Skip Navigation
TechTakes @awful.systems
David Gerard @awful.systems

Generative AI runs on gambling addiction — just one more prompt, bro!

video version

12 comments

  • This post is gold.

    • many ideas cribbed from rants by our valued comrade @fasterandworse, who rants about Hooked at the slightest provocation

      • David pulled it all together perfectly here. It's amazing how well it fits the hook model, whether they intended it or not

  • Curious what a study on relative AI uptake by Parkinson's patients and my fellow adhd'ers looks like. The same dopaminergic systems whose underperformance drives symptoms also makes us more likely to get hooked on gambling than average.

    Edit: also, AI-to-cocaine is the new cloud-to-butt

    • @o7o7 @techtakes

      I always wonder if an ADHDer gambling while a stimulant dose is active would find gambling even more addictive than a non-ADHDer because of the artificially increased amount of dopamine floating around. Or the same but with other potentially addictive behaviors.

      • In my experience, stimulants mean more self-control!

        ADHD is a broad church, but most of us generally show reduced dopaminergic activity around the frontal lobe, which can manifest, as an example, executive dysfunction. The right stimulant medication at the proper dosage helps balance that out. My unmedicated brother can drink a latte and go to sleep immediately.

        Gambling, browsing, etc feel good for the same reason, and I will never touch casinos, cocaine or street meth because I suspect that it'd ruin me.

  • This reminds me of the "Compulsive Programmer" chapter in "Computer Power and Human Reason" (c. 1976), and also of how I used to write code when I first started - way before LLMs were a thing and also before I studied proper engineering. That kind of unfortunately common type of programmer follows exactly the hook-loop model, except instead of relying on an LLM to randomise the result of each loop iteration you do it yourself by proceeding without really trying to understand the problem.

    I think this is a basic feature of programming, where a single iteration of trial and error is very fast and cheap, and where you can very easily have something that looks like it works without knowing why or even if it does. ChatGPT removes technical barriers and friction, sure, but programming was already kinda cooked. I would be interested in whether generative tools make this approach feasible in other more mature technical disciplines.

    Also, that chapter in Computer Power is well worth a read on its own, as a finely aged sneer at computering under the assumption that enough computering is a good substitute for understanding anything else.

12 comments