4 ms·

I think a better question is: “what problem in my existence does this solve”. This one is a definite none.

• ein0p 2y ago
  There are multiple problems this could solve, actually. I can think up of at least one every day that a sufficiently advanced agentic AI could do for me, freeing me up from mundane bullshit. Imagine having a secretary onto whom you could offload bullshit tasks using freeform human language without any accommodations needed for the AI, and with near perfect contextual knowledge of what you’re currently dealing with. That’s the promise. But that promise is years away and in the meanwhile I’d be inclined to conserve my runway a bit more, especially when entering the era where borrowing costs are relatively high.
  • cjk2 2y ago
    I had a human one of them. I spent more time interacting with it and dealing with misunderstandings and corrections than I gained from not solving the problems myself. As pointed out elsewhere I was gaslit into changing the subject by an LLM the other day and then it got stuff wrong. End game is it created more work than not using it. And that’s what we need to be careful of. The promise versus benefits of technology are not always obviously swayed towards the benefit end.
    • ein0p 2y ago
      You’re probably not using GPT4. If that’s the case, you aren’t really using an “LLM”. With GPT5 that gap will be even more obvious.