I think OP's point is that, if the attacker already has root, they haven't gained any new capabilities from the CVE. They could just do whatever they wanted much more directly. Therefore, the inherent risk seems fairly low.

Yes if that was an accurate description of the supposed vulnerability then I would agree that it's bogus, but we're filling in a lot of blanks here and something tells me there's more to it than that.

But they do that via things in their control. For example, banks that don't follow FATCA cannot access the US banking system. It's rare for the US to send law enforcement to banks that don't comply

The UK would have been perfectly justified in saying 4chan can't access any of their systems.

He's the "doing ui in c to piss of react devs" guy, right?

Ooh that's interesting, can you point me to the exact videos?

that's literally what it's called on YouTube

tsoding is

Because human review is a serious bottleneck and optimizing something that isn't the bottleneck isn't helpful

Which makes sense.

In a conversation about code, the most likely answer is often some comment about backwards compatibility.

In a conversation about a legacy enterprise app, the most likely answer is often going completely insane

I still think this is because, on a new release, it works on some prompts the previous ones did badly at, because new weights do well on a different set of prompts.

Then after a few days you notice the prompts that it does badly on that the old ones did fine with and everyone is convinced there's a regression when it's just a different part of prompt space

I'd put Kevlin Henney up there with him. His talk on 1968 alone is pretty amazing

Sure, but the new patterns are several years old at this point.

Not adapting is a choice

So that you can name a section of code without polluting the namespace.

It's training data might have a ton of examples of people hurrying and screwing more after being yelled at