Discussion across the Atmosphere
Why AI hasn’t replaced software engineers, and won’t
normaltech.ai
Coding agents as normal technology
Read on normaltech.aiLooking across the Atmosphere…
1 linkblog note · 47 Bluesky posts · 2 margin.at saves · 8 Semble saves
Highlights
-
AI will create massive structural shifts in how software is produced, which will have big impacts on which software engineers stand to gain or lose
-
seth @seth.computer
It seems clear that the amount of time people spend on getting computers to do new things will increase over time. This might take the form of building software, or managing complex workflows using agents, or something else. It will require a mix of software skills, AI skills, and domain expertise.
Discussion
-
Tim Disney @disnetdev.com
> Coding agents as normal technology
-
Chris Kanich @kaytwo.org
Really love the headline graphic for @randomwalker.bsky.social and @sayash.bsky.social's article here: I keep coming back to *taste* and *accountability* being core (human) contributions to knowledge work, and they map pretty perfectly onto "decide" and "deliver."
-
Ranjit Mathew @rmathew4bs.bsky.social
Coding has never been the bottleneck (or the only thing that SWEs do): “Why AI Hasn’t Replaced Software Engineers, And Won’t”, AI As Normal Technology . Via HN: #AI #Coding #SoftwareEngineering #Productivity #Jobs #Layoffs
-
boxwave 🐀 @maskhate.bsky.social
"it would be surprising for that not to correlate with productivity" Not to me, I'd find it unsurprising, because of the work of Prof Arvind Narayanan & Dr Sayash Kapoor There is also "slop grenades" that make other jobs harder so overall productivity stagnates
-
Frank Pasquale @frankpasquale.bsky.social
“We can think of many kinds of knowledge work, including software development, as a “decide-execute-deliver sandwich”. AI compresses the “execute” layer…but the other two layers resist automation in a way that will not be overcome by capability improvements alone.”
-
Pedro de las Heras Quirós @pheras.bsky.social
Coding agents as normal technology Arvind Narayanan and Sayash Kapoor Jun 11, 2026
-
draglikepull @draglikepull.bsky.social
This is a good post that touches on that point.
-
Jeremy Morgan @jeremymorgan.com
In the first year New York's WARN Act added an AI-layoff checkbox, 160-plus companies filed and not one checked it. Narayanan and Kapoor use the data to argue writing code was never the bottleneck. Deciding what to build and owning what ships is.
Saved in
-
-
seth @seth.computer
-
-
User & Agents @userandagents.org
-
Tim Disney @disnetdev.com
Connections
-
Wesley Finck @wesleyfinck.org
this related Taste Is All That's Left | Blog NotAShelf
-
Fred Jawa @fredjawa.bsky.social
this related AI is removing the middle class of software engineering Blog - Florian Herrengt
As it becomes easier to produce and deploy code, judgement becomes more important, not less.
-
zafarali.me @zafarali.me
this helpful What will be left for us to work on?
-
zafarali.me @zafarali.me
Why AI hasn’t replaced software engineers, and won’t Simon Willison’s Weblog explainer this
-
zafarali.me @zafarali.me
Writing Code vs. Shipping Code: Productivity Effects Across Generations of AI Coding Tools National Bureau of Economic Research supports this
-
Ian Vanagas @ianvanagas.bsky.social
The golden rules of agent-first product engineering PostHog opposes this
Argues code-writing was never the bottleneck, so agents won't replace engineers — counters the agent-does-the-doing premise.