it's hard to distinguish what is actually remarkable product or research idea from an overhyped this-can't-actually-ever-work-like-this thing.
I forgot where I read this but I think there's a proverb that I suspect applies to the current hype: people overestimate impact of new technology on short term but underestimate long term.