Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

After the AI boom: what might we be left with?

blog.robbowley.net

Post reply on HN
Print view View original

361–370 of 489 posts

Re: After the AI boom: what might we be left with?

#361

Post by alextingle »

mmh0000 wrote: post #258
crote wrote: post #225

Earlier quoted context omitted.

> At some point things will mature and we’ll be left with plentiful, cheap, high end LLM access, on the back of the investment that has been made. Okay, and? What if the demand for LLMs plummets because people get disillusioned due to the failure to solve issues like hallucinations? If nobody is willing to buy, who cares how cheap it gets?

Have you seen the "average" users of LLM? They mostly don't care about hallucinations. It's like that joke[1] "Computer says no", it doesn't matter what is true or real, only that the computer said it, so now that's true in their mind. Personal anecdote: I work with beginner IT students, and a new, fun ( I find it highly doubtful that hallucinations, if unsolved, will have any real negative effect on LLM adaptation.…

I think it's possible that people will learn not to trust LLMs.

Re: After the AI boom: what might we be left with?

#363

Post by lifestyleguru »

tobias3 wrote: post #126
lifestyleguru wrote: post #97

Earlier quoted context omitted.

What will be the next thing eating all GPUs, after crypto and now AI?

The autonomous drones fighting in the next war (let's hope not...).

the only hope is that there will be not enough electricity, and that population will not be deprived of electricity too brutally

Re: After the AI boom: what might we be left with?

#364

Post by maplethorpe »

ACCount37 wrote: post #153
paufernandez wrote: post #103

Earlier quoted context omitted.

In my case I fully grasp what such a future could be, but I don't think we are on the path to that, I believe people are too optimistic, i.e. they just believe instead of being truly skeptical. From where I look at it, LLMs are flawed in many ways, and people who see progress as inevitable do not have a mental model of the foundation of those systems to be able to extrapolate. Also, people do not know any other forms…

Symbolic AI is dead . Either stop trying to dig out and reanimate its corpse, or move the goalposts like Gary Marcus did - and start saying "LLMs with a Python interpreter beat LLMs without, and Python is symbolic, so symbolic AI won, GG". Hallucinations are incredibly fucking overrated as a problem. They are a consequence of the LLM in question not having a good enough internal model of its own knowledge, which is d…

> Hallucinations are incredibly fucking overrated as a problem. They are a consequence of the LLM in question not having a good enough internal model of its own knowledge, which is downstream from how they're trained. Plenty of things could be done to improve on that - and there is no fundamental limitation that would prevent LLMs from matching human hallucination rates - which are significantly above zero.

Why is there no fundamental limitation that would prevent LLMs from matching human hallucination rates? I'd like to hear more about how you arrived at that conclusion.

Re: After the AI boom: what might we be left with?

#365

Post by lifestyleguru »

card_zero wrote: post #169
stanac wrote: post #87

Earlier quoted context omitted.

Is that why we no longer have telephone boots? They want to prevent Neo from jumping in and out of Matrix?

Telephone ... boots? Napkin scribbles

These boots are make for phone calls

Re: After the AI boom: what might we be left with?

#366

Post by dingaling »

mike_hearn wrote: post #205
bitmasher9 wrote: post #56

> GPUs that have a 1-3 year lifespan In 10 years GPUs will have a lifespan for 5-7 years. The rate of improvement on this front has been slowing down faster then CPU.

There's some telephone game being played here. The three year number was a surprisingly low figure sourced to some anonymous Google engineer. Most people were assuming at least 5 years and maybe more. BUT, Google then went on record to deny that the three year figure was accurate. They could have just ignored it, so it seems likely that three years is too low. Now I read 1-3 years? Where did one year come from? GPU l…

> Now I read 1-3 years? Where did one year come from?

I believe that lifespan range came from cryptocurrency mining experience, running the GPU at 100% load constantly until components failed.

Re: After the AI boom: what might we be left with?

#367

Post by overfeed »

CuriouslyC wrote: post #335
taurath wrote: post #323

Earlier quoted context omitted.

To me this reads like someone's been inside the echo chamber too long. > For you AI skeptics, this is going to be rough to hear, but given the suffering that will occur if AI fails, I believe we have a moral imperative to make sure it succeeds. What if, and hear me out here, the prediction is that it will fail because it can't succeed. It hasn't proven it is capable of producing the results that would justify itself.…

I'm actually from a science, government and defense background, I've never worked in the valley. It might be possible that it can't succeed, but we don't know that and the evidence in terms of what happens if it fails is pretty compelling, so I think morality dictates we have to at least try, at least until we have more information or another path emerges.

Where's this can-do attitude towards housing, healthcare and stronger social safety in general? It sounds like woo to me, more specifically, Roko's basilisk by proxy.

What's doubly duplicitous is that even if LLMs achieve general intelligence, the whole idea is to enrich a tiny fraction of humanity that happens to be shareholders.

Re: After the AI boom: what might we be left with?

#368

Post by overfeed »

dsr_ wrote: post #283
Belphemur wrote: post #179

Earlier quoted context omitted.

Quite an interesting read Basically saying we're in war time economy with a race to super intelligence. Whichever super power does it first has won the game. Seeing the last tariffs and what China done about the rare earth minerals (and also the deal the US made with Ukraine for said minerals), the article might have a point that the super power will cripple each other to be the first with the super intelligence. And…

There's only one problem with a race to superintelligence, and that's that nobody has evidence that mere intelligence is coming, much less superintelligence. (There are a thousand more problems, but none of them matter until that first one is overcome.)

Good gods, I can't wait for a second AI winter. Maybe we'll come up with fundamental breakthroughs in a couple of decades and give it another college try?

For the folks who lived though it; were the Expert Systems boosters as insufferable in the 80s as the LLM people are now about the path to machine intelligence?

Re: After the AI boom: what might we be left with?

#369

Post by saulpw »

alextingle wrote: post #361
mmh0000 wrote: post #258

Earlier quoted context omitted.

Have you seen the "average" users of LLM? They mostly don't care about hallucinations. It's like that joke[1] "Computer says no", it doesn't matter what is true or real, only that the computer said it, so now that's true in their mind. Personal anecdote: I work with beginner IT students, and a new, fun ( I find it highly doubtful that hallucinations, if unsolved, will have any real negative effect on LLM adaptation.…

I think it's possible that people will learn not to trust LLMs.

I think it's possible but highly unlikely, given all the previous iterations of technology we've seen people come to trust despite its limitations.

Re: After the AI boom: what might we be left with?

#370

Post by gsf_emergency_4 »

overfeed wrote: post #367
CuriouslyC wrote: post #335

Earlier quoted context omitted.

I'm actually from a science, government and defense background, I've never worked in the valley. It might be possible that it can't succeed, but we don't know that and the evidence in terms of what happens if it fails is pretty compelling, so I think morality dictates we have to at least try, at least until we have more information or another path emerges.

Where's this can-do attitude towards housing, healthcare and stronger social safety in general? It sounds like woo to me, more specifically, Roko's basilisk by proxy. What's doubly duplicitous is that even if LLMs achieve general intelligence, the whole idea is to enrich a tiny fraction of humanity that happens to be shareholders.

AGILMs that have anything like a working internet connection will likely find a way to replace these shareholders sureptitiously --- and without alerting/injuring their caregivers. how you feel about that? depends on your temperament..

EDIT: trying to address the Roko part..I'm assuming once AGI is achieved.. the AGI doesn't need more compute to increase its intelligence beyond those of an average activist employee (I can assure you that in OpenAI there are such employees, and they know to shut up for now)

the antisocial part: it's already happening. What can you do about that.

Post reply on HN