Because as we know, the security industry isn't a firehose of fear about everything that isn't airgapped (and then airgapping itself). Security recommendations come so thick and fast that even security researchers themselves don't bother following them. Even the bloke in your link explicitly apologises for using Word for some things, and he's considered a saint in the industry.

What all the commenters here doing a hatchet-job on Torvalds are missing is that he's saying "show me the money"/"perfect is the enemy of good". The hatchet-jobbers are too busy doing the usual tittering over his language to actually absorb the point.

Security industry != crypto academia.

Clearly, Linus didn't know at the time. If he'd know, he would have chosen another hash in 3 seconds —no additional complexity, no additional effort involved in not choosing a hash the security community starts to have doubts about.

Don't forget the extra computation cost with a more complex hash.

Nobody cares. Hashes are plenty fast. Even the code complexity doesn't matter: I've implemented both SHA-256, and SHA-512, take less than a hundred lines of code. And of course, Linus would have use some existing implementation.

Also, "more secure" doesn't necessarily mean "more complex", or "slower". Blake2b for instance is as fast as md5 and has a simple RAX (xor, rot, add) core from Chacha20.

It wasn't like that in 2005.

Which is largely irrelevant on desktops. Nobody's using git on embedded devices, so hash computational cost shouldn't be the deciding factor. Maybe people storing BLOBs will notice.

There are blobs in the kernel and speed was a very important motivation when kernel switched to git.

So now we need to assess whether hashes are a significant bottleneck.