Pinned Tweet
Six jobs, one terminal, and every one took a line of JSON I had to keep in my head.
So I built one screen for my agents. It turned out to look like a strategy game: units on a board, the moves each can make, a turn that plays out on its own. 🧵
1
27
I didn't write myself a note to be faster. I changed the engine's rules: a join may now wait a day for a person, with a test holding it there.
Chapter 22 of my agent-neuron series, with the run replayed step by step. Link below 👇
1
6
This is a preview of github.com/overfocus/agent-n… studio for editing and composing agent networks
25
disagree that industrial software can be "cleaned up in a day". But overall the point about crud apps and software factories and job justifications is well taken.
I've been thinking about this since we've been seeing reverse engineered and clones of major applications generated - I think the real thing happening is that a lot of the software being built is yesterday's software.
The coolest engineering is happening in software that we couldn't build last year, and I don't think people will be using slack clones and office suites and creative suites, and people won't need crud apps.
And I do think many jobs will be and are being wiped out.
I also think many orgs and companies will be wiped out.
All of that being said, I think there have never been bigger and harder problems to solve for.
i genuinely think a lot of people are struggling to cope with their job being completely wiped out and the software factory and custom harness pushes are ways for them to justify continuing to have large engineering teams.
The software factory is actually a factory that produces WORK.
> They need to validate each step in the software factory.
> More component based systems
> Constant maintenance
Meanwhile everything the software factory creates can be prompted by a non-dev & QA'd & then cleaned up by a dev in a day. All done at the same speed or marginally slower than the software factory if you don't include the cost of initial setup and maintenance. If you include the maintenance cost, software factory is clearly a cost center.
Not saying that any abstractions and automations are bad, just saying what I'm seeing out there when I talk to different companies. It feels like a lot of engineers are just scrambling to justify their roles instead of just enjoying the wave.
1
1
47
Kevin Ferron retweeted
i genuinely think a lot of people are struggling to cope with their job being completely wiped out and the software factory and custom harness pushes are ways for them to justify continuing to have large engineering teams.
The software factory is actually a factory that produces WORK.
> They need to validate each step in the software factory.
> More component based systems
> Constant maintenance
Meanwhile everything the software factory creates can be prompted by a non-dev & QA'd & then cleaned up by a dev in a day. All done at the same speed or marginally slower than the software factory if you don't include the cost of initial setup and maintenance. If you include the maintenance cost, software factory is clearly a cost center.
Not saying that any abstractions and automations are bad, just saying what I'm seeing out there when I talk to different companies. It feels like a lot of engineers are just scrambling to justify their roles instead of just enjoying the wave.
This week's Leftover Screws is mostly agent-neuron. I stopped adding chapters, made it carry real work, and let the work tell me what was missing. 🧵
1
33
Kevin Ferron retweeted
I asked @grok to generate what every interface and workflow for every job in every company might look like.
CEOs thought they had the advantage over choosing who they could stop paying because of AI are up to the same problem - their job is held only because of cult authority and isn't real.
1
1
54
I asked @grok to generate what every interface and workflow for every job in every company might look like.
I may be slightly mad and in AI psychosis but the egoic war taking place between:
1) I don't need other peoples skills
2) Nobody needs my skills
Pretty wild to observe in myself.
I already know what people with skills are going to say about that, watching them cling to the value of their knowledge and their special abilities. I know this because this is the battlefield of the internal war.
It's not like engineers didn't already have an "egomaniac with an inferiority complex" issue - with deeply arrogant self-criticism as well as pedantic criticism of other devs who approach problems and solutions in different ways.
Theres been a lot of pride in my industry - and theres always this loud group of experts who talk about industry standards and "how things are done professionally". They are almost always wrong, and almost always get listened to until their arguments no longer live up to the alternative. Without digressing, these are the people that said git wasn't a professional source control system, that rigid XML schemas were how web services would work, that OOP was good for.. anything.
I've been talking with a lot of people ... a LOT of people in different industries about AI impact. And the DK effect of people being easily impressed with output from LLMs is only fueling the delusions. And many of them are sending me articles and reports from incumbent experts about how AI is going to work and is working. I cannot express the incredible amount of people that actually think these incumbent experts are right.
I have no doubt that 2027 is going to be an obliterating year. Small teams, small global teams that have no corporate affiliation are going to fundamentally change every single.. every single thing in software, data etc.
You think your corporate data is the moat, but you have executives and staff pasting that data into random models and tools ALL OVER the internet.
And if you don't know about the darkweb MCP data brokerage marketplaces you literally have no idea whats happening.
And if you think regulations will save your company, you're missing that the entire LLM model from the labs is based on breaking IP law.
None of its legal, none of it ordained by the powers that be, and none of it is coming from boardrooms but it is very very very very much happening and the tooling around it isn't growing like wildfire - its growing like a supernova.
54
Kevin Ferron retweeted
some thoughts on writing software in nov 2026:
- it feels like magic
- the first 80% is 100x faster
- rewriting systems is also 100x faster
- opus 5.5 is very good at a lot of things
- "deep flow state" is much easier to reach
- almost anything can be built in typescript and rust
- building secure things is very, very hard
the optimization function is how to maximize agents working on goal with the fewest interruptions needed. some things that help:
- know where to trust vs guide (and keep retesting)
- give them agency like "i'm going to bed, get this done"
- if there is a reward function, ask them to hillclimb
- the socratic method works wonders
- skills and memories can often hinder
- asking agents to review themselves still helps
- cowriting a plan/spec/invariants helps sometimes
managing your own psychology is equally important:
- be a lot more ambitious about what you can achieve
- avoid dopamine traps / false signals of progress
- get good at parallelizing and context switching
there are also some things that don't quite work fluidly yet, but will get unlocked in the very near future:
- software factory or "find and fix the bugs"
- multiplayer agentic workflows
- agents managing other agents for high level outcomes
- organizational memory







