When I read part 1 of Joel's talk at Yale, I thought maybe he was referring to Phil Haack's recent post about TDD.  But, like Phil, I also got the impression later on that it was just a coincidence.

I do like something Phil said in his post about this coincidence though...

It’s no proof of correctness, but can be a proof of incorrectness when a test fails.

That is an excellent statement! I am not a TDD advocate, but I see its benefits.  Sadly, I've heard from a lot of people who seem to think TDD does prove correctness.  But, as Phil said, it doesn't.  It can only make you aware of failures against scenarios that are expected to succeed.