Hacker News Reader

Comment by JKCalhoun | Hacker News Reader

> The project entails more than 2 gigawatts of computing capacity—Zuckerberg said it could eventually expand to 5 gigawatts—programmed to train open-source large language models.

Given that the human brain takes much longer to "train", I wonder how the energy efficiency pans out — comparing the two.

View on HN

How long does a human brain take to train?

Biological systems are wildly energy efficient, that's kind of their whole thing. The average human will consume approximately 75kwh worth of calories in their lifetime. There are electric cars with bigger batteries.

[Edit] ok, yes, please. I get that i missed the k in kcal. The point stands. Biological training is massively more efficient, even when you forget to multiply by 1000

You're off by about three orders of magnitude.

A human consuming 2000 kcal/day (conservative estimate) uses about 2.32 kWh per day. Over 75 years, that's roughly 64,000 kWh.

Oh, right i did a conversion wrong. Woops. In any case, a rounding error when talking about gigawatts of generation capacity

We're efficient once we have the energy, sure. How much energy does it take to go from raw sunlight to a calorie your body is actually able to use, and finally to your dinner table?

All of our food was alive before we ate it. All calories used by living things are efficient. Life is an end unto itself. It does not need to justify its existence by the moral code of technocrat materialism. The fact that this discussion is being had on this board in good faith is morally condemning of our worldview.

Since the original point of this chain was a comparison between the energy efficiency of biological vs machine learning, then we need to be trying to understand if the machine is more efficient than the human. You don't need to make some moral or philosophical argument about existential justification to accept that taking a more efficient approach is better, in that it generally enables more life for the same energy.

If the true, total cost of a machine to perform some task is less than a person to do the same task, then the machine should do it and the person should move to do what the machine cannot. This means more energy is available for everything else, living included.

This is wrong by at least three orders of magnitude. Very roughly, a human requires 2000 kcal a day = 2 kWh a day so 75 kWh is enough to cover about a month, putting aside the upstream losses in the energy supply chain (which are far greater for humans).

In general, saying that biological systems are "wildly efficient" is... wildly wrong. Some biological processes are optimized by evolution... most are not. There are no bicycles in nature.

Your forget that a biological system has approximately 0 throughput in work done.

Nearly everything a biological system accomplishes depends on massive external machinery.

Humans are only intellectually interesting because of their use of tools.

> The average human will consume approximately 75kwh worth of calories in their lifetime. There are electric cars with bigger batteries.

Doesn't pass the smell test. I think I could push an electric car at least a mile a day if that's what I spent most my extra calories on. If I did that I'd surpass its range in well under 2 years, much less than my lifetime.

Reply on news.ycombinator.com