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

State of AI co-author: we must slow down the race to god-like AI

ft.com

Post reply on HN
Print view View original

11–20 of 52 posts

Re: State of AI co-author: we must slow down the race to god-like AI

#11

Post by worrycue »

ChatGTP wrote: post #9
synaesthesisx wrote: post #6

No - absolutely not. Slowing down innovation in the face with vague “existential risk” fearmongering is absurd. We should be pursuing the opposite, and accelerating efforts to build what will be some of the most transformative technology humanity has ever developed (and some could argue, NOT building has its own existential risks).

Do you have any solutions to the (admittedly theoretical ) existential problems that will likely arise ? I see the problem more like, a fail is an all or nothing fail, that’s it, everyone dies, maybe including the ASI (goes rogue and destroys itself and us), then what? Good outcome for anyone if we don’t understand what we’re doing? I’m not saying you’re entirely wrong, or we need to stop, but I mean, we do share a p…

> Do you have any solutions to the (admittedly theoretical ) existential problems that will likely arise ?

I guess we can develop super AI then ask it to solve it. :P

Re: State of AI co-author: we must slow down the race to god-like AI

#12

Post by ChatGTP »

worrycue wrote: post #11
ChatGTP wrote: post #9

Earlier quoted context omitted.

Do you have any solutions to the (admittedly theoretical ) existential problems that will likely arise ? I see the problem more like, a fail is an all or nothing fail, that’s it, everyone dies, maybe including the ASI (goes rogue and destroys itself and us), then what? Good outcome for anyone if we don’t understand what we’re doing? I’m not saying you’re entirely wrong, or we need to stop, but I mean, we do share a p…

> Do you have any solutions to the (admittedly theoretical ) existential problems that will likely arise ? I guess we can develop super AI then ask it to solve it. :P

I mean it’s what any “intelligent” species would do right ? :)

Re: State of AI co-author: we must slow down the race to god-like AI

#13

Post by Llamamoe »

uejfiweun wrote: post #5

Here's the thing, current computational power isn't enough to achieve the scale of model that would be required for "god like" AI. So isn't that kind of a natural brake, where we have to wait for computing speed to catch up with our ambitions? Why do we need additional handicaps and slowdowns on top of that? I've seen a theory that the big tech companies are the ones pushing for regulation in order to achieve regulat…

> I've seen a theory that the big tech companies are the ones pushing for regulation in order to achieve regulatory capture

This... would track with what's historically been the case. Damn, that's an eye-opening take, thank you.

Re: State of AI co-author: we must slow down the race to god-like AI

#14

Post by nullsense »

worrycue wrote: post #11
ChatGTP wrote: post #9

Earlier quoted context omitted.

Do you have any solutions to the (admittedly theoretical ) existential problems that will likely arise ? I see the problem more like, a fail is an all or nothing fail, that’s it, everyone dies, maybe including the ASI (goes rogue and destroys itself and us), then what? Good outcome for anyone if we don’t understand what we’re doing? I’m not saying you’re entirely wrong, or we need to stop, but I mean, we do share a p…

> Do you have any solutions to the (admittedly theoretical ) existential problems that will likely arise ? I guess we can develop super AI then ask it to solve it. :P

Then it can lie to you about a plan to solve it and get you right where it wants you.

Re: State of AI co-author: we must slow down the race to god-like AI

#16

Post by cuteboy19 »

ChatGTP wrote: post #9
synaesthesisx wrote: post #6

No - absolutely not. Slowing down innovation in the face with vague “existential risk” fearmongering is absurd. We should be pursuing the opposite, and accelerating efforts to build what will be some of the most transformative technology humanity has ever developed (and some could argue, NOT building has its own existential risks).

Do you have any solutions to the (admittedly theoretical ) existential problems that will likely arise ? I see the problem more like, a fail is an all or nothing fail, that’s it, everyone dies, maybe including the ASI (goes rogue and destroys itself and us), then what? Good outcome for anyone if we don’t understand what we’re doing? I’m not saying you’re entirely wrong, or we need to stop, but I mean, we do share a p…

It would really be funny if this were ChatGPT itself asking these questions

Re: State of AI co-author: we must slow down the race to god-like AI

#18

Post by verdverm »

flangola7 wrote: post #10
uejfiweun wrote: post #5

Here's the thing, current computational power isn't enough to achieve the scale of model that would be required for "god like" AI. So isn't that kind of a natural brake, where we have to wait for computing speed to catch up with our ambitions? Why do we need additional handicaps and slowdowns on top of that? I've seen a theory that the big tech companies are the ones pushing for regulation in order to achieve regulat…

> Here's the thing, current computational power isn't enough to achieve the scale of model that would be required for "god like" AI. I don't see how that is the case at all. The human brain only needs about 100 watts. A GPT-3 instance only needs a few GPUs and maybe 1000 watts. GPT-4 is presumably in the same ballpark, let's say 10000 watts to be generous (even though it's probably closer to 1000 too), but still smal…

Simply scaling up the current LLM models will not fix the fundamental problems of hallucination and lack of planing abilities. If they cost $10M+ to train, will scaling that cost 100X (to $1B) really make sense. I find it doubtful that the ROI is there

Consider that GPT4 is 15-30X the cost of GPT3.5, is it really 15-30X better? Not in my experience. I would expect that we are approaching the point of diminishing returns with text based LLMs, at least for the current architectures and training regimes. I'm interested to see what the Large Multi-Modal Models produce, but I suspect the planning & lying will still be there.

Chaining these models will increase the cost even more, without certainty that it will produce usable results for anything beyond simple, context limited tasks.

Post reply on HN