Login
You're viewing the aleph.land public feed.
  • Frank van Puffelenpuf@c.im
    Oct 24, 2026, 10:53 PM

    @bit101 The argument is certainly valid, and it really does happen. I just find most of the discussions where it is brought up are where both sides are yelling at each other without listening.

    I have plenty of concerns about (the amount of) AI usage and our growing reliance on it, but I am also realistic enough to acknowledge that the programming skill might be done for - vs software engineering, which is very much still needed (for now at least).

    Every an optimist, I am excited about what these new tools enable people to do. I'm also *very* anxious about it, but prefer to focus on the excitement in public discourse - and keep the discussion about the anxiety inducing topics for 1:1s (either in person or in DMs), just to avoid the yelling mentioned above. :-)

    💬 1🔄 0⭐ 0

Replies

  • bit101bit101@mstdn.social
    Oct 24, 2026, 10:59 PM

    @puf Yeah, I'm also worried that "programming" is going away. And while I may have emotions about that, at the end of the day, I have zero interest in prompting an AI to write code for me. I'm also moving towards retirement, so just trying to get by for a few more years as I also have no interest in starting a new career.

    💬 1🔄 0⭐ 0
  • Frank van Puffelenpuf@c.im
    Oct 25, 2026, 3:43 AM

    @bit101 I still like the software engineering part, and for now find it fascinating to see how I can get different models to generate something useful - and how to reasonably validate and verify that. I miss typing the code of course, but it's still interesting enough to do.

    But I must also admit that I'm doing that in pretirement, which makes it easier.

    I'm in talks with a company to help with some DevRel work though, and there I'm really lying awake at night: what do I teach the folks that are now just starting? 😳

    💬 1🔄 0⭐ 0
  • bit101bit101@mstdn.social
    Oct 25, 2026, 12:01 PM

    @puf
    Eventually the engineering won't matter, only the outcomes. At least that's the goal of these things. Any part that "a human is needed for" is a target for automation. It might totally fail. But I expect it will be barely adequate and save shareholders enough money that it will go forward.

    💬 1🔄 0⭐ 0
  • Frank van Puffelenpuf@c.im
    Oct 25, 2026, 3:02 PM

    @bit101 Yup, same here. For software tasks, it's currently to the point of giving meh-to-ok level results for a ridiculous cost. But as the cost goes down and (hopefully) the results keep getting better, it becomes feasible for more such tasks for more people.

    What I am most curious about is how this'll change the devices. None of things I've made these tools build for me so far will ever see an app/play store. I've limited to use-cases that can run in a browser so far, but that won't always be feasible (especially looking at iOS there). I was expecting an escape hatch to appear, but recently also started seeing more people buying cheap devices and "writing" custom firmware solutions. There are so many fascinating options!

    💬 0🔄 0⭐ 0