If you can optimize it in see Python and then feed it into pipi, maybe you could go exactly faster still, right? That would have run at about like, as opposed to from 10 hours to one hour. But we didn't really feel like spending them spending the time to add that to the pipeline. Like I said, it was unnecessary. And definitely was fast enough already. It seems like, if you need it to go faster, it seems like you could just keep iterating on this. Or use like scython, for example, which compiles Python down to C with the weirdly Python plus C syntax. If you have to, right? Yeah, I've never