News Former Google CEO says climate goals are not meetable, so we might as well drop climate conservation — unshackle AI companies so AI can solve globa...

DS426

Upstanding
May 15, 2024
254
190
360
This idea of "throw everything we have at AI now and it will solve the problem before it's too late" notion is ridiculous. We don't need AI to tell us that we need to start changing things now to result in even lower impact down the road, i.e. many of our impactful systems have a lot of inertia. Moreover, AI isn't going to fix politics, so little wins need to be taken wherever they can be.

Kind of tangential but this over-reliance on AI will just be speeding up idiocracy.
 

kkthebeast

Distinguished
Nov 24, 2015
8
5
18,515
He is right, we cannot do it ourselves (if you even agree it needs to be done..., nor is even controllable by humans) either way the solution will be via advancements in science/engineering/methodology/material sciences and not via social means. There is nothing you can do to fix the problem without destroying QOL or freedoms. The solution WILL ONLY ever happen via innovation, optimization, atmosphere tech scrubbing on a tera forming level. Anything less is a smooth brain emotional response and an Utopian ideology at the absolute BEST. "IMHO" lol.
 
  • Like
Reactions: jp7189

watzupken

Reputable
Mar 16, 2020
1,175
660
6,070
And how does he think AI will help here? AI is a collection of human knowledge scrapped and gathered from whatever sources they can to train the it. So I am not sure what about AI that can meaningfully solve climate issues. Even if the AI comes out with a plan, it also requires commitment to get there, At the end of the day, you make a plan, you try and press on to see where you will land. You may not get there, but at least you may make some improvement. Instead, his idea is that you can't make it, so that's go the other way and make it worst. This is a seriously dumb comment.
 

Notton

Commendable
Dec 29, 2023
865
764
1,260
That dude is so out of touch with reality.

Climate change is not a binary solution, there are varying degrees of intensity.
If we can't hit 2.5C, then our next goal should be 2.6C.

Although, seeing as how this guy thinks, it's no wonder Google dumps the baby with the bathwater from their product stacks.
Okay, I thought the obituary was shorter, that's a lot of things Google has killed.
https://killedbygoogle.com/
 
  • Like
Reactions: KyaraM

jp7189

Distinguished
Feb 21, 2012
512
298
19,260
And how does he think AI will help here? AI is a collection of human knowledge scrapped and gathered from whatever sources they can to train the it. So I am not sure what about AI that can meaningfully solve climate issues. Even if the AI comes out with a plan, it also requires commitment to get there, At the end of the day, you make a plan, you try and press on to see where you will land. You may not get there, but at least you may make some improvement. Instead, his idea is that you can't make it, so that's go the other way and make it worst. This is a seriously dumb comment.
He not talking about LLMs. AI (machine learning) is one of the hottest topics for climate modeling. Bigger datacenters = bigger models = more accurate and faster simulations. That means you can test solutions faster and with more accuracy to help choose the best path forward in the real world. ..and he's right that that's ultimately more important than if you put the wrong type of plastic in the wrong bin.
 
What's happened is that the tech industry has stumbled onto this major cash-cow that has the unfortunate side effect of consuming staggering amounts of energy at a time when energy consumption and carbon footprints are getting more and more attention. And these companies don't just have to consider justifying how much energy they're using now: they're thinking about the amounts they're going to have to justify using in just a few years' time, while other companies and private consumers are getting leant on to use less. (That's also why he tries to claim it's self-regulating because "No large company wants to have a huge power bill.", when he knows full well that if a company is guaranteed to make $2 profit for every $1 spent on power, it'll be happy with the biggest power bill possible.) The game here is to try and avoid governmental restrictions or financial punishments on their energy consumption.

One way is to try and lump all the AI in together, good, bad and frivolous. Make it sound like all this energy is being used to do good stuff, but who knows how it divides up? ChatGPT alone reportedly consumes in a year enough energy to power an entire European country for a day or two. Is that sort of energy also being expended on protein folding models? Seems unlikely. And people aren't asking ChatGPT how to solve the climate crisis, they're asking it to help with their homework or recommend a TV show to watch or restaurant to go to. When Schmidt talks about "unshackling AI", I'm very sure he's not advocating leaving LLMs and deepfakes out of that.

So he makes out the huge catch-all of "AI" is part of the solution to climate change, not part of the problem, so leave us alone to keep doing whatever we want please. Of course the only answers people are going to be interested in are going to be ones that don't cost them more money or require them using less energy. I'm not sure AI can work out that miracle of arresting climate change pain free. But as long as the money's rolling in, its not Big Tech's immediate problem if nobody likes the answers.

I wonder what Schmidt's reaction would be if part of AI's solution to solving climate change was turn off all the non-scientific AI? Try and bury it, probably.