dibujaron

say you have a suspect, somewhere in a building, with a gun. If you send in a cop to the building, and the suspect points their gun at the cop, the cop must shoot the suspect, and if they shoot they must shoot to kill. So now the suspect is dead.

if you send in a mobile robot first, then you don't have to send in a human to scout. You can find the suspect and assess the situation without having to put a human at risk (and therefore in a situation where they have to shoot to defend themselves). This helps both the cop and the suspect remain alive.

"Blue collar workers and service industry workers aren't in immediate danger of being replaced by AI"

I wouldn't be so sure about this; robotics is advancing rapidly. I do think there'll be a substantial market for "human-made" products and services as luxuries, though, so at least some humans will retain these jobs.

The OP could've done more to explain it, but it really is a fascinating subject. Saba is an extremely steep island which, for most of its history, had no point of access except for a set of 800 stone stairs called "the ladder". everything brought to the island had to be carried up the ladder on foot, including large items of furniture. Given this, construction equipment wasn't an option.

excel can edit csvs, definitely, no problem at all. It will lose any formatting and column size adjustments if you save back as csv, but that's expected.

I'm not an expert but I think this sort of thing is relatively traceable for two reasons. One, datacenters are difficult to conceal. Two, the supply chains for many of the relevant materials are difficult to conceal. Some of those supply chains still require western components, I believe, so if you know how much of X component was sent to china, you know how much compute they have.

North Korea is geopolitically useful as a buffer state between the United States' sphere of influence in South Korea, and China. China has defended it pretty determinedly, historically.

This article makes Odin sound extremely well-known. I've never heard of it before, and I feel like I keep up with programming topics pretty diligently. Admittedly I don't work at the systems programming layer, but I've definitely heard plenty about Rust and c++ topics.

Curious if others feel similarly, or maybe I just happened to miss it?

I read recently that corporate taxes used to be a lot higher on large conglomerated companies, which used to deter monopolization somewhat. Under this automatic-stake idea, it'd be interesting if the government's stake in corporations increased as they got larger? Companies try to avoid this and so don't combine so much?

Of course once the government does have a large stake in the largest monopoly-like companies, it's immensely motivated to keep them large. Hmm. This idea isn't good.

If it is, would that bad? Seems like a person who might really have strong personal investment in the situation. Using the oil companies' profits to try to shrink them, seems good.

If it's actively building the next generation of itself, I'd say that counts. It's more like a parent raising their kid well than it is like a parent modifying their own mind, but the result is still that you have a better model in a year than you do now.

a heightmap is three-dimensional, where the third dimension is usually represented with color or contour lines.

Does this strike anyone as bad api design to anyone else? I don't know what their definition of "much" is and it may not be the same as mine. behind the scenes they must have a real number like page_visits_in_last_30_days, which they're checking against a threshold value or "much". It'd be much more useful to just go ahead and return page_visits_in_last_30_days.

(to be fair, maybe the api does offer this too and it's just not the focus of the article)

It's hard to know who was right. All of these things can be true: it made it back ok; it had a high chance of making it back ok; it should've had a much higher chance of making it back ok. Most of the concerned people were stressing this last point, that it should've been safer than it was. They still thought it had a quite high chance of making it back ok. It took a lot of shuttle missions before Columbia failed.

A less cynical explanation: It's heavily trained to ask follow-up questions at the end of a response, to drive more conversation and more engagement. That's useful both for making sure you want to renew your subscription, and also probably for generating more training data for future models. That's sufficient explanation for the behavior we're seeing.