frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Ask HN: Do you offload your coding to AI, or keep your skills sharp?

5•Hex08•1h ago
I used to be a software engineer, but these days I find myself being more of an agents manager: constantly switching between 3-4 tasks that my agents are working on, and making sure that the features get shipped, the bugs fixed, with correct and clean code. There's no doubt that I'm producing more value for the company, but I wonder if that's the best I can do for myself if the goal is staying relevant in the coming years.

I wonder: if the AI is writing all the code, what value am I bringing?

In an attempt to keep my critical thinking and coding skills sharp, I sometimes foolishly try to debug errors or implement a feature myself. But it feels like wasting time when an agent can do it in a fraction of time.

What is gonna be sought after in the coming years: someone that persevered with manual coding and debugging but is slow and uncomfortable around agents, or someone that has little coding and debugging skills left but knows how to orchestrate dozens of agents effectively?

I'm curious to hear what other software engineers on HN's approach is.

Comments

rgbrgb•1h ago
imo the valuable thing we bring as software engineers is and has always been solutions to problems with the contextually correct quality/speed/cost tradeoffs. with agentic coding, these factors all shift and we're all in the process of recalibrating. namely the speed dial can now be shifted way down.

to answer directly, i live in claude code now. and while it's a different shape of daily work, i'm still engineering, still evaluating those 3 factors and figuring out how to triage and solve user-problems much faster than i could 5 years ago.

ddxv•57m ago
I heavily use browser based LLM for bugs I care about. Code that is close to business logic like SQL is all 'by hand'.

Frontend stuff for me, which I'm bad at, I let agents do all of and I test the end result.

Other things are in-between, but time wise it's probably mostly me writing code. Lines of code is the LLM for sure.

charlesfrisbee•51m ago
I have the agent grill me/ask me questions about what has been implemented, often on a fairly technical/granular level of the feature is important enough.This is obviously way slower than just vibing the whole thing and ticking off a passing test suite (that the agent also wrote). But I find that doing this has allowed me to retain most of my problem solving and reasoning skills (which have definitely atrophied).

People are kidding themselves thinking they fully understand what is going on at a deep level for 3-4 (or more) parallel tasks constantly context switching. It's also an anxiety-inducing way to work and has led to a proliferation of agent management tools enabling max efficiency.

Focusing on systems thinking/observability/knowing when NOT to build something/using agents to fix/learn new things quickly seem to be where I see devs still bringing the most value compared to any other cowboy.

tripleee•17m ago
> if the AI is writing all the code, what value am I bringing?

What wouldn't an average person off the street be able to do?

I think you're mistaking typing characters on a keyboard for coding. Designing a system is coding. Knowing what refactorings to do to ensure things evolve in the right direction is coding.

Being able to recall all the syntax has lost value due to AI, but designing and leading the feature development hasn't and that can't (yet) be trusted to AI.

I'm not too worried about keeping skills sharp - if I'm ever put in a situation where they become relevant again I'm confident I can just relearn them.

RF applications in particle accelerators (2024) [pdf]

https://cds.cern.ch/record/2953465/files/2601.16760.pdf
1•joebig•38s ago•0 comments

Show HN: Upgrade to Neuro+ GBrain

https://gemini.google.com/app/b4b90aeb20c60c14
1•FDX2018•43s ago•0 comments

Euler's Eyes

https://valeman.medium.com/eulers-eyes-14dcad8fe694
2•ibobev•2m ago•0 comments

Engagement Speak is the hen-house with a growth chart attached

https://notsocommonthoughts.com/blog/engagement-speak/
1•kohlhofer•2m ago•1 comments

Virus-Like 'Jumping Genes' Became Our Partners in Evolution

https://www.quantamagazine.org/how-virus-like-jumping-genes-became-our-partners-in-evolution-2026...
1•ibobev•3m ago•0 comments

Biology Might Not Be Quantum, but Its Math Is Quantumlike

https://www.quantamagazine.org/biology-might-not-be-quantum-but-its-math-is-quantumlike-20260923/
1•ibobev•3m ago•0 comments

Code Critique: Intent, Drift, and Spotlight for AI-Generated Diffs at Scale

https://arxiv.org/abs/2607.29516
1•raahelb•3m ago•0 comments

Owners mourn spoiled food after firmware update bricks Samsung smart fridges

https://arstechnica.com/gadgets/2026/09/owners-mourn-spoiled-food-after-firmware-update-bricks-sa...
3•healsdata•3m ago•0 comments

Opaquer .NET Obfuscator

https://opaquer.net/
1•Rustemsoft•6m ago•0 comments

Show HN: AgentRun: DSL to turn agents into Workflows

https://github.com/Parcha-ai/agentrun
4•miguelrios•6m ago•0 comments

VSArena – The official open browser benchmark for embodied AI

https://vsarena.vercel.app
1•AranKair•7m ago•0 comments

Earth's magnetic field alters lifespans of fruit flies in surprising ways

https://www.404media.co/earth-magnetic-field-effects-on-aging-study/
1•anigbrowl•9m ago•0 comments

Designing a domain-specific research harness

https://sparsethought.com/2026/09/23/how-to-design-the-right-domain-specific-harness/
1•galsapir•13m ago•0 comments

"Good" AI

https://thoughts.wyounas.com/p/what-does-good-ai-mean
2•simplegeek•14m ago•1 comments

GoAccess – real-time web log analyzer

https://github.com/allinurl/goaccess
1•nobody42•15m ago•0 comments

Nuvio – Open-Source Alternative to Netflix for Your Own Media Sources

https://digitalescapetools.com/tools/nuvio.html
2•xabd•15m ago•0 comments

Algorithms of Isolation

https://boris.schapira.dev/notes/2026-09-algorithms-of-isolation/
2•speckx•16m ago•0 comments

Support for Local AI Models in the Antigravity SDK

https://developers.googleblog.com/introducing-support-for-local-ai-models-in-the-antigravity-sdk/
1•theanonymousone•16m ago•0 comments

Anna's Archive 502

https://annas-archive.gl
2•figbert•17m ago•0 comments

Show HN: I built a Bitcoin full node from scratch in Rust

https://github.com/realtonypark/mini-bitcoin
1•realtonypark•19m ago•0 comments

Senate Hearing on Flock Cameras [live]

https://www.youtube.com/watch?v=klE498xFjOY
3•Cider9986•20m ago•0 comments

Xilinx

https://www.xilinx.com/.html
1•MaxBareiss•21m ago•1 comments

So You Want to Build Your Own Data Center

https://blog.railway.com/p/data-center-build-part-one
1•gk1•21m ago•0 comments

Show HN: Suparelay – Call any phone in the world from a browser tab

https://suparelay.app
1•azyc•25m ago•0 comments

How we made claude.ai 3x faster in two weeks

https://claude.dev/blog/how-we-made-claude-ai-faster/
27•matthieu_bl•25m ago•4 comments

Show HN: I made a free and open source Muse/Instinct

https://github.com/jayzalowitz/skytwin
1•jayzalowitz•26m ago•0 comments

Show HN: Ceptile AI Search platform which combines web results with AI chat

https://www.ceptile.com/
1•ShauryaAgarwalr•26m ago•0 comments

Toys 'R' Us is opening 120 new stores as grownup collectors fuel toy sales

https://www.nbcnews.com/business/consumer/toys-r-us-opening-120-new-stores-rcna598543
1•bko•28m ago•1 comments

Stop Wasting Time. Learn Agentic RAG Patterns

https://medium.com/@NikoZero11/rag-is-easy-until-the-questions-stop-being-easy-a76186e11394
1•Natashiiy•29m ago•0 comments

MiMo-V3 is getting a new architecture. The core of it, HySparse2, is out today

https://twitter.com/_LuoFuli/status/2102766365190901957
1•ksec•30m ago•0 comments