> This gap between perception and reality is striking: developers expected AI to speed them up by 24%, and even after experiencing the slowdown, they still believed AI had sped them up by 20%.
I wonder what could explain such large difference between estimation/experience vs reality, any ideas?
Maybe our brains are measuring mental effort and distorting our experience of time?
The "economic experts" and "ml experts" are in many cases effectively the same group-- companies pushing AI coding tools have a vested interest in people believing they're more useful than they are. Executives take this at face value and broadly promise major wins. Economic experts take this at face value and use this for their forecasts.
This propagates further, and now novices and casual individuals begin to believe in the hype. Eventually, as an experienced engineer it moves the "baseline" expectation much higher.
Unfortunately this is very difficult to capture empirically.
It was fun to watch, it’s super polished and sci-fi-esque. But after 15 minutes I felt braindead and was bored out of my mind lol
What if agentic coding sessions are triggering a similar dopamine feedback loop as social media apps? Obviously not to the same degree as social media apps, I mean coding for work is still "work"... but there's maybe some similarity in getting iterative solutions from the agent, triggering something in your brain each time, yes?
If that was the case, wouldn't we expect developers to have an overly positive perception of AI because they're literally becoming addicted to it?
https://softwarecrisis.dev/letters/llmentalist/
Plus there's a gambling mechanic: Push the button, sometimes get things for free.
My issue with this being a 'negative' thing is that I'm not sure it is. It works off of the same hunting / foraging instincts that keep us alive. If you feel addiction to something positive, it is bad?
Social media is negative because it addicts you to mostly low quality filler content. Content that doesn't challenge you. You are reading shit posts instead of reading a book or doing something with better for you in the long run.
One could argue that's true for AI, but I'm not confident enough to make such a statement.
I wish there was a simple way to measure energy spent instead of time. Maybe nature is just optimizing for something else.
This bit I wasn't at all surprised by, because this is _very common_. People who are doing a [magic thing] which they believe in often claim that it is improving things even where it empirically isn't; very, very common with fad diets and exercise regimens, say. You really can't trust subjects' claims of efficacy of something that's being tested on them, or that they're testing on themselves.
And particularly for LLM tools, there is this strong sense amongst many fans that they are The Future, that anyone who doesn't get onboard is being Left Behind, and so forth. I'd assume a lot of users aren't thinking particularly rationally about them.
I would think this could contrast with agentic coding, where the AI keeps generating code, and then you iterate on this process to get the AI to fix its mistakes. With normal human code review, it takes longer to get revisions and can feel like a slog. But with AI that's a much tighter loop, so maybe developers feel extra productive from all these dopamine hits from each interaction with the agent.
When manually coding and in flow state I'd think it's a more consistent level of arousal, less spiky. Probably varies by person and coding style though, which might also explain why some people love TDD and others can't stand it?
It’s also not a sample size of 1, it’s a sample size of however many tasks you do because you don’t care about measuring the effect AI has on anyone but yourself if you’re trying to discern how it impacts you.