Ask HN: Dealing with Vibe Coding Depression?
16 softirq 23 6/4/2025, 7:31:48 PM
While originally I was an LLM skeptic, I was also eager to gain insight into it’s true capabilities, and recently I’ve reached the tipping point of existantial dread - I no longer feel any joy while coding. I’m no longer an artisan enjoying the journey of creating, I’m now truly a cog designed to review factory output until even that role is no longer required.
My biggest feeling right now is an immense sense of loss. My belief was that the purpose of one’s life is found through acts of creation. The painter finds joy in painting, and the result is valued because of the effort involved. This feels like an attack on all intellectual pursuits, including the arts, but it’s especially hard considering the technology seems to have the most value at replacing its creators.
Where do we go from here? So many of my friends have talked about switching fields, as we watch this miracle field edge towards becoming a facsimile of itself. I am personally left with many questions about my own future.
The creators at these companies are making all their bets that the switch between interpolate and extrapolate will happen at sometime in the very near future.
If and when that fails to materialize (as I think you could argue some companies are already recognizing is coming, ie Microsoft), the bubble will burst.
Where we go depends on the philosophy we apply, and there's way too little philosophy for what to do when AI upends your intellectual craft profession.
All businesses exist to serve a customer, and to help them achieve the outcomes they desire, so to the extent that you can apply these skills to those outcomes then you'll continue to be gainfully employed in a business context.
I think zero sum applications of software like internal tools with fixed scopes will rapidly be automated away with these tools. By comparison, positive sum product engineering of novel technologies seem like they are still hiring.
Would focus on bridging your technical skills into product engineering and deeply understanding a specific customer domain, invest in EQ through coaching and therapy and continue adopting AI tools to optimize your work.
I haven't experienced this at all, especially when you can ask the AI to generate unit tests based off of specs.
I'm manually testing AI output twice as much as if I wrote the code myself, but I'm generating code 10x faster so I'm still gaining massive amounts of productivity.
That's where we are right now. The best bespoke hand-crafted coders are far less valuable than they used to be.
It sucks for the weavers who loved to weave, but this is the consequence of technological progress.
It’s not a loom, or a car, or a printing press or python vs punch cards.
It is very different, and the facility “Management” is hoping to replace, is decision-making, and not some predictable mechanical repetitive tasks.
The thing itself is more similar to an idiot savant with no memory or understanding. Getting it to do the right thing is sometimes easy, sometimes more difficult than doing it yourself, and sometimes catastrophic. Letting it on without supervision is almost guaranteed catastrophic.
“Management” trying to replace people with it is a function of perceiving people as cogs.
Is it? Isn't this just what happens in a capitalist system?
me the first time my boss forced me to unit test my code
...
The best thing you can do is listen to your gut and try to act as rationally as you can.
Talk with trusted mentors if you've got them. Don't listen to me and for the love of god don't listen to people on HN or reddit or Youtube or any other social media.
Nobody knows what they're talking about and they certainly don't know how it'll impact you.
If somebody is making you feel afraid, left behind/out, inferior -- they're trying to sell you shit. Don't listen to the bullies and con artists.
You're entitled to your opinion. If you think AI output is crap, it's crap. Don't be pressured to conform. This is supposed to be hackernews after all. There are plenty of companies using java 8 today. You won't be unhireable.
> The painter finds joy in painting, and the result is valued because
Engineers are not painters in one very fundamental way!
Painter’s product is an asset. Software engineer’s code is a liability.
Painter is an artist, creating art. Its primary purpose is in itself.
Software engineer is a craftsman, he creates a mean to an end. A tool to reach product/business goal.
End product is an asset, not the implementation details. That's the truth, regardless of how you or I feel about it.
That doesn't mean there is no beauty in the craft of writing the code etc. But it's not the key part.
For folks who love software, it absolutely is a key part. Donald Knuth titled his book series "The Art of Computer Programming" for a good reason. It taps into the human creative impulse under a compelling set of technical constraints.
Generative AI in any form has the effect of forcing us toward the vanguard, to create something genuinely new or humanly beautiful if it is to be at all valued by others -- whether that novelty/beauty is because of the unique constraints of your company's internal software ecosystem or because you're striking out and building something others haven't built before (even a novel combination of existing ideas).
Novelty is very hard for most people, but perhaps the beauty (in the classical sense) of software projects[0] can still be recovered. Human agency always will be a powerful thing. (And there will always be other people for whom to create things.)
[0] 37Signals has always exhibited a compelling sense of joy in the process of creating beautiful software for people.
Those are the folks from early days when it was more about math (which is an art) and exploration rather than about actual software _engineering_.
If you're building new algorithms, new computers etc. - yes, your product is the tool itself. But this is a very limited case these days.
Unfortunately many folks drag this attitude in their business software domain and this leads to overengineering and redundant complexity.
With experience you can smell if this part of codebase was mostly made by folks who just like engineering for the sake of it, rather than people trying to reach the business goal in a reasonable time/cost. And this smell is terrible.
But just like an architect designing a beautiful bridge that becomes a landmark, with the right management and product-market fit, it can be possible to profitably build beautiful software products for people. There are real, profitable businesses that do this, and people love their software.
People enjoy and value beautiful things; as a software business culture we ought to value creating them.
Instead of spending days updating a software package to the latest version, to get the exact same features I already had, I can focus my time prioritizing features and designing the code infrastructure.
And when I did use LLMs, I found that they were wasting time by spitting some code that I have to then piece together, troubleshoot or debug.
With abundance coming around the corner for coding, anyone write code, which means software engineers will be needed more than ever.
There will always be new jobs.