[aurareturn]

The optics are bad. Sam needs a good PR firm/coach. He also needs a Sheryl Sandberg in my opinion.

The message behind it is that yes, training a new model uses a lot of energy but the benefits are enormous once trained. I think his humans use energy too analogy likely came from him hearing messages about how training pollutes the earth. In his mind, he thinks training a new model is as useful or more useful than 3,000 people on average. Even if he is right, he shouldn't have used this analogy due to how sensitive people are right now with AI, energy, potential job loss, etc.

Having worked in San Francisco, I know how insensitive tech bros are to the outside world. Perfect example is the 2015 Airbnb ads: https://www.sfgate.com/bayarea/article/Airbnb-apologizes-for...

[palata]

> In his mind, he thinks

If people have to analyse what "he probably meant" and how "he probably thinks" because what he says is completely stupid, to me it's not a good sign.

> Even if he is right

Well he is obviously not, his analogy is completely stupid. What you try to rationalise as being "probably what he meant" cannot exactly be right or wrong, it's a bit philosophical. If the goal is to get as productive (or "useful") as possible before we as a species completely collapse, then maybe training a new model is as productive as 3000 people on average.

Now if the goal in life is that most living things are better in the long run, being productive is the opposite of what works. Again we are living a measurable mass extinction right now, that is happening orders of magnitudes faster than the famous one of the dinosaurs. We are failing at surviving as a species, and we are so good at it that we are making most other species fail to survive.

If "his philosophy" (again, the one we have to interpret because he is incapable of articulating something that makes sense) is that "the goal of our species should be productivity and not survival", then I can confidently say I disagree. But that's already giving him a lot of credit from this analogy.

[aurareturn]

He's thinking like an AI CEO. I don't think his goal is to eliminate 3,000 human births so he can use the energy to train a new model.

In his mind, he must be thinking that I can train a model that uses as much energy as 3,000 humans so that it can be used by billions of humans. He thinks it's a positive statement.

Again, the optics are bad. Read between the lines. I can see how most people are offended by his statement.