For me it’s not about trusting the hotpatching; I know it can work. What I don’t trust is my own ability to not only do it properly but also to do it in less time than it takes to carefully close all my programs, reboot and then get them all started again. And so, reboot.
And to save making another comment elsewhere: About a week. It would have been longer, but I was having a configuration issue and a reboot seemed like a good idea at the time.




Sometimes a person’s brain is only capable of operating in a conversation mode. Full, natural language sentences. Or sentence fragments (like these), I guess.
Then there’s the fact that some people can’t discern the line between the artificial and the real, or else are able to ignore it where they can see it because the LLM makes that easy.
I occasionally bounce ideas off free LLMs and I’ve been mostly conversational when I’ve done so, but I’m aware of things like 1) it’s a crutch 2) they’re mostly wrong about a lot of things and 3) any praise they give is invalid, so I’m not yet into the trap of thinking they’re people.
… but I still feel kind of bad if I drop a conversation when I’m done with it without saying goodbye.