• 0 Posts
  • 6 Comments
Joined 4 months ago
cake
Cake day: March 13th, 2024

help-circle


  • I use LLMs all the time for work and hobbies, but my work and hobbies are well suited for LLM assistance.

    Writing boilerplate documents. I do this for work. I hate it. LLMs are very good at it.

    Writing boilerplate code. I do not like writing docstrings, making my code more maintainable, enforcing argument types, etc. I do a lot of research code and I need to spend my time testing and debuging. I can feed my spaghetti into an LLM and it will finish out all the boilerplate for me.


  • I do not want my information filtered through an opaque algorithm. My worldview is much too important to surrender to some corporation. I want to understand and have some control over any feed I use. My media diet includes Lemmy, AP news, PubMed/science journals, and conversations with friends and coworkers.

    I am very happy with Lemmy so far. Some have pointed out there is less content on Lemmy, but that is a bonus in my book. It is not healthy to spend hours scrolling.



  • This is a pretty complicated topic that touches video games, gambling sites, social media algorithms, and marketing in general. It also touches fundamental philosophical questions like the existence of free will.

    We have lots of established law on which sort of “mind tricks” are fair play and which aren’t, but we have not advanced those laws to keep pace with the science. Currently, lying is really the only thing off limits and is covered by fraud statutes. We also have some limits on marketing to children. But one could argue that there are several “persuasion” tactics that can be just as effective as outright lies in manipulating the behavior of others. In fact, licensed therapists are ethically barred from using these tactics, yet we allow salesmen, marketers, etc to use them at will.

    I don’t really have an opinion on this lawsuit, nor do I feel qualified to offer a solution. But let me give you an example of how the human mind works which underpins addiction to gambling.

    Dopamine is a signaling molecule that regulates a lot of our reward responses. If I find honey in a honeycomb, dopamine gets released and now I am more likely to seek out honeycombs in the future. You can see how this is evolutionarily beneficial. Dopamine release reinforces behavior that increases survival. But let’s say that only about 1/3 of all honeycombs have honey. Now I have a lower chance at a reward, so does that mean the dopamine release is likewise diminished? No, the opposite is true. Dopamine release skyrockets. Evolutionarily this makes sense, we do not want to miss out on a reward simply because the probability is diminished, so the high dopamine release counterbalances the diminished probability such that reward seeking behavior is reinforced so long as the probability of reward is reasonable (it peaks at about 1/4). In fact, dopamine is released even when the honeycomb has no honey. You can draw a direct line between this physical phenomenon and gambling addiction. What people don’t appreciate is that this physiological response is very similar to addictive drugs in effectiveness. It can be hard to acknowledge that one of the reasons you are not a gambling addict is simply that you didn’t start gambling to begin with, not that you are somehow superior to those that are addicted.

    We have lots of behavioral quirks like this that can be exploited. At what point does this manipulation cross the line? That is a hard question. For me, gacha games cross that line. But if we want to enact meaningful regulations we need to acknowledge that these mind exploits exist and confront the fact that free will may not be as free as we hope.