Cal Newport

Like they showed all these exploits they found.

They didn't count on a lot of security researchers said, well, wait a second.

Why don't I get like a much smaller, cheaper model aiming at that same source code and say, can you find any vulnerabilities?

they could find the same ones.

So the evidence that it's finding vulnerabilities better, we don't have any way of knowing that's true.

And if anything, we actually are getting a lot of reports that they were paying big bounties for security researchers.

I'm going to give you access to mythos.

I'm going to pay you

for any bugs you can report that you found with it.

So they had security researchers, just who knows how many false positives were coming out of that.

And then on the exploitation side, we only really have one study.

It comes from AISI, who I do not trust, but it's the only independent study.

The fact that they gave them access itself should make us maybe a little bit suspect, but it basically just showed like normal progression, no massive leap.

Model by model gets a little bit better on some of these tests and benchmarks, and Mythos has no out-of-scale leap.

It's just like on some, it's about the same.

On some, it's a little bit better.

And yet it got covered as if we had just turned on Whopper from the movie War Games.

Like we had just some new entity that was like on its own undermining security.

And I do not think that – I think that was highly credulous coverage of what almost certainly is just like a standard slight jagged move forward on these various capabilities that we've been seeing for the last three years.

That is the number of people who take the stairs when there is also an escalator available.