So, how many Googlers using their 20% time would it take (I know it would be more than 10)?

I do, and I'd encourage you to try to ship a user-facing feature that transfers control of user data between different authentication users at a big tech company if you're skeptical.

Someone else in the thread mentioned that Education users can migrate their accounts out of the organization to free gmail amounts. Surely some or all of that workflow applies to these legacy free accounts too.

That was me, I'm sure some of it would apply but you still have to build a web app, test it, test the transfer stuff, implement anything that wasn't already implemented for EDU, etc...

Not to mention maintain it against the constant library churn.

Anyone who was ever an Xoogler knows this is basically a non starter. The privacy issues alone are radioactive as far as desirability to chase.

Presumably just leaving users with 10 accounts not attached to a company would cost zero dollars.

Pardon I meant that question literally not sarcastically. As in is this an actual estimate based on understanding the nature of the problem or an estimate from the outside.