d123

[d123 @dfh13571]

@Aging_Scientist @williawa @s_r_constantin Do you still believe significant longevity and bio progress can happen in a slow-takeoff world? Seems pretty hard to me if AI performance in these fields doesn't dramatically improve.

[d123 @dfh13571]

@doomslide #defundScience

[d123 @dfh13571]

@nihilunbounded slopdropped ones, so I can go tourture my AI to explain them to me

[d123 @dfh13571]

@chaitinsgoose @francoisfleuret I heard AI isn't really good at topology.

[d123 @dfh13571]

@EricCompeau @tranche_cheddar @teortaxesTex thoughts on this? Seems like a pretty articulate and consistent set of reasonings on why current AI are not intelligent.

[d123 @dfh13571]

@EricCompeau @tranche_cheddar I feel like I’ve baited you into writing several essays at this point.

[d123 @dfh13571]

@EricCompeau @tranche_cheddar I’m actually curious, is there an open problem that ai can solve and if ai solves it, you would call it intelligent?

[d123 @dfh13571]

@EricCompeau @tranche_cheddar I guess I will call current AIs ASI, artificial sub-par intelligence.

[d123 @dfh13571]

I think I agree with it; current AIs are probably not superintelligence and still have a way to go. It's just that I wouldn't call it not intelligent, like in the original post, because most people can't really write a math paper as beautifully and elegantly as a professional, even with training.

[d123 @dfh13571]

@EricCompeau @tranche_cheddar Also relevant

https://t.co/7W5yylUnim

[Jaime Sevilla @Jsevillamol]

AI can get away with terrible decisions in eg software design, because it makes up for them with relentlessness and it's ability to hold many things in context.

AI solutions in the short term are going to look terrible to experts, because it in fact doesn't need to deal with human limitations.

[d123 @dfh13571]

I'm not a mathematician, so I will shut up about whether the AI solutions to math problems are interesting or not, but one thing I think is important to keep in mind is that these things aren't replicas of human brains and thus have different sets of limitations and advantages compared to us.

[d123 @dfh13571]

@YafahEdelman @GregHBurnham I feel like at this point it’s just a bet on how much can AI accelerate biology research in the next few decades is it not? Care to explain why are you bullish on this?

[d123 @dfh13571]

@EricCompeau @nihilunbounded @doomslide @billionsmustliv @teortaxesTex @GregHBurnham @ElliotGlazer @ChrSzegedy @chaitinsgoose @littmath @AcerFur @__alpoge__

[d123 @dfh13571]

@EricCompeau Is it fine with you if I @ some people here?

[d123 @dfh13571]

@EricCompeau yeah I think I need sometime to think about all this before I can response. This is definitely a well articulated thought.

[d123 @dfh13571]

@EricCompeau and what role will ai play in mathematics and will the types of works that mathematicians change do stay the same or change in to a different form?

[d123 @dfh13571]

@EricCompeau basically how do you think mathematics will be like in the age of AI, how things will change, and how will people adopt it.