‹ BackHN Continuity

Thread

Plan mode is dead

591 points · 510 comments · jmvldz

  1. taurath · · focus · HN ↗
    I'm actively watching understanding slip away from developers, code review getting paired down to no comment checkmarks, and codebases go to bloated messes that nobody can read. Axioms like engineers must understand and take responsibility for the code they ship are getting torn down, and the products coming out are reflecting conway's law, becoming impenetrably obtuse and always "so complex there are no obvious deficiencies" (as opposed to "so simple there are no obvious deficiencies" which used to be the aim).

    The one thing plan mode helped is for the humans to get an understanding of the strategy, and be able to poke around and look at the design and architecture. You can achieve this with some self discipline and keeping shorter leashes on agents, but it feels like a losing battle. The best devs still put out good code, but the poor devs are learning nothing while their metrics look great. I can't help but think we are racking up immense amounts of debt that will very soon become due.

    1. greenowl · · focus · HN ↗
      And what happens when it becomes due?
      1. solenoid0937 · · focus · HN ↗
        Nothing, the people worrying about AI tech debt on HN have been worrying about it for years. 2024 HN threads had the same comments.

        The reality is that models just keep getting better and are very good at cleaning up the debt they created. The "tech debt" bill never came due. It won't.

        1. archagon · · focus · HN ↗
          Until some catastrophic data loss event or data leak that nobody understands or has any mitigations for. Oops! Guess bankruptcy and prison time are the ultimate uncaught exceptions.
          1. solenoid0937 · · focus · HN ↗
            It's not 2024. Models today are far better than humans at spotting the bugs that would cause this.
            1. trinix912 · · focus · HN ↗
              I remember this being said as far back as 2022 yet we're still here fixing things manually when the models inevitably code themselves into a hairball convoluted enough they can't find a way out.
              1. solenoid0937 · · focus · HN ↗
                No one at a FAANG is fixing things manually these days.

                If it turns into a hairball, just have a few agents rewrite it.

                1. trinix912 · · focus · HN ↗
                  Do you have any source to back this claim? I haven't heard of any Facebook, Instagram, Google or macOS rewrites being done with agents.
        2. taurath · · focus · HN ↗
          I’m not worried about tech debt, that’ll be my career making money from the laziness of vibe devs.

          > The "tech debt" bill never came due.

          Companies paying $200k a day for coding models to churn on what the coding models are messing up is one thing.

          I’m not even worrying about tech debt, I’m talking full on defects, production incidents, security holes, and reputational damage.

          1. solenoid0937 · · focus · HN ↗
            Do you think the average dev is better than frontier models when it comes to spotting defects and security holes?
Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.