egregiousRac

in case anyone is curious, those comments are not a modern addition. They are a faithful reproduction of the original scanned copy of the code.

That’s the fatal flaw with AI. If you can replace the base-level workers in a skilled field, competition will lead to most places doing that. 5-10 years later, the industry has a shortage of mid-level workers.

Nah, opinions of them have been sliding for a while. The software licensing stuff is just a topic that attracts certain vultures, making coverage suddenly explode.

Intercept isn’t the right word, but what they are doing is shady. The Bing search result for “Google” has results pushed down until they are nearly off the screen. They place what looks like a Google doodle (exploiting the fact that Google replaces their logo all the time) and a search box in the middle of the screen. It tricks the inattentive user into thinking it did what they meant for it to do, while still keeping them in Bingland.

The claim that they are doing a clean-room implementation is bullshit. The only way any of these models are able to make any working code is by being trained on every bit of code that could be scraped from the internet. Unless the project you are cloning was released after the model was trained, it was trained on the code. It may be a tiny fragment of the training data, but it still saw it.

I haven’t used RCN/Astound since COVID, but they were pretty decent before then. Far better than Comcast.