I lean towards this, despite being a guy currently heavily invested in AI tools.
Cluster of things that all seem true to me:
I am 100% addicted to vibecoding in a straightforward “this is fun and dopamine inducing way”, which is making it hard to reason about.
As fun hobbies go, it does seem fine, all else equal
You’re broadly right about the 80%/20% rule, and microoptimizations not being worth it.
But:
There are some infrastructure that probably will make sense to have as the AIs scale up, which AI companies either won’t provide, or, you probably don’t want to trust. (This is a bet, I’m not that confident)
Learning to wield AI is going to be important (for at least many people. I think it’s more straightforwardly important for software engineers than theoretic researchers).
Some of that infrastructure is stuff that AI can basically one shot. I think the skill/habit to cultivate here is “check if it 1 or 2 shots it. If not, bail.”
I am 100% addicted to vibecoding in a straightforward “this is fun and dopamine inducing way”
I’m curious, how does that work? What mindset are you approaching it from? What sorts of projects (in terms of their… emotional felt-sense, I guess) are you attempting with it?
I think I would like to be able to engage with it as with a hobby, but it’s not been fun for me.
For me it’s like “I type some quick stuff in, and then, like, agency comes out and I get to see stuff get built, and it works great 20% of the time, okay 60%, and fails 20% of the time, but, that produces a kinda skinner-box slot machine element to it.” (to be clear I think the skinner-box bit is bad, the “stuff comes out with little effort” part is great. It’s like jamming with a partner who can do most of all the tedious parts of the work)
My impression from your other posts is that you are mostly just getting a much worse hit rate (because yeah if it’s not really set up to excel in a domain, it’s a lot less workable)
My impression from your other posts is that you are mostly just getting a much worse hit rate
No, the hit rate sounds mostly similar. I think it’s more that I may have unusually strong anti-gacha instincts? Like, if I’m doing something, momentarily reflect on it, and recognize that it’s equivalent to playing a slot machine, this immediately causes negative feelings in me and sours the whole experience. Which I guess is usually a good adaptation to have, but may or may not be be anti-helpful in this specific case.
I lean towards this, despite being a guy currently heavily invested in AI tools.
Cluster of things that all seem true to me:
I am 100% addicted to vibecoding in a straightforward “this is fun and dopamine inducing way”, which is making it hard to reason about.
As fun hobbies go, it does seem fine, all else equal
You’re broadly right about the 80%/20% rule, and microoptimizations not being worth it.
But:
There are some infrastructure that probably will make sense to have as the AIs scale up, which AI companies either won’t provide, or, you probably don’t want to trust. (This is a bet, I’m not that confident)
Learning to wield AI is going to be important (for at least many people. I think it’s more straightforwardly important for software engineers than theoretic researchers).
Some of that infrastructure is stuff that AI can basically one shot. I think the skill/habit to cultivate here is “check if it 1 or 2 shots it. If not, bail.”
I’m curious, how does that work? What mindset are you approaching it from? What sorts of projects (in terms of their… emotional felt-sense, I guess) are you attempting with it?
I think I would like to be able to engage with it as with a hobby, but it’s not been fun for me.
For me it’s like “I type some quick stuff in, and then, like, agency comes out and I get to see stuff get built, and it works great 20% of the time, okay 60%, and fails 20% of the time, but, that produces a kinda skinner-box slot machine element to it.” (to be clear I think the skinner-box bit is bad, the “stuff comes out with little effort” part is great. It’s like jamming with a partner who can do most of all the tedious parts of the work)
My impression from your other posts is that you are mostly just getting a much worse hit rate (because yeah if it’s not really set up to excel in a domain, it’s a lot less workable)
Thanks!
No, the hit rate sounds mostly similar. I think it’s more that I may have unusually strong anti-gacha instincts? Like, if I’m doing something, momentarily reflect on it, and recognize that it’s equivalent to playing a slot machine, this immediately causes negative feelings in me and sours the whole experience. Which I guess is usually a good adaptation to have, but may or may not be be anti-helpful in this specific case.