dboreham

That's the literal definition of a loyalty test.

The other two don't actually make chips though, so there's that.

I've been counting inbound packets with options set on my edge routers for a couple of days. So far: zero.

As a business owner I get these surveys. In times past I'd spend a bunch of time responding, on the theory it's my duty to help the government gather accurate statistics. Now I delete the emails on arrival since the government is under the control of goose-stepping imbeciles who wouldn't know good data if it smacked them in the head.

Also CISA emails. Full Orwell.

Also the court follows laws. There are numerous cases where weird stuff happens with supreme court rulings only because congress never passed any law regarding whatever it is.

I think you misunderstand the direction of the trade.

No. Which may mean they don't know.

I've never heard any such claims but meanwhile in my actual job I'm seeing unbelievable productivity gains (in terms of actual cash spent before and after vs the same kind of work being done).

Well I think we can replace economists working for PE companies with AI now!

It's amazing to still see posts like this in late 2026. I suggest you recalibrate your ideas on what both AI and human intelligence is and can do. My take: both are trained systems. Both are as "statistical" as the other. Any idea that "because statistics" means that humans are different or better is bogus. Both can do things based on their training data. You are totally correct that a human has access to training data that AI doesn't (the smell of new mown hay, for example, or the fear of being crushed by the garbage compactor room in Star Wars). But there's no "humans have more training data therefore better" argument to be made about fields like mathematics. In fact the opposite argument works quite well: AI can have all mathematics ever invented in its training data, something that's pretty hard to achieve as a human.

I won't even start on "AI is unconscious"...

Just to add a data point: we added an edge router rule to drop inbound IP packets with options set. Interestingly a) that rule has a hit count of zero so far and b) it was somewhat tricky to test because it turns out every host I could think of to run an nping command from is on a network that filters the packets anyway. Or perhaps our upstream provider is doing the filtering. I had to send packets from a machine where the path didn't go through anyone else's router in order to see hit counts increasing and thereby test the rule.

Sony Pictures is still in business.

Based on email I got from another big hosting provider, they're all patching for the same checksum metadata vulnerability (the CVE they mentioned is the same). There is however a root RCE PoC according to the researcher so...potentially bad.

Current models are amazing but production code incorporates unknowable information through decades of diagnosis and bug fixing (and sometimes the opposite: working around bugs in other software that it talks to). Put even the most amazing coder that was ever born onto improving a complex old codebase, and you can guarantee they are going to break stuff. AI is better but not infinitely better.

I think we can all agree that the task as set forth in the PR puff piece is impossible. That said however, current AI tools are amazing for cleaning up and reanimating old codebases. I've been using Claude on a very large Java protocol proxy product and it's been able to do amazingly useful things for me. I originally wrote much of the code but Claude now knows much more about it than I do today, because people forget stuff over decades. It will sometimes remind me "as you noted in November 2016 this feature was designed to...", based on its analysis of the code and git commit messages.

It's good, but it can't create information that is undiscoverable.

Indeed. Making a perfect clone of such software is I suspect provably impossible because you have no way to assemble all the test cases. There might be some peer box out in the wild somewhere that is the only thing that will exercise some code path. You have no idea what that box is or what it does.

See also: https://en.wikipedia.org/wiki/Purple_hat_therapy

Like it was written by PR people's AI.

Plus: publicity!

No but their no doubt highly paid lawyers should have sought relief from their trademark being infringed, and demanded the extension was pulled long ago.

No, you misunderstand (not surprising since 99% of the coverage on this story is wrong). What happened is: someone published an extension called "Anthropic.claude-code". That someone was not Anthropic and never was Anthropic. They just used the string "Anthropic" in the name of their extension. The comment you are responding to is saying that Cursor (the organization that manages the extension marketplace) should have prohibited someone from using a popular AI lab's name in the title of their mystery meat extension. They're probably right about that.

No. You'd need to know how Binance worked in some very deep level of detail. E.g. that there's a way to force signing of high value transactions with no rate limiting or countersignature, then how to mount an attack on that code through lax controls via a developer machine. That's not something you learn through casual scanning.

And almost certainly an inside job given the knowledge of Binance systems required.

Binance is a bank.

Yes, but as someone who worked in the cryptocurrency industry for a few years: that's not how it actually works. Smart contracts have bugs, it isn't peer to peer so someone has to run and pay for big servers, humans are sometimes criminals or Faustian bargain takers.

All cryptocurrency is permissioned. This is because governments can make it illegal for someone to turn your cryptocurrency into real currency. You need their permission to turn it into pizzas and yachts.

Thanks. Always wanted to know what the secret Styrofoam material in US thermonuclear weapons was...