Ask HN: How to avoid skill atrophy in LLM-assisted programming era? | Hacker News Reader

Ask HN: How to avoid skill atrophy in LLM-assisted programming era?

18 pointspy445 comments

Will technical skill even matter at all?

What skills are atrophying that would be useful in the future?

If you're letting LLMs do more than assisting, don't. That's my advice. But if like you're title they're just assisting you, then what skills are atrophying? You still review the code and understand it right? You still second guess the LLMs proposed solutions and look for better approaches right?

Articulating how LLM assistance is different than junior programmers writing code and assisting would be useful, everyone has different setups and workflows, so it's hard to say in my opinion.

Let's say you want to make an architectural change. There are two options:

1. Ask AI to come up with the different options and let you review it

2. You think about the options and ask AI for feedback

#1 is much faster but results in atrophy (you are not critically coming up with the architecture changes)

#2 uses your and AI skills but it's gonna be slower.

which one will you choose? currently i'm doing #1

I agree... I think the process of coming to a conclusion yourself is different than having that solution proposed to you and accepting it.

> What skills are atrophying that would be useful in the future?

Well for once, tech companies are still at large hiring via leetcode/livecoding interviews. I feel much less prepared now that I was a year ago.

Were you really using anything in your day to day work that had any relevance to preparing for tech interviews?

In my experience, many heavy LLM users do not review their own code. They'll blindly open PRs full of slop, making it the reviewer's problem.

IMO the code itself has become much less valuable. Most people in this thread are telling you to stay in the code but I would argue you need to stay current on how to architect a good project. What supporting infra do you need? Did you pick the right language? Did you break the project up into appropriate tasks? You need to become a really great PM.

Learn to wrangle your agent better than everyone else. Don't rely on the chat too much, break up your project into tasks, learn to use sub-agents.

Learn to use the new tools well.

This tool seems obvious but its message is really that what you prompt is profoundly important.

https://developers.googleblog.com/conductor-introducing-cont...

Have a personal site and passion (read: not side gig) projects you work on outside of work. Hand code, get frustrated, be ambitious, don’t open Claude every time you forget a tailwind class

If you don’t have ideas, spent more time away from the screen, they will come.

> If you don’t have ideas, spent more time away from the screen, they will come.

Love that, and you stated a fact. Or, rethink other products!

Ask HN 1800: How to avoid losing spinning wheel skills in new spinning jenny era?

Ask HN 1920: How to avoid losing farrier skills in new automobile era?

Ask HN 1980: How to avoid losing typewriting and shorthand skills in new microcomputer era?

Ask HN 1990: How to avoid losing assembly language skills in new C++ era?

Ask HN 1995: How to avoid losing DOS TUI app dev skills in new Windows era?

Ask HN 2000: How to avoid losing Visual Basic skills in new web application era?

(The answer, btw, is if you are still interested in such niche skills, then you just have to practice on your own, or find a niche product or marketplace).

The younger generation discovering TUIs has been amusing.

Funny enough, I lived through all of those eras starting with assembly in 1986 in sixth grade),

1996 - C and Fortran on DEC VAX and Stratus VOS mainframes

2001 - C/C++ on PCs and mainframes and starting to work on VB

2006 - JavaScript/C#/some Perl

2011 - C# on Windows ruggedized devices

2016 - .NET Core

2021 - Working as an L5 at AWS (ProServe)

2026 - staff consultant at a 3rd party consulting company. Every single project I’ve done has had Bedrock (AWS service that host most of the popular models) and I constantly have three terminal sessions open - one to run code, one running Codex and the other running Claude.

Ask HN 1647: How to avoid losing ship captaining skills in the tulip trading era?

How to avoid skill atrophy? Easy. Limit your use of LLMs. Intentionally practice. It's what I do.

You're losing if you're handing your brain over to LLMs right now, because companies would prefer to hire someone with more up-to-date coding skills, even if they then force them to use LLMs. So the winning move is to resist using LLMs for as long as possible.

Stop fanboying the industry's attempted commodification of your work, and get back to the basics.

Man its really sad to see that this place seems to embrace LLMs with open arms and seems to have no care for the implicit costs and side costs of it.

I have no interest in SWE - I focus on other fields. But, LLMs are a complete disaster of a product, as the more you use them the less you are engaging your own brain to tap into the knowledge you have to get shit done and move fast. LLMs are a mirage and the fatal flaw of a human is laziness.

This lack of brain engagement is deadly. People dont realise how tough it is to get back once you've started to lose it. Its akin to the gym and muscles.

If your only “skill” is “I codez real gud and turn well defined requirements into code”, you were commoditized a decade ago.

Huh? When did I say that was my only skill? Did you reply to the wrong comment?

Well an LLM only helps you code, coding is not a competitive skill in 2026. If your “work” can be commoditized by a next word predictor, it was going to be commoditized by someone willing to work for less than you make anyway

1. No, implementing well defined requirements were not commoditized a decade ago. You still have to come up with the design and proper (efficient,correct,...) solution that respects the requirements. it was and still is the skill set of a L4/L5 SWE.

2. If you think LLMs cannot help with navigating ambiguity and requirements, you are wrong. it might not be able to 100% crack it (due to not having all the necessary context), but still help a lot.

You realize you are arguing my point? We are in complete agreement about #1.

As far as #2, I came into a large project at my new at the time company last year one week before having to fly out to a customer site. I threw everything I could find about the project into NotebookLM and started asking it questions like I would ask the customer. Tools like Gong are pretty good to at summarizing calls. I agree with you on #2.

I am at a point now where I am the first technical person after sales closes a deal and I lead (larger) projects and do smaller projects myself. But I realize remotely, my coworkers from Latin America are just as good as I am now and cheaper.

I’m working on moving to a sales role when I see the time coming. It’s high touch and the last thing that can’t be taken over.

I would never have trusted any L4 or L5 SWE I met at AWS anywhere near one of my customers (ProServe). But they also wouldn’t let me put code into a repo that ran an AWS service. Fair is fair

If I remember correctly, the leveling guidelines were (oversimplifying).

An L4 should be able to handle a well defined story

An L5 should be able to handle a well defined Epic where the what is known bit not how

An L6 should be able to lead a more ambiguous longer term project made of multiple Epics.

I was saying it was not commoditized a decade ago, but i feel it's getting commoditized *now*. So you seem to be basically saying SWE is over and it's time to move on to something that is primarily based on human-human interaction?

Yes it has to be. LLMs are getting to the point they can do everything else. What they can’t do, cheaper non US labor can.

For context, the software developer market in the US is very bimodal, most developers are on the enterprise dev side (including most startups like YC companies). I’m referring to this side - not FAANG and equivalent

By commoditization back then, I knew there was nothing I could do on that side of the market that would let me make more than around $150K-$165K. My plan then was to get on the other side of the market in 2020 after my youngest graduated and out of enterprise dev.

“Commodization” now means too many people chasing too few job. In 2016, I could throw my resume up in the air and get three or four random enterprise dev job offers within less than a month - now not so much.

I discovered AWS belatedly later that year and my thesis was changed to I want to do #1 that you said above - customer focused, using AWS as a tool, and bringing a developer mindset to cloud implementations.

It just magically happened in June 2020 that both felt into my lap - cloud consulting full time opportunity at BigTech (no longer there thankfully).

Simply don’t use an llm to assist you? You don’t have to, I don’t. I don’t even use an lsp.

How is doing either going to keep you competitive in the market when everyone is coding faster than you using modern tools?

That stance honestly sound like me not using a compiler and doing everything in assembly like I did 40 years ago in my bedroom in 6th grade on my Apple //e.

I might be an old guy at 51. But I’m not that old guy. I’m the old guy who didn’t have to worry about “ageism” in 2020 when I got my first (and hopefully last ) job at BigTech in 2020, another after looking for two weeks in 2023 (with 3 offers) and another in 2024 when looking just by responding to an internal recruiter - I’m a staff cloud consultant (full time) specializing in app dev.

Not claiming I’m special. But I like to eat, stay clothes and stay housed. I do what it takes

Are they coding faster? The METR 2025 study shows LLM users feel faster but are actually slower. If LLMs make programmers more productive, then awesome LLM–written software is available everywhere for a low price — so where is it?

simple example: Claude Cowork was written entirely with claude code

Did they make it faster than without Claude Code?

Why would any company past the savings on to you? Besides “no one ever got fired for buying IBM”. No company is going to replace SalesForce with a vibe coded alternative they find out about on “Ask HN”. Coding has never been the issue with having a successful project. Look at all of the companies who were crushed by BigTech just throwing a few developers at the same problem and having a good enough alternative.

Competition forces prices lower.

Your technical skills are shaped by market demand, and they always have been.

One way is to not give in to LLM hype and ignore LLM grifters.

LLM is not hype. it has made and my colleagues who are NOT working on CRUD, way more productive

What is it that you are working on?

llm training/inference stack

If you believe that, then what's the point of this thread? You've decided (wrongly imo but that's not the point I guess) that the LLM is better than you and should be trusted to to the job. If you start from that position, then of what use are the skills you wish to keep fresh?

You don't have to think the AI is better than yourself. Many coding tasks are just repetitive boilerplate... pretty simple stuff. Sometimes you have to set 20 fields on an object, refactor 10 functions to return a default value, write a generic SQL statement that returns data in a specific shape, center a div, or any number of relatively simple tasks. I wouldn't use it for the high level architectural decisions. Just a fancy context-aware autocomplete. Even though I can spell just fine, I use autocomplete on my phone all the time just to save time. I think it's a similar thing for code, if you use it properly. Of course many do just offload all the thinking and do not critically review it's work, but I think that is the wrong approach.

Yes it is hype.

The skills that atrophy are the ones you weren't using anyway. If you let the LLM do the interesting/engaging parts, that's on you, not the tool

if it's able to do the interesting/engaging part faster than me, i don't see why i should not outsource to it (The same argument as why use LLM-assisted programming at all, you don't want to miss the productivity boost)

What then is your interest in avoiding skill atrophy? It sounds like you realize that outsourcing your programming work to AI will likely result in skill atrophy, but you are so happy with the results that you are okay with this. (And so are a lot of people! Not saying it's a bad decision.)

What change are you after?

i'm trying to see what i can do to stay relevant

Aren't you making yourself irrelevant and the first group to be cropped out of the market or do you see others who don't use llms as much as dinos who will be filtered away first because your method offers more productivity?

I personally rarely have been paid for productivity. How fast I can put out features rarely earns me extra money. What people want is someone who understands what they want and finds a way to deliver when we agreed to and spots pitfalls along the way.

If LLMs are the best programmers then programming is obsolete so why do you want programming skills?

you need to be able to review what the LLM is writing.

I honestly think a better title would be "how to stay relevant in LLM era"

Reply on news.ycombinator.com