News for hackers. Stuff that matters.
Re:The Problem with AI Welfare by iDont17 on Thursday June 12, 2025 @10:43AM (#44258410)
Why make AI then if intelligence is electrical activity in a substrate is everywhere.

We’re engineering nothing novel at great resource cost and appropriation of agency.

Good job we made the models in the textbook “real”?

Wasted engineering if it isn’t teaching us anything physics hadn’t already decades ago, then. Why bother with it?

Edit: and AGI is impossible… ones light cone does not extend far enough to accurately learn; training on simulation is not sufficient to prepare for reality. Any machine we make will eventually get destroyed by some composition of space time we/the machine could not prepare for.

Re:The Problem with AI Welfare by vonneumannstan on Thursday June 12, 2025 @11:12AM (#44258747)
This is a strange position.

>Wasted engineering if it isn’t teaching us anything physics hadn’t already decades ago, then. Why bother with it?

Why build cars and locomotives if they don't teach us anything Horses didn't...

>and AGI is impossible… ones light cone does not extend far enough to accurately learn; training on simulation is not sufficient to prepare for reality. Any machine we make will eventually get destroyed by some composition of space time we/the machine could not prepare for.

This could be applied to human's as well. Unless you believe in some extra-physical aspect of the human mind there is no reason to think it is different than a mind in silicon.

Re:The Problem with AI Welfare by tomrod on Thursday June 12, 2025 @10:48AM (#44258470)
AGI may not be impossible. But next token prediction won't get us there.
Re:The Problem with AI Welfare by vonneumannstan on Thursday June 12, 2025 @11:13AM (#44258755)
It's actually really unclear that this is true. If you brought GPT-o3 back to 1990 I have a hard time believing the world wouldn't immediately consider it full AGI.
Re:The Problem with AI Welfare by ACCount36 on Thursday June 12, 2025 @12:14PM (#44259421)
If you told a person from 1990 that in the year 2025, they have this thing, and described OpenAI's o3 - strengths, flaws and all? That person would say "yep, your sci-fi future of year 2025 has actual AI!"

But if someone managed to actually make o3 in year 1990? Not in some abstact sci-fi future, but actually there, available broadly, as something you could access from your PC for a small fee?

People would say "well, it's not ackhtually intelligent because..."

Because people are incredibly stupid, and AI effect is incredibly powerful.

Re:The Problem with AI Welfare by vonneumannstan on Thursday June 12, 2025 @12:29PM (#44259642)
I'm very confident that if someone in 1990 used o3 they would be absolutely astonished and would not pull the 'well actually' thing you think they would.
Re:The Problem with AI Welfare by ACCount36 on Thursday June 12, 2025 @12:39PM (#44259764)
Nah, AI effect is far too powerful. Wishful thinking of this kind is simply irresistible.

In real life, AI beating humans at chess didn't change the perception of machine intelligence for the better. It changed the perception of chess for the worse.

Re:The Problem with AI Welfare by iDont17 on Thursday June 12, 2025 @11:37AM (#44259004)
dang is a little bitch who can mock others emotional sensitivity like a sanctimonious prick but censors those who do the same to his world

Fucking coward