Live data from Hacker News

Anxiety as an algorithm

adamjuliangoldstein.com

41–50 of 50 posts

Re: Anxiety as an algorithm

#41
post #37

Earlier quoted context omitted.

Evolution is a very simple self-optimising algorithm that's been allowed to run for a very long time. The way I see it, algorithms often have room for improvement. Turns out, self-optimisation with no oversight for 400M years produces something with a lot of room for improvement.

What even is 'optimal' though. As far as evolution is concerned, we are surviving to pass our genes on to the next generation, so we already are 'optimal'. Going any further than that is just bringing in social and cultural biases about what 'optimal' is.

Optimal is what we say it is. And we generally don't consider "surviving to pass our genes onto the next generation" as a sensible function to optimize for.

Re: Anxiety as an algorithm

#42

Earlier quoted context omitted.

Evolution is a very simple self-optimising algorithm that's been allowed to run for a very long time. The way I see it, algorithms often have room for improvement. Turns out, self-optimisation with no oversight for 400M years produces something with a lot of room for improvement.

See, I don't buy this (I'm not arguing with you, just initiating a conversation). There is no doubt we are incredibly complex, so much so that we've not been able to crack the puzzle that is "us". We invent crude medicines to carpet-bomb some bacteria/virus, which can potentially harm another part of the body, we staple cadaver-parts onto our knees to move well again, we're incredible crude in a bio-hackish way. We'v…

Worth remembering though that the only way we can gain more understanding is by pulling strings and seeing what happens (and comparing the result with what we thought would happen).

Re: Anxiety as an algorithm

#43

I think this post makes some interesting points, but anxiety is a slightly more complex subject. As a founder of a big company, there are ways it can go wrong, and being worried about those things is probably correct. A rational anxiety is not really an anxiety at all, but a fear. When the fear response is decoupled from threats in reality, that is when a response is considered anxious. Unlike in the case of the immu…

No analogy is ever perfect, but auto immunity seems to be a good extension to cover anxiety: identifying a threat from a disease or a noisy bush is helpful, but overfitting reality to find threats that aren't there is the opposite.

Re: Anxiety as an algorithm

#44

It's interesting that the article mentions the immune system, because I've always considered if allergies are the physical equivalent to mental anxieties.

That's a great insight. I never experienced chest caving anxiety ever, until I did once. Now it's much more likely.

On the flip side there is treatment to allergies, controlled small doses to triggers. I think anxiety may work similarly. For example, public speaking more often with smaller audiences will reduce initial anxiety in the long run.

Re: Anxiety as an algorithm

#45

I have noticed a tendency (especially amongst HN articles and readers) to continually compare our bodies, minds, souls etc. to machines, as if they could be "biohacked" into perfection with just the right patches or software updates. Or that we could easily be replaced by the "correct AI" algorithm, if we could just get the simulation close enough. Not that I don't believe we aren't complex systems, we are, to be sur…

I definitely agree, and I think this mindset develops in many technically inclined people only after you spend a significant amount of time living in the real world.

I imagine many of us from a young age find an interest in math and science because the logic makes sense. We build on first principles and everything has an objective truth. Perhaps in our early lives we experience things that don't "make sense" in this way, so we throw ourselves into a world of logic that we can more easily understand. Great for academic success, and after all, when we're young we're probably in a world where academic success is lauded (school or college). We are quite literally rewarded with good grades and honor societies for thinking in this way, so it reinforces the idea that this way of thinking is "right". Or at least, this was the case for me throughout high school and college. I remember thinking at the time, "All of these people around me partying, having fun, getting into trouble, living illogically -- they're doing things wrong."

But then at a certain point we leave these structured environments and are on our own. And we continue to run into experiences where we try, again and again, to apply algorithms to life, and they almost always fail. It seems like we can come to one of two conclusions: Either the algorithm is still incomplete and the parameters need tweaking, or we realize that there is no algorithm, or at least no knowable algorithm. We're just too complex, too irrational by nature, to be described by some knowable set of rules.

Every time our "life algorithm" fails it hurts, because it feels like we failed. So we do whatever we possibly can to avoid failure in the future, and yet those failures keep coming and hurt more and more each time.

Until, hopefully, we arrive at that second conclusion which comes with both relief and regret. Relief because instead of trying to describe life with an algorithm, we can actually start living it. And regret because we no longer have the freedom to make those same sorts of mistakes that help us build character when we're young, because they probably have real consequences now. Plus it's really damn hard to give up the way we've been living since childhood. Things like practicing mindfulness and meditation help a lot.

Anyway...that's my two cents.

Re: Anxiety as an algorithm

#46

I have noticed a tendency (especially amongst HN articles and readers) to continually compare our bodies, minds, souls etc. to machines, as if they could be "biohacked" into perfection with just the right patches or software updates. Or that we could easily be replaced by the "correct AI" algorithm, if we could just get the simulation close enough. Not that I don't believe we aren't complex systems, we are, to be sur…

I didn't read the article as being about how to hack your brain for an extra 10% job performance, I read it as a model for dealing with mental illness and a prescription at the end which bears some resemblance to existing therapies. There are anxiety disorders which can have a crippling effect on someone's ability to function and turn their life into a living hell. There are many entrepreneurs who suffer from them. E…

That's right, I'm writing this essay series for founders whose anxiety is interfering with their happiness and effectiveness.

Re: Anxiety as an algorithm

#47

I think this post makes some interesting points, but anxiety is a slightly more complex subject. As a founder of a big company, there are ways it can go wrong, and being worried about those things is probably correct. A rational anxiety is not really an anxiety at all, but a fear. When the fear response is decoupled from threats in reality, that is when a response is considered anxious. Unlike in the case of the immu…

Thanks for sharing! I agree that certain anxieties are "rational." As I mentioned at the end of this piece, my second essay in this series will be about why people get anxious about different things. Look forward to hearing your thoughts on how my model for that compares to your research.

Re: Anxiety as an algorithm

#48

Earlier quoted context omitted.

A world of only perfect beings does sound pretty boring. I'm not sure if anything worthwhile would actually happen in such a place. That being said, mental (rational) tools that help one recognize the origin of anxiety and reduce its deleterious effects are quite welcome. I think this article could help programmers/founders do just that.

If you had a world of perfect human beings, then they would create imperfections to stop themselves from getting bored. I know, I would.

The desire to do that sounds like an imperfection to me...

Re: Anxiety as an algorithm

#49
post #6

I only skimmed so I might have missed these points being discussed... The 'mental immune system', which uses imagination to search for potential paths to classify likely future outcomes suffers from one very critical issue. Selection bias / neural network (literal) training set problems. Repeated failures without success lead to the weight of that success against the failures skewing the prediction algorithm. Reality…

There's a lot of research to suggest that this is why kids from lower social-economic backgrounds have lower levels of self-control. They're so used to negative outcomes that they've been neurally trained to immediately take a positive outcome even if a more positive outcome would present itself if they waited.

The classic experiment is the cookie experiment, where the outcome is by resisting a plate of cookies they can have as many as they like but if they take one straight away they only get one. But it generalizes to many other outcomes too (not sure if outcome is the right word but you get what I mean?)

Re: Anxiety as an algorithm

#50

I think this post makes some interesting points, but anxiety is a slightly more complex subject. As a founder of a big company, there are ways it can go wrong, and being worried about those things is probably correct. A rational anxiety is not really an anxiety at all, but a fear. When the fear response is decoupled from threats in reality, that is when a response is considered anxious. Unlike in the case of the immu…

Thanks for sharing, I'll definitely read that. How do you balance between the threat being real and the anxiety not necessarily being helpful? I can think of many situations where my emotional reaction to things are the correct reactions but are disproportionate to the point of being unuseful
Post reply on HN