Twitter had been around for a long time and could very well be considered feature complete and run with a skeleton crew.
They don't own any mission critical software and in the days it went down after Elon started pulling the plugs, the only thing that changed was the people going to reddit to complain about stability.
I would say that in general dynamic type languages are problematic in a large codebase without strict safeguards (Any everywhere, untested paths, lack of test coverage, large methods with different return scopes, etc.).
I've worked a long time in C++ land in large codebases and the issues there are different, but to undig a project from the spaguetti land is like pulling teeth.
It's 100% a HMI and moving costs to the other end of supply chain.
We can have optimized automation in warehouses/logistics, but if you talk to any site manager you learn very quickly that no one wants any downtime or impact to their operation to introduce new machinery or optimize traffic, etc. If it is not built with that from the start it's very hard to introduce it later on unless there is a very clear deployment path and cost structure.
And boy, robotics currently has any of those today. Sure, move those billions in to R&D. Time will tell.
I believe Microsoft biggest achievement is being capable to stay relevant for the past 50 years, largely due to enterprise.
If you take a close look as an user, all their products is half-baked in some way (inconsistent behaviors, dark patterns, poor support, etc.), good enough so they can lock you in and hold your data hostage with time.
Gmail's +tag (and the .) is nice in theory, but terrible in practice. It's super easy for malicious actors to just drop them and there are a few services out there that simply are not able to work with the +tag, potentially getting you locked you out of your own account. Not gmail's fault, but I would recommend against using it.