Seems like there’s broadly two ways to use LLMs for coding - either as a way to generate the same code you would have written but faster, or as an opaque program-generator where you have no idea what the code is doing. One of these methods results in roughly the same amount of understanding and the other one radically less.
Why write code in 2026
61–70 of 321 posts
Re: Why write code in 2026
#62> For example, have you ever seen an agent follow the boy scout rule? Where they leave code better than they found it? And would you WANT them to try to do this?
Yes, it's in the rules; run profiles, check code coverage, do a critical review, post the report and follow up tasks. 90% of people I've worked with did not follow these boy scout rules nearly as well as today's frontier LLMs.
Is the author implying this is bad?
Re: Why write code in 2026
#63All this debate around use LLM or not is tiring and just black and white thinking.
Can I use agents to code a SWE project? yes, with nuances.
Can I write code for a SWE project? yes, with nuances.
Its more options now, I'll write code about projects I deeply care and will use llm at work where its shared slop and forced usage.
Re: Why write code in 2026
#64Recently, even a tourist lost to OAI's model in competitive coding. To be honest, I haven't been able to beat AI at coding since around 5.2. People often say 'AI can't write good code,' but in reality, the quality of AI's output is layered depending on the level of the prompt input. The deeper the prompt, the better the code actually gets. Usually, when people say AI code is terrible, it's because they either don't u…
I don't write code anymore, because AI writes better code than me. I could write code, but the next AI would find 10 ways to make it better and more consistent with the rest of the AI-generated codebase. So I just let it write all of it. However, I inspect it all carefully and I constantly asks it to reflect on the code quality, to refactor, to reorganize, to make the code as good as possible. The end result is code…
If AI is generating better code than you can, that does say a lot. I've tried various AI's available and outside of being a better Google or a fast reference for Stack Overflow; LLM's do not generate better code nor better running code.
Re: Why write code in 2026
#65Why is this even a legit question? I need to keep writing code to stay relevant, not to forget my craft, be able to review code... So many reasons. AI doesn't change a thing.
Re: Why write code in 2026
#66I write code all the time I can, outside the KPI metrics that everyone is being pushed to, I only care about AI for smarter code completion.
Re: Why write code in 2026
#67i write code because i love it. it's something that makes me genuinely happy, so why would i give that up?
Exactly this. I got into this field because designing programs and writing code is enjoyable. I'm probably behind on using AI and need to get more up to speed, but I never want to stop coding by hand.
Re: Why write code in 2026
#68Note as some may be confused by the "1 hour ago" with comments older than that: this submission was rescued by dang when a previous discussion existed: https://news.ycombinator.com/item?id=48883341
fwiw I think the rationale behind it is counterproductive because the only difference between a OP submitting their article link and someone else submitting their article link is internet points.
Re: Why write code in 2026
#69> For example, have you ever seen an agent follow the boy scout rule? Where they leave code better than they found it? And would you WANT them to try to do this? Yes, it's in the rules; run profiles, check code coverage, do a critical review, post the report and follow up tasks. 90% of people I've worked with did not follow these boy scout rules nearly as well as today's frontier LLMs. Is the author implying this is…
That quote is the lead in to an example I talk about after this quote illustrates what I often see.
> Agents bias to making the current change as safely as possible. I had a situation in a previous codebase where one morning, pre-caffeinated, my meat brain mentioned using browser local storage. So some random state was managed in local storage. Everything else through a backend database. When I looked at the code, the amount of wrapping and indirection to preserve this idiotic human mistake probably tripled the LoC. Agents can amplify our one-off bad decisions by being so conservative.
You can of course solve this many ways. And many of boils down to just how a particular humans brain works. Some will solve this by not reading code. Some will read / write code.
Whatever works for you is great. But many there is upside to the precision of not having code intermediated through the LLM for many.
Re: Why write code in 2026
#70Note as some may be confused by the "1 hour ago" with comments older than that: this submission was rescued by dang when a previous discussion existed: https://news.ycombinator.com/item?id=48883341 fwiw I think the rationale behind it is counterproductive because the only difference between a OP submitting their article link and someone else submitting their article link is internet points.
I was fairly confused myself as the author :)
This was actually my original submission last week. There was a front page submission last night from someone else (hence the comments). Then my old post got re-upped just now (1 hr ago)