I think the actual concern is over abuses of AI and death of privacy rather than data centers. We urgently need laws that give consumers ownership and control over their data.

Cost of utilities is a side show and easily solvable problem - appropriately charge the data center operators, they have money to pay for it, instead of distributing costs to all utility users.

I should have asked why the OpenAI non profit would give those people equity instead, they are probably making the decision. Still, it's a bit of a handout. If they were just given profit shares in the public company they would almost certainly be worth far less.

I understand that Sam&co expect a high percentage of equity to be evenly divided per profit share and Sam just getting some as a signing bonus, but what is in it for the investors to give it to them? Unless there is a huge vesting period a lot of them will take early retirement and they have no legal right to equity, just the expectation.

Sam has an under the table agreement to threaten to walk if they don't hand out tons of equity?

Thank you. That is how I read it also.

Not just not accurate but wrong.

That's like saying the price of the battery in an electric car is that car's price minus the price of a comparable ICE car. No, it isn't. There are more differences than just the battery.

And yes, of course they recoup their development costs. But that doesn't mean that the OP is right in this context.

Seems like this all boils down down definitions. What does a grade mean?

If a grade means understanding the material, there's no reason every student couldn't get an A. Sure, many won't, but when we're talking about Harvard students, especially at lower-level courses, the barriers to get into the school are so high that it makes sense most students would be able to master the material.

If grades are relative to other students, even if every student understands the material perfectly there's still going to be the curve, some A's, B's, C's, and some must fail.

errr... Apple doesn't charge for its operating system. macOS literally doesn't have a price.

If that's your interpretation then you know it wasn't sincere.

It is called Model Collapse, and avoiding it is a hot research topic.

If you punish companies for firing, you get less hiring.

Countries with inflexible labor markets tend to have higher unemployment.

I moderately to heavily use AI for my work because it is capable of speeding up routine time-consuming tasks. I do that so I can use my time more effectively on other productive tasks. However, I did rough calculation and it costs about 4$/hour in tokens for that. That is subsidized costs where LLMs are offered at a loss to capture market share. True costs are easily double that. This is not trivial cost if everyone in a company starts doing that.

I know people that still expose their lives to Google, but I am not one of them. Especially now, at the start of the age of AI where all information is used to profile you and used against you, from salary negotiation to loan applications, it is absolutely crazy to want any product at any price, including free, from Google.