It depends on what the account is posting prolifically. For example, I’ll happily read silence7’s interesting links any time, but Patnou just asks weird questions over and over and I blocked them after a while. Weird questions may produce more engagement in the form of comments, but that’s a flawed metric because overall it’s just not interesting to see in the feed.

throwing away your old ICE car

But most people wouldn’t send their old car to the scrap heap, they would sell it on the secondary market to someone else (or a dealer, who would auction it into the secondary market). The old car would then continue to burn gas for likely many more years, until it “falls apart” anyway.

Stepping back, your old car may be the first in a chain of older (or more falling-apart) vehicles getting traded out, all the way down to one that really does get fully retired, or replacing one that was totaled in a collision. So choosing to keep it deprives someone else of its availability and thereby drives up used prices slightly.

For any study of this type of net effect, the authors need to pick a boundary for what gets considered. How many secondary market transactions do you study in that replacement chain, and what do those buyers substitute when the original ICE vehicle is not replaced with an EV? How far do you go in the pollution and other supply chain effects of manufacturing a new vehicle? I didn’t read a machine translation of your Swiss link, so I don’t know where the study authors drew the bounds, but I suspect it’s easy to choose and defend framing that supports either conclusion.

In my personal calculation, I can only see one step down the used chain, wherein my old vehicle would continue to be driven by someone else, so replacing it with an EV wouldn’t make a substantial difference. I love my old car with no surveillance, so I’m in no hurry to switch even though I’ll presumably buy an EV eventually.

Ultimately this is just one more example that that’s no ethical consumption within capitalism.

I’m kinda looking forward to the collapse of the Ai bubble.

Once upon a time, a lot of telecom bubble startups went under and left behind lots of dark fiber optic cable that others snatched up for far less than the capital cost of construction. But eventually, over subsequent decades, most of it was put into use. I expect these AI data centers will be similar, except a bunch of them are being built in stupid ways like with their own on site electricity generation.

I’m not that excited about what people will do with all of the super cheap bankruptcy auction surplus (formerly AI) data compute capacity. But at least it will be a novel problem.

If LLMs don’t turn into AGI and instead the AI bubble pops, maybe some of the data centers don’t get finished. Even so, many of the rest will likely continue to operate and consume water.

The US federal government will soon start decreeing how the overdrawn Colorado River will be resolved and things will get very ugly. Western water rights law is an unsustainable construct and all real solutions hurt rich people (even as they also hurt everyone else). Maybe the current administration is waiting until enough of the data centers are built before “solving” the water crisis by prioritizing AI over crops and humans in the name of national security.

But we’re still living in a bad B movie. If AGI actually works and we get SkyNet in the end, then the water infrastructure in most of those regions is probably plenty for the machines once all of the humans are gone.

Autoclanking

First off, this was a shitpost that only one person managed to figure out. LLMs are garbage and the whole premise here is fairly ridiculous.

I agree that he owes us nothing. But I’m sure he owes the money people at Bantam or Random House something. Like actually, contractually, owes them. They will complete it posthumously if needed, using his draft material. It’s incumbent on him to finish it while he still has creative control.

Finally lol

Funny, that’s very interesting.

And hey, it looks like AI might wind up a pretty savvy tool for writers wanting to keep track of their own complex plot maps and character webs, too.

See that’s exactly the kind of housekeeping help that might actually boost GRRM’s personal productivity without compromising his vision. I’ve never had to wrangle thousands of pages of my own narrative but I can imagine it’s inherently really slow.

GRRM deserves every insult for being unfocused and putting other work ahead of his task. Sanderson may be the one to finish extending his piles of chapter drafts instead of Claude, but someone is definitely going to pick it up after GRRM posthumously cedes creative control because he can’t be assed to finish it himself.

That ending was dreamed up by Weiss and Benioff who were desperate to crap out something, anything, so that they could go make Star Wars movies instead. It’s not the real ending by the real author, and the TV series had already diverged from the books by that point.

Vyvanse or Adderall

He has claimed to have already written lots of the subsequent material, he just doesn’t like how it’s organized. Having an LLM refine that unpublished material very different than asking it to project all of the storylines itself.

Given the mountain of unpublished work, if he has a heart attack tomorrow, someone else is going to refine it and publish it posthumously. There’s too much money at stake for that not to happen.

Actually I think the problem is that he cares way too much. He has written himself too many sprawling storylines that all need to come back together in a satisfying convergence, which is very difficult. The further he gets in the story, the harder it becomes to continue. He’s also a perfectionist.

He’s a bit of a jerk for chronic underestimation of what it will take him to finish anything, and possibly for taking on other projects that further delay work on his magnum opus. But I understand procrastinating one important task by throwing yourself into work on different tasks.

I know but there are two more left and the penultimate one just might make it.

It depends somewhat on how you’re planning to watch it all. If you’re sailing the high seas or watching blurays you already bought years ago, just go until it starts to feel dumb again. But if you are going to pay real money or spend streaming subscription time, don’t watch any of it. Weiss and Benioff did such a bad job in handling the end that the entire series should be considered financially toxic. Not a penny, not a groat.

A better plan would be to listen to the audiobooks and hope that some residuals will help GRRM focus his ADHD long enough to finish another book while he’s still alive.

There are almost certainly several such instances, gobbling up content and feeding it to LLM training, ignoring all deletes.

You might enjoy The Ministry for the Future, by Kim Stanley Robinson. It’s fiction, and written a few years ago, but it begins with heatwaves and crumbling infrastructure in India and you will probably get something out of it.

@whiskey_pete@dod.mil gets up to all kinds of hijinks.

Not to discount the usefulness or complexity of curl in any way, but it’s not one of the larger codebases out there. It’s also pretty darned good. Firefox seems to have a very positive experience with Mythos, but they also had their own internal test harnesses from prior work, ready to utilize LLM analysis at scale. It was far more intensive than having a third party run something on their behalf and produce a report.

Language isn’t even shown as an option through Voyager. I don’t know about other third party clients.