The FTC was seeking an injunction, which has been denied. That means the acquisition can go ahead, and any further action by the FTC would be up against a fait accompli.
The FTC has appealed the ruling on the basis of nothing. They have filed a notice of appeal but not given grounds. Presumably there will be a reason given at some point, but they need to think of it first.
There is only one company in the world (ASML based in The Netherlands and probably the most important company in the world now) that makes the most advanced equipment for producing silicon chips, and they're also a key supplier even for less-advanced devices. So this not only prevents China from making chips on advanced processes of 7nm and below, it will cripple the country's ability to produce chips at 14nm over time. It already has machines since they were not previously restricted, but now it can no longer buy more or procure replacement parts.
That pushes them back to 28nm (the 20nm node sucked except for Intel's proprietary version) and 28nm when TSMC is ramping up 3nm will not get you anywhere.
The restrictions also hit flash memory, DRAM production, and logic chips like microprocessors.
Can China build its chipmaking tools? Sure. When not sure, it could be a few years or two decades. Even if they steal the designs, which they probably already have, they currently need the factories to make the parts to make the machines to make these machines.
How did they achieve this miracle? They trained phi-1 using textbooks rather than on the internet.
What does it means? It means you can produce an AI that is smart enough to perform simple tasks and small enough to run on your laptop or, probably, your phone.
What else does it mean? It means to score 85% on that test using the same approach as GPT-4 you'd need something like 2 quadrillion tokens, which would cost billions of dollars to train even if you could find that much data. And then years to "align" to get it to stop giving obviously wrong answers because you stuffed it full of nonsense.
The second paper highlights two problems with the first.
1. - 4% of the problems in the test set cannot be solved with the information provided, or in some cases, at all:
>Below you are given the delays for the different gates you were permitted to use in part D above. Compute the propagation delay of your circuit from D.
That's the entire question. There is no part D above, yet the claim is that GPT-4 answered this question correctly. There are many questions like this in the test set - this second paper links to a spreadsheet with the complete list of questions, good and bad.
2. - the answers provided by GPT-4 are scored by GPT-4. If GPT-4 tells GPT-4 that GPT-4 got the question wrong, GPT-4 gets to try again indefinitely.
Supposedly the answers were verified manually, but if so, they did a poor job because they missed all the wrong questions.
3. - not included in the paper, but posted today on Twitter, the original code used to run the tests leaks the answers used for verification by GPT-4 to the GPT-4 instance answering the questions.
It's a lot. Most of the big default subreddits, which, to be fair, are all troon-ridden shitholes. The default subreddits that new users are subscribed to will be going private so that new users can't access them at all.
Which will improve the site, but I don't know if Redditors will see it that way.
This story came out a couple of days ago. Still, I held off on it because it sounded more like a morality play than the kind of accident that happens with advanced weapon testing, which tends to be loud and messy and not relegated to a presentation at the Royal Aeronautical Society.
And there were two issues with the way this story was first reported:
1. It was a simulation rather than a live test of actual hardware.