TWeaK

Wales implemented it country wide. Almost all 30mph limits are now 20mph.

1 second is not 3 car lengths unless you're going really fast. Also, when pulling away, you shouldn't sit stationary and wait for the gap to establish, as by the time you're going you will be even further back. Instead, pull away at the same time as the vehicle in front, but restrain your acceleration so they still pull away from you. When they stop pulling away, you stop accelerating and you'll have more or less the correct gap.

Standard advice is usually 2 seconds in the dry, 4 in the wet - "only a fool breaks the two second rule" takes about 2 seconds to say, then "but if it pours, make it 4" brings it to 4 seconds.

And breathing has a non-zero chance of causing death.

And they got away with it Scott free.

So where is the revolving door of corporate lobbying sending him?

It explicitly doesn't.

If you have a hard copy book and someone steals it, you're not only losing out on the potential sale price of the book, but the tangible value you have already paid to produce that copy.

Say the book is $12, you get $2, the publisher gets $5 - the book store buys it for $7, and sells for $12 making $5 profit. If you steal from the book store, they've lost a potential profit of $5, but more importantly they've actually lost the $7 they already paid for it. This is what theft is about, the value of a possession taken away, not the potential value.

With a digital book, each individual copy costs nothing. It costs something to make the original, but making a copy is free. Thus the only thing you've lost is the potential profit, which arguably you wouldn't get anyway as the person didn't want to buy from you to begin with - just because they downloaded it for free does not mean they would have paid full price if a free download wasn't an option.

With theft, you have a tangible loss. With digital piracy, the only loss is opportunity to profit.

If a person studies a text then writes an article about the same subject as that text while using the same wording and discussing the same points, then it's plagiarism whether or not they made an exact copy. Surely it should also be the case with LLM's, which train on the data then inadvertently replicate the data again? The law has already established that it doesn't matter what the process is for making the new work, what matters is how close it is to the original work.

But the thing is the law has already established this with people and their memories. You might genuinely not realise you're plagiarising, but what matters is the similarity of the work produced.

ChatGPT has copied the data into its training database, then trained off that database, then it runs "independently" of that database - which is how they vaguely argue fair use under the research exemption.

However if ChatGPT can "remember" its training data and recompile significant portions of it in certain circumstances, then it must be guilty of plagiarism and copyright infringement.

US government: "Make us an app that people can use to subvert suppressive regimes."

Developer: *makes Signal*

US Citizens: "Hey, that's a useful app.." *installs*

US Government: "No, not like that!"

There was a Smarter Every Day video just published where he gave a talk to some people from NASA, some of the things he revealed and how little the people in the room seemed to know were a bit shocking. Like, no one knows exactly how many Starship launches it will take for refueling everything to get Artemis and supported hardware to the moon (the latest paper says "at least 12"), and how the ways they're looking to do are very complex meanwhile the Apollo engineers literally wrote the book on how they did it - TL;DR keep it simple, build in a ton of redundancy.

They claim it's not stored in the LLM, they admit to storing it in the training database but argue fair use under the research exemption.

This almost makes it seems like the LLM can tap into the training database when it reaches some kind of limit. In which case the training database absolutely should not have a fair use exemption - it's not just research, but a part of the finished commercial product.

But the fact is the LLM was able to spit out the training data. This means that anything in the training data isn't just copied into the training dataset, allegedly under fair use as research, but also copied into the LLM as part of an active commercial product. Sure, the LLM might break it down and store the components separately, but if an LLM can reassemble it and spit out the original copyrighted work then how is that different from how a photocopier breaks down the image scanned from a piece of paper then reassembles it into instructions for its printer?