On AI
41 points by thang
41 points by thang
It’s like being Sisyphus, but there’s another hill nearby and people are having a blast watching a crane lifting boulders to the top of it with questionable accuracy while it hits some spectators every now and then.
One must imagine Sisyphus pissed.
I laughed out loud when I read this.
Only loosely related: I found John Green's recent video about Sisyphus inspiring. You might too. https://youtu.be/--3Qi0Z0ASU?si=SziZh3ZuRslRWIK0
It really sucks that the marketing around "AI" has sucked all the meaning out of automated tooling that can be handy. Voice transcription, typo's are things(off the top of my head) I'd like to play around with, but it's hard to navigate things because the marketing on LLMs overshadows their technical value. This makes it even harder to have a productive discussion about LLMs.
The vibes in the world of vibecoding just isn't something I'm interested in, so I think the conclusion the author ends up on hits really close to home, speaking as somebody who likes to programming, just because it's fun in its own way.
After all the big tech controversies, the tech community really was like "let's make two more absurdly huge influential tech companies", while we use their botting TUI that slurps up 200+ mb on idle and call the process ✨ innovation ✨.
I am somewhat jaded, but I do feel like I have so much more fun in smaller, more niche, communities.
Nice post. I really feel all of that right now as well.
Regarding that last Andreas quote at the end, something that I don't think we've come to terms yet is that a lot of our learning was happening incidentally, as part of our everyday tasks. You couldn't help but learn while you were implementing some random feature, or doing the yak shaving to set up CI for a new compilation target, etc.
And now, we can magic those bits away. We can just "create" whatever is in our head. But we can't just incidentally learn anymore. It's actually really hard to stay engaged while the LLM churns through a bug for you.
I don't know what the conclusion of this is. But I don't think that saying "I actually like coding by hand" will be virtue signaling, any more than saying "I like to lift weights" is virtue signaling. You have to put in some time somewhere now to learn and get better.
It's easy for someone like Andreas to pivot to using LLMs because he has built up the skill to ensure they're doing quality work and solving the right problems, and that's not going to atrophy immediately. But I don't know how the next generation of programmers ever gets there if they have LLMs from the outset, and the simple types of work that build skill are also the easiest to automate away. As an industry we are collectively shooting ourselves in the foot.
I think the author is on point. If you enjoy writing code by hand, then you do it for the sake of it, because the journey is the point. The fact that LLMs exist should not stop anybody from writing code because they find the activity to be stimulating.
While I agree that this sounds nice, and is something I genuinely want to believe in, I also feel pretty deeply the despair of being a quiet ship in a cacophonous storm that the post alludes to as well. To protect my sanity I end up looking at what other people are doing less often, and with less vigour, and I mostly don't talk about or share computer things I'm doing myself anymore. I'm sure there will always be small communities of software carpentry around the fringes, but I expect they'll be difficult to locate and become a part of, as they're probably sheltering themselves too.
I don’t know. I believe the author’s point is that two incentives that could previously be coupled (depending on the problem you pick of course) are now decoupled for good, so you have to pick one.
Not the worst problem you might have, but definitely a new thing to consider, for sure.
Any new form of automation is inherently disruptive. What used to be a necessary skill is now becoming more of an art form. I don't think there's anything inherently wrong with that, but it will be an adjustment period to be sure.
LLMs existing may stop one from being able to make a living being a "software carpenter" though.
Doing something difficult for no practical reason is the purest form of “art”, and the fundamental motivation for making “art” is personal. Nothing anyone else does can make it less authentic or valuable.
Note that this doesn’t apply to art made for practical reasons, like selling to other people, in which case other peoples’ opinions are highly relevant. But I don’t think making a MacOS 9 web browser is in that category.
I learned to program as a kid and have done it literally as long as I can remember. I feel compelled to make things and have always looked up to people like OP who do things that are difficult because they have the same urge that I do.
Using an LLM robs you of this and I don't have the same respect for it. It's not about what you produce at all. I want to celebrate when someone writes the thousandth NES emulator, even though no one will likely use it, because they harnessed some inner passion to grapple with a challenge that was unfamiliar to them, and came out stronger for it, and we have the evidence that they succeeded. You're probably not going to add much novelty in your implementation, but it's yours, and you've proven yourself to be the type of person who could make it. Using an LLM to generate an NES emulator on the other hand is entirely pointless, because there are already many of them and you learn little that way.
I hope people keep practicing the craft even if everyone who never had that passion deserts it. It's one of the things that makes the world beautiful to me.
My personal issue with using AI is that I find I retain much less when using it. I might get the task done, but I don't learn nearly as much, or remember nearly as much.
Are there good strategies for retaining the learning side of building while using AI tools?