Per the linkedin post: the persons goal is to replace c/c++ with rust. That's not a plan; it's the team's north star. The headlines don't reflect what the person wrote, and it pisses me off. Slashdot should be better than that.
Here's my bet for the reality: Microsoft wants to have a agile, easily updated code base with a team of motivated programmers.
But the actual code base is giant and old. I've worked with two different teams where the people were younger than some of the code they touched. The code supports a bunch of features that aren't relevant to the world any more (for example: we still have code to work with corporate servers for regional dial-up connections. That's not a thing any more, and hasn't been a thing for decades).
Worse, the code base, being old, has old code patterns that strongly resemble security issues, and the teams spend a bunch of time chasing down bugs filed by the automated security analysis tools. It's a drain on the teams, and reduces morale.
So management spins up a team to try out AI on large code bases. It's a risky bet with a big payoff. The worst case is that some code gets rewritten. And the best case is that Microsoft gets a nice new billion-dollar business in "rewriting old code".