Hacker News Reader

Comment by tintor | Hacker News Reader

"... AI will avoid climate change."

xAI's Colossus 1 and 2 data center mobile gas turbines beg to differ.

AI data centers are driving electricity consumption in 100s of GW, and most of that energy is not clean renewable nor nuclear.

View on HN

That doesn't seem like an actual quote in the original article, but this a fair point to bring up and a reminder that Gates has been problematic on his views on climate.

First, in the essay he mentions "Using AI, low-income farmers will soon be able to get better advice about all these things than even the richest farmers get today and increase their output substantially.". There may be some truth in this, but it's worth reminding that this is more of a 'productivist' worldview, and historically productivity gains in agricultural practices have benefitted farmers that were already wealthy (see books like The Globalization of Wheat by Marci Baranski [1] on the Green Revolution where "Borlaug’s new technologies ultimately privileged wealthier farmers, despite assurances to politicians that these new crops would thrive in diverse geographies and benefit all farmers").

Gates has also been criticized in the past for being influenced by figures like Bjorn Lomborg, where he had a (now-deleted) essay [2] with him, and there's a decent thread by energy entrepreneur Jigar Shah on Gates being off on climate issues for 15 years [3]

[1]: https://upittpress.org/books/9780822947349/

[2]: https://web.archive.org/web/20230601175906/https://www.gates... (June 2023)

[3]: https://bsky.app/profile/jigarshahdc.bsky.social/post/3m4del...

The Green Revolution most certainly did improve the lives of tens of millions of low income farmers in India. The high yield crops also permanently removed the spectre of famine from India, saving tens of millions from starvation.

To retcon this into “only rich farmers benefited” is weird. Certainly there are many rich farmers out there. Of course their incomes increased when they could plant more productive crops. But so what if they make money? They’re feeding everyone!

I encourage you to do a deeper dive into the sources of famine in India - specifically impacts of Churchill in WW2 just before the Green Revolution. See https://www.cnn.com/2019/03/29/asia/churchill-bengal-famine-... (based on https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2018... ) along with this documentary https://www.youtube.com/watch?v=T31MFNLJKJw . While there are drought induced famines, I still wonder what other approaches could be done (ranging from land-use changes, etc.)

Also ask yourself what the tradeoffs were in terms of biodiversity losses and agricultural diversity losses due to Green Revolution, and areas where it didnt work out like the attenpts of it on Balinese subak rice agriculture practices (the book 'Priests and Programmers' should have more info)

Why would you assume I don’t know about the Bengal famine? I know it intimately. But it’s really not relevant to the discussion of whether more productive crops are better for everyone, farmers and consumers alike.

It’s really weird to attempt to school me on Indian history too. “Ask yourself” indeed. If you have something to say on the subject, say it.

I’m almost certain you have a very surface level understanding of Indian history if you’re trying to link the Bengal famine during WWII instead of talking about the very real food shortages in India post-1947.

People were dying because they didn’t have enough food to eat. India had to beg for food aid, which America provided after extorting and humiliating India. Thanks to these crops that never happened again. Biodiversity is an afterthought when children are dying of starvation and you sound very out of touch for bringing it up.

It is from the Gates video in the article.

Those emissions are nothing. Its a drop in the bucket compared with other industrial uses. A 50% reduction would change nothing, reduction to 0 would change nothing.

This is an absurd argument. Any one source is a drop in the bucket.

Any given optimization reduces running time of a program, but you should still focus on your hot loops.

That’s an argument about sequencing, not of necessity.

And as an analogy it breaks down because we need to reduce/eliminate all sources of pollution, not just the current biggest one.

Sequencing and necessity are not separable in practice. Time and resources are finite, so prioritization is mandatory. If you define everything as "necessary" and try to do it all at once, you just fail even harder.

Edit: Perhaps a more interesting question is whether we should rate pollution sources by absolute amount or amount per actual utility produced. Gas-turbine-powered AI datacenters surely rate abysmally low on utility efficiency.

The original comment I replied to states: “A 50% reduction would change nothing, reduction to 0 would change nothing.” When talking about individual data centers.

My point is that this is true of basically any since source of pollution.

You can argue about sequencing if you’d like, but every individual source must be dealt with.

And when it comes to the atmosphere and the cumulative effects of each tonne of CO2 has on the climate, it is better and cheaper to prevent adding new CO2 than it will be to remove it once it’s added.

I think international shipping produces significantly more CO2 than any single other source

Yet the AI boom relies on global shipping to be produced , so “AI” is “additive”.

But not when you look at just one ship, or just one datacenter.

Its not an absurd argument. You should think why this source of emission demands so much focus because its not emission related. If it were emission related you would be focusing on reductions in 100s of other places before bothering with datacenters.

They've got the conclusion (they want datacenters gone) and they are going searching for possible any reason.

I believe we are on autopilot towards essentially free renewable electricity, and so this doesn't feel like a big concern when held against the potential end-of-humanity potential that AI has.

Climate change in general has very little end-of-humanity potential. Unlike AI tech.

Reply on news.ycombinator.com