I've been wondering for many years before the first traffic camera appeared, why the toll-roads aren't enforcing the speed limits automatically. The time you enter and exit the highway is recorded down to a second. The distance between these two points is known — your average speed could be computed on the spot even with the early 90-ies technology...

The polite police officers would be standing right behind the toll-booths issuing tickets without the drama of hiding in the bushes, then chasing you at highway speeds...

And, yeah, you could lower it by stopping at a rest area — but it'd still be a tremendous disincentive to speed.

I was and continue to hope, that such universal enforcement, affecting all voters, would cause the limits to go up to reasonable figures — or even be abolished completely...

I've run into these people, they're the worst

When have you last checked your compiler's output?

produces higher quality sperm

That's really important for the sperm destined for the septic system, sure...

I don't think you understand the OPs point. LLMs are trained on everything the model creator can get their hands on. That means all Internet available open source (and many non-open source) code, including chardet. An existing AI can't perform a clean-room implementation because even if you don't show it the code, *it's already seen it*. And since the training data is encoded in the weights in a very non-trivial manner, you can't specifically remove a set of training data free the fact. You'd have to train a completely new model from scratch, with a dataset very carefully cleaned of any bits of the original project code. It's not impossible, but it would be hard, and vastly more expensive than hiring human coders to perform the clean room implementation.

It's not injecting any wealth into rural communities. It's injecting wealth into a single or a small group of large landowners, who upon receiving said wealth will immediately pack up and move to a large city somewhere and live the high life until they go bankrupt a year later.

They are just trying to coerce you into using the Hotmail web client in Edge instead of the classic Outlook app, which they no longer want to spend the money to maintain.

They can't just buy a looser fitting outfit, the outfits are regulated by the Olympic comittee. And according to the scientific study mentioned in the summary (you did read that, right?) the more fabric in the suit, the farther jumpers go. So, anything that adds material to the suit could (theoretically) provide an advantage. Are they actually doing this? Maybe. Olympic level athletes have been known to do some pretty insane things for even the tiniest advantage. I have no doubt some would try this if it provided an advantage (maybe even if they only thought it provided an advantage).

While I don't trust the CDC these days either, the numbers are for 2024, i.e. before Trump and his minions started their American fire sale. The numbers actually make Trump look pretty bad: not only due to the effects of COVID (which they claim wasn't a serious disease), but because the decrease in deaths from fentanyl shows that their entire "anti-drug" campaigns are horseshit, and that Biden-era policies were actually reducing deaths from drugs without engaging in wholesale slaughter of Venezualen fishermen or murdering nurses on the streets.

"likely to push printing toward cloud-locked, subscription-based systems"

And there we have the real reason for passing this law.

Something isn't "transformative" when it can literally spit out the original work, word for word. Which it absolutely can with LLMs, because that's how they work. All they are, at heart, is a non-linear mathematical regression to a dataset, in this case the corpus of human writing. You can actually look at them as a lossy compression algorithm (technically *all* mathematical models are compression algorithms, after all, since the entire goal is to construct a model that can reproduce the input data set using relatively few parameters. They're just rarely looked at that way). The trained model literally contains a compressed (albeit to some extent lossy) version of the training data set. In some cases, it's not even all that lossy, as ChatGPT has spat out word-for-word reproductions of NYT articles, and image generators will sometimes reproduce watermarks from the source images.

Yeah, it must be a falling birth rate. It couldn't possibly be decades of assault on education and progress on equal rights by the far right and the current administration. Not at all.

Because they have to get everyone to click through a waiver that prohibits them suing Verizon for anything that has happened or may ever happen.