The problem isn’t that the energy is too cheap, it’s that there’s too much of it, which is why it’s so cheap. An electrical grid can only support so much power and there is no cost effective way to store enough energy to run the grid for any appreciable amount of time, so it all must be used or else the system becomes unstable.

Because they would have brownouts overnight and when the weather was bad.

Don’t give them any ideas

What’s the thermal impact of a ram module? Don’t they use like 2 or 3 watts even in a desktop? Can’t be much…

Because ram is incredibly cheap and developer hours are incredibly expensive. I think it’s a bit silly too but there’s just no financial incentive for companies to care about memory usage when they know most consumer devices have tons of extra headroom.

*proceeds to wrap everything in unsafe {}

Sounds like a good way to get the feds interested in your otherwise not very notable property crime.

No, motion sickness.

I read the second half of this in Heath Ledger’s joker voice

We’ll that’s when I’m on social media…

For a horrifying take on this check out this short story by qntm

https://qntm.org/mmacevedo

It isn’t, but the GDPR requires companies to scrub PII when requested by the individual. OpenAI obviously can’t do that so in theory they would be liable for essentially unlimited fines unless they deleted the offending models.

In practice it remains to be seen how courts would interpret this though, and I expect unless the problem is really egregious there will be some kind of exception. Nobody wants to be the one to say these models are illegal.

I don’t understand why I would want a bunch of usb c ports? On a phone where there obviously isn’t space for a full sized port sure, but I find that fiddling with the one usb c port on the back of my desktop is a pain in the ass and the port really struggles to keep a good connection when attached to a stiff or heavy cable.

The mentioned but unsupported link to “general intelligence” reeks of bullshit to me. I don’t doubt a modified LLM (maybe an unmodified one as well) can beat lossless compression algorithms, but I doubt that’s very useful or impressive when you account for the model size and speed.

If you allow the model to be really huge in comparison to the input data it’s hard to prove you haven’t just memorized the training set.

It’s not that it’s too hard to maintain, but that in order to make sense it would have to be cheaper than building and maintaining the solar panels on some larger and less valuable patch of land 30 minutes out of town.

As long as you can find devices which let you do local only setup…

Don’t forget consumer protections!

Uber somehow managed to convince people that it’s not their fault if their drivers don’t follow traffic laws, drive intoxicated, and assault people.

*brakes

I wonder how much (if any) leverage this gives Checo for contract renewals. If RBR let him go and he immediately sues them for discrimination it’s going to be a bad look.

That would be a very on brand thing for them to do