“But it also takes a lot of energy to train a human,” Altman said. “It takes like 20 years of life and all of the food you eat during that time before you get smart. And not only that, it took the very widespread evolution of the 100 billion people that have ever lived and learned not to get eaten by predators and learned how to figure out science and whatever, to produce you.”
So in his view, the fair comparison is, “If you ask ChatGPT a question, how much energy does it take once its model is trained to answer that question versus a human? And probably, AI has already caught up on an energy efficiency basis, measured that way.”
Aren’t humans and biological creatures in general found to be extremely efficient with energy? Given the computing power in our brains the fact it runs on so little is amazing no?
Yes, it’s disingenuous for him to bring up all the time used for humans to evolve as well. If we’re going to go that far, we also ought to include the energy/time used by the engineers who created ChatGPT, and all the energy used by plants/animals in the evolution leading to those engineers. Not to mention all the time/energy/training of all the people who created the training data over the past few centuries.
Frankly, at that point, any human artist is more “efficient” than AI - they’re able to master their field in mere decades.
Doesn’t the human brain do what it does on like the same electricity as a lightbulb?
12 watts maybe. But there is no currents and flashes like internet bullshit images.
What kind of bulb are we talking here?
In Altman’s case it’s a dimbulb.
I really don’t think he meant it that way. Think of it like this - if I want to generate some images, my GPU will run at 100% for few minutes. If I want to play cyberpunk, my GPU will run at 100% for hours.
I think if he meant it that way he would have said that, instead of talking about the energy that humans use and particularly talking about food.