I've worked with similar tools and while they're impressive, it's too often the case that the LLM literally makes up fake but realistic looking data and pretends that it's real. This includes pretty deep fakery like setting up mock database connectors so that it looks like you're fetching data from the right place, but it's just getting synthetic data
I agree that a lot of progress in math comes from refactorings and novel concepts that generalize neatly. My point is that those breakthroughs don't happen through refactoring in a vacuum, they happen because the refactoring is undertaken with a specific motivation. It often ends up having more general applications than the original motivation, but doing it without a specific motivation doesn't usually yield progress.
It's the same as with software refactoring. If you refactor without a sense of what you want to get out of the refactor, how do you know whether you're refactoring the right things?
Besides the legal requirement, the reason these companies go public is often to provide liquidity for early investors or employees. So they do want to have as good of a margin story that they can, at least in terms of unit margin.
> Just serialize stuff to text if somebody wants to "edit" and then parse it back when they are done. What happens to the text in between of course is the programmer's problem.
I think that's working against the way that a lot of people write code.
You need things like syntax highlighting and code completion even while typing something like "my_var = " on a line. A normal parser doesn't work here, and bailing out to make it the programmer's problem leaves the programmer with a lot of work
Until it had backup storage. Which ended up being useful in 2011 when tens of thousands of mailboxes were deleted due to a software bug and needed to be recovered from tape...
How does this guard against that?