I’m not disagreeing with you. I’m just stating that a broken unplayable game objectively has no value. The publisher has forced that value to 0 if they turn off their servers without support, regardless of if there was any value there before or not.

Edit: I realize we might be talking about different things when saying “stop supporting”. I meant that to mean when the servers are turned off, not when they stop releasing updates or delist it from stores.

But that assumes that the (live service) game loses value after the company stops supporting it

Well yeah. Obviously the game losses value BECAUSE it’s not being supported anymore. There’s no value in a paperweight.

My understanding is that this would force games to be sold as either a good (lasts forever) or a service (lasts a specific, advertized amount of time). It does not prevent service games from existing, it just stops them being sold as goods with an unspecified expiration date. The problem is consumers are uninformed about the lifetime of the game they are purchasing.

AI definitely would have done better. This looks like a child drew a smiley face on a lump of clay by poking their fingers in it.

I find it fascinating how this glint from the sun appears multiple times at the top edge. The gaps get smaller towards the edge. There’s probably some beautiful math formula for this involving the golden ratio.

These “euro-kkkolonizers” were all several generations ago. Maybe you want to pretend nothing has changed, but things have gotten significantly better since then.

I won’t pretend we don’t still have problems. People of color are still statistically lower income, and they’re still affected by all the same capitalist problems that come with that.

The problem is also not the same across the country. Every state has their own top issues.

I’m not talking about history. I’m talking about the US today.

China also kinda just forces anyone out who’s in the way. To build any new infrastructure the US ends up getting slowed down to a crawl because of red tape and beurocracy. Land owners have a lot more rights in the US.

Thanks for the links. I was able to find the original source for that claim, which has actually usage numbers: https://iea.blob.core.windows.net/assets/18f3ed24-4b26-4c83-a3d2-8a1be51c8cc8/Electricity2024-Analysisandforecastto2026.pdf

0.3Wh / request for Google 2.9Wh / request for ChatGPT

That does however reference the same paper as your linked articles, which I can’t find without a paywall: https://www.sciencedirect.com/science/article/abs/pii/S2542435123003653?dgcid=author
I’d love to know how they came up with that number for ChatGPT, but it looks like I was a bit off with my estimates regardless. There’s probably some scaling efficiencies they’re taking advantage of at that size.

I’m not sure that’s even a valid comparison? I’d love to know where you got that data point.

LLMs run until they decide to output an end-of-text token. So the amount of power used will vary massively depending on the prompt.

Search results on the other hand run nearly instantaneously, and can cache huge amounts of data between requests, unlike LLMs where they need to run every request individually.

I’d estimate responding to a typical ChatGPT query uses at least 100x the power of a single Google search, based on my knowledge of databases and running LLMs at home.

Damn, $1500 is actually a great price for a 4090. If only…

Edit: I don’t know how I got here. I just realized this post is 4 months old.

Pretty sure at that range they’re dead no matter what. May as well crack one last joke.

You can search for them, but I think the one I have is similar to a Uni Kuru Toga pencil. I don’t write as much as I used to, but it’s awesome for taking notes.

Mechanical pencils for the win! Did you know there’s even ones that rotate the lead for you as you write so there’s always a sharp point?

Logic and reasoning were never on the table unfortunately.

This graph actually shows a little more about what’s happening with the randomness or “temperature” of the LLM.
It’s actually predicting the probability of every word (token) it knows of coming next, all at once.
The temperature then says how random it should be when picking from that list of probable next words. A temperature of 0 means it always picks the most likely next word, which in this case ends up being 42.
As the temperature increases, it gets more random (but you can see it still isn’t a perfect random distribution with a higher temperature value)

It’s what Microsoft would do in the same situation. It’s only fair

That does seem to be the case. As long as any modifications to the source are publicly available. Which is pretty reasonable.

It’s in the RSALv2:

You may not make the functionality of the Software or a Modified version available to third parties as a service

You may not X in a way that Y implies that You may X in a way that does not Y, and is more specific (and changes the meaning of the license) vs You may not X

The legal distinction in this case allows for distributing the software for example as source code, but not as a service.