As a first approximation, the answer to all "why aren't people talking about X" questions is the same: because there are a lot of potential topics to discuss, people are busy, and no-one, including you, has written a post about it yet. If you want to discuss X, it is generally better to just write a post about X, rather than a meta-post about why no-one is discussing X.

[Also EAs have discussed this a bunch! Just not on the forum.]

I can’t speak for others on the EA Forum and I imagine many would disagree, but just speaking for myself, I don’t trust Anthropic on this and would want to see confirmation from other, independent and more reliable sources.

Mainly, my thoughts are: 1) is this even true, at all, in the first place?^[1] And 2) even if it is true or partially true, to what extent is Anthropic exaggerating the significance of this?

Just one reason I don’t trust Anthropic: in mid-March 2025, Anthropic’s CEO Dario Amodei predicted that by mid-September 2025, 90% of all code would be written by AI. This didn’t end up happening, obviously. But Dario didn’t admit he was wrong. He tried to deny it. That makes Dario lose trust and credibility, in my books. If Anthropic can’t even be trusted to acknowledge clear, straightforward facts that are publicly known, how could it possibly be trusted on something like this?

1. ^{^}

   Edited at 11:15 PM Eastern on 2025-11-16 to add: Just to be clear, I would be very surprised if the whole story was an outright fabrication by Anthropic. It's more that I don't trust their judgment and I don't trust them not to exaggerate or distort things. 

This is a slightly orthogonal response but I'd love it if people wrote 'discussion thread' posts where they explain a bit of relevant context, ask a question and kick off a discussion. 

Maybe it's just me, but this looks like a win for Anthropic. Bad actors will do bad things, but I wonder why they would choose to use Anthropic instead of their own Chinese AI, where I would assume the security is less rigorous, at least to their own state actors, no? I had Claude quickly dig this up for me, and from what he said, it occurred as far back as mid-September 2025, which would indicate this release had intentional timing. Anthropic chose to announce during peak AI governance discussion, framing it to emphasize both the threat and defense value of their systems. The delay between September detection and November announcement allowed them to craft a narrative that positions Claude as both the problem and the solution, which is classic positioning for regulatory influence. Nothing wrong with that I suppose...?