> It's still easy for many good engineers to switch jobs for a 40% pay increase, and in ML people are making more than they ever have at large US tech companies.

Uh... I'm not sure that's the case here otherwise those folks would have been reassigned instead of laid off.

Layoffs have to be across the board / department, otherwise that would open the company up to even more lawsuits. They can’t special-case ML engineers out. This does not mean the company does not need or is aware it needs ML engineers elsewhere in the company.

> They can’t special-case ML engineers out. This does not mean the company does not need or is aware it needs ML engineers elsewhere in the company.

In the case of META, Amazon, some of the layoff target was shutting down the whole Org/Div/Product Group. In these cases, shouldn't they be able to re-assigned a few folks to another team?

But like I said, they can move these ML if they actually truly need it.

Reassigning folks would defeat the point of having a layoff, which is to be seen as having a layoff (taking difficult actions in the light of new macroeconomic conditions, etc, etc).

This tracks with my experience (WARNING: Anecdotal evidence incoming). I'm currently aggressively hiring for development roles (in particular around data). Despite the layoffs our candidate pipelines are only modestly bigger and the quality level is similar to before. Layoffs haven't really moved the market down very much except for some pretty major outliers where someone was being paid pretty far above market.