The point is that non-deterministic outcomes can be a solution to the problem.
We are at the stage that Donald Knuth is releasing literate programs with LLM assisted proof. They aren't stochastic parrots. They are a useful additions to the thinking mans arsenal of tools.
The market loves it too, MSFT is up 6% this week.
There are likely other things that drive this but my hope is that the share price is increasing on the idea that these patches are so disastrous that even Microsoft should be able to see that their current strategy isn’t delivering a good product.
[1] https://collections.vam.ac.uk/item/O1227026/concrete-stereo-...