How much of that compute was for the ads themselves vs the software useful enough to compel people to look at the ads?
It is like an arms race. Everyone would have been better off if people just never went to war, but....
A lot of advertising is telling people about some product or service they didn't even know existed though. There may not even be a competitor to blame for an advertising arms race.
You're tilting at windmills here, we can't go back to barter.
The AI boom has completely changed that. Data center power usage is rocketing upwards now. It is estimated it will be more than 10% of all electric power usage in the US by 2030.
It's a completely different order of magnitude than the pre AI-boom data center usage.
It might help to look at global power usage, not just the US, see the first figure here:
https://arstechnica.com/ai/2024/06/is-generative-ai-really-g...
There isn't an inflection point around 2022: it has been rising quickly since 2010 or so.
Figure 1.1 is the chart I was referring to, which are the data points from the original sources that it uses.
Between 2010 and 2020, it shows a very slow linear growth. Yes, there is growth, but it's quite slow and mostly linear.
Then the slope increases sharply. And the estimates after that point follow the new, sharper growth.
Sorry, when I wrote my original comment I didn't have the paper in front of me, I linked it afterwards. But you can see that distinct change in rate at around 2020.
Yes, data center efficiency improved dramatically between 2010-2020, but the absolute scale kept growing. So you're technically both right: efficiency gains kept/unit costs down while total infrastructure expanded. The 2022+ inflection is real though, and its not just about AI training. Inference at scale is the quiet energy hog nobody talks about enough.
What bugs me about this whole thread is that it's turning into "AI bad" vs "AI defenders," when the real question should be: which AI use cases actually justify this resource spike? Running an LLM to summarize a Slack thread probably doesn't. Using it to accelerate drug discovery or materials science probably does. But we're deploying this stuff everywhere without any kind of cost/benefit filter, and that's the part that feels reckless.
"yeah but they became efficient at it by 2012!"
No, but it puts some perspective on things. IMO Google, after abandoning its early "don't be evil" motto is directly responsible for a significant chunk of the current evil in the developed world, from screen addiction to kids' mental health and social polarization.
Working for Google and drawing an extravagant salary for many, many years was a choice that does affect the way we perceive other issues being discussed by the same source. To clarify: I am not claiming that Rob is evil; on the contrary. His books and open source work were an inspiration to many, myself included. But I am going to view his opinions on social good and evil through the prism of his personal employment choices. My 2c.
Data centers are not another thing when the subject is data centers.
When people have choices and they choose the more harmful action, it hurts their credibility. If Rob cares so much about society and the environment, why did he work at a company that has horrendous track record on both? Someone of his level of talent certainly had choices, and he chose to contribute to the company that abandoned “don’t be evil” a long time ago.
There is no upside to vast majority of the AI pushed by the OpenAI and their cronies. It's literally fucking up economy for everyone else all to get AI from "lies to users" to "lies to users confidently", all while rampantly stealing content to do that, because apparently pirating something as a person is terrible crime govt need to chase you, unless you do that to resell it in AI model, then it's propping up US economy.
If there is any example of hypocrisy, and that we don't have a justice system that applies the law equally, that would be it.
And before the LLM craze there was a constant focus on efficiency. Web search is (was?) amazingly efficient per query.
I mean, buying another pair of sneakers you don't need just because ads made you want them doesn't sound like the best investment from a societal perspective. And I am sure sneakers are not the only product that is being bought, even though nobody really needs them.
