AetherSpire
It will creep in any domain that is verifiable due to how reinforcement learning works.
I used to work at a company doing AI at its core, back before it was a dinner table topic - before every LinkedIn post had "AI-powered" in it.
This week my husband, a mechanical design engineer, mentioned his manager has started using Claude to help draft sales quotes.
Mechanical design engineering. If you'd asked me to rank departments by "least likely to touch AI first," that would've been near the top.
He said it's decent but makes errors sometimes. I told him that's exactly why I still block off an hour every week just to check whether AI has quietly broken any of the workflows I've built or started behaving differently than before.
He just stared at me. "You babysit AI?"
Pretty much, yeah.
Anyone else noticing AI creeping into workflows/roles you genuinely didn't expect?
Curious what that looks like in other fields.
It will creep in any domain that is verifiable due to how reinforcement learning works.
I go absolutely nuts debugging with some of my co workers. They way they type everything out, don't use CLI history search, don't copy paste or use hot keys. Small things add up if you do them hundreds of times a week. I stay quiet but I could never do that lol
I don't think many people here realize how advanced the latest gen of AI is, like Claude Fable 5. They don't even rely on accurate prompts anymore. All they need is to see something, like via screenshots, to analyze it and spit out a human-readable response.Feed it a problem, it outputs an analysis and response in seconds.We're on the cusp of big things with AI based on what I'm seeing it can do now. Just absolute ridiculousness how powerful it has become.