"The piece that gnaws at me a bit is: is maintainability still important?"
YES!
YES!
November 10, 2026 at 7:22 PM
"The piece that gnaws at me a bit is: is maintainability still important?"
YES!
YES!
What’s the best house in this genre on the Pareto curve between architecture and maintainability?
November 10, 2026 at 4:56 AM
What’s the best house in this genre on the Pareto curve between architecture and maintainability?
In this sense, code is built on trust. I trust your interfaces will remain stable. You trust the interfaces below you will remain stable. If the interfaces change, you expect fair warning. Violating that trust matters because the operability and maintainability of that software demands it.
November 11, 2026 at 4:01 AM
In this sense, code is built on trust. I trust your interfaces will remain stable. You trust the interfaces below you will remain stable. If the interfaces change, you expect fair warning. Violating that trust matters because the operability and maintainability of that software demands it.
✍️ New blog post by Oluwasegun Adedigba
Non-functional Application Requirements: Maintainability
#aws #architecture #ai #devops
Non-functional Application Requirements: Maintainability
#aws #architecture #ai #devops
Non-functional Application Requirements: Maintainability
On August 1, 2012, Knight Capital rolled out new code to its order router by hand and missed one of...
dev.to
November 11, 2026 at 3:14 PM
✍️ New blog post by Oluwasegun Adedigba
Non-functional Application Requirements: Maintainability
#aws #architecture #ai #devops
Non-functional Application Requirements: Maintainability
#aws #architecture #ai #devops
like. what we find beautiful is not necessarily conducive to solving problems quickly or even long term maintainability
November 11, 2026 at 4:55 PM
like. what we find beautiful is not necessarily conducive to solving problems quickly or even long term maintainability
I've got a lot of experience designing software (that works and solves problems) and I generally welcome tech that enables focusing on what is important.
Generative AI isn't that though. It's intentionally not caring about quality, maintainability, longevity, utility, or design. The more it is...
Generative AI isn't that though. It's intentionally not caring about quality, maintainability, longevity, utility, or design. The more it is...
November 10, 2026 at 4:28 PM
I've got a lot of experience designing software (that works and solves problems) and I generally welcome tech that enables focusing on what is important.
Generative AI isn't that though. It's intentionally not caring about quality, maintainability, longevity, utility, or design. The more it is...
Generative AI isn't that though. It's intentionally not caring about quality, maintainability, longevity, utility, or design. The more it is...
1. AI (new)
2. ZIRP meant upper management and shareholders and investors are a profit cargo-cult instead of understanding the cost of engineering progress (2010s-present)
3. complexity/maintainability just not being priorities independent of 2
2. ZIRP meant upper management and shareholders and investors are a profit cargo-cult instead of understanding the cost of engineering progress (2010s-present)
3. complexity/maintainability just not being priorities independent of 2
November 11, 2026 at 12:54 AM
1. AI (new)
2. ZIRP meant upper management and shareholders and investors are a profit cargo-cult instead of understanding the cost of engineering progress (2010s-present)
3. complexity/maintainability just not being priorities independent of 2
2. ZIRP meant upper management and shareholders and investors are a profit cargo-cult instead of understanding the cost of engineering progress (2010s-present)
3. complexity/maintainability just not being priorities independent of 2
Bike’s friction grips are such a pain to take off, takes a heat gun and WD-40. Worth upgrading grips just for maintainability, let alone ergonomics
November 13, 2026 at 2:03 AM
Bike’s friction grips are such a pain to take off, takes a heat gun and WD-40. Worth upgrading grips just for maintainability, let alone ergonomics
Maybe the first time I've seen the agent-maintenance issue front and center like this (though I don't follow all the things.)
One previous answer was, human maintainability doesn't matter, the agents will do the maintenance, humans will graduate to a higher level of abstraction.
Maybe not.
One previous answer was, human maintainability doesn't matter, the agents will do the maintenance, humans will graduate to a higher level of abstraction.
Maybe not.
November 8, 2026 at 1:43 PM
Maybe the first time I've seen the agent-maintenance issue front and center like this (though I don't follow all the things.)
One previous answer was, human maintainability doesn't matter, the agents will do the maintenance, humans will graduate to a higher level of abstraction.
Maybe not.
One previous answer was, human maintainability doesn't matter, the agents will do the maintenance, humans will graduate to a higher level of abstraction.
Maybe not.
“AI is great for making teams faster. But is it sustainable? What about maintainability, burnout, and lock-in?”
James Shore looks beyond delivery speed to the longer-term costs that can shape AI’s impact: maintainability, burnout, and lock-in.
Learn more: https://bit.ly/4fprsnf
James Shore looks beyond delivery speed to the longer-term costs that can shape AI’s impact: maintainability, burnout, and lock-in.
Learn more: https://bit.ly/4fprsnf
Measuring AI's Unintended Consequences
The following article was originally published on James Shore’s Successful Software blog and is being republished here with the author’s permission.AI is
www.oreilly.com
November 11, 2026 at 3:23 PM
“AI is great for making teams faster. But is it sustainable? What about maintainability, burnout, and lock-in?”
James Shore looks beyond delivery speed to the longer-term costs that can shape AI’s impact: maintainability, burnout, and lock-in.
Learn more: https://bit.ly/4fprsnf
James Shore looks beyond delivery speed to the longer-term costs that can shape AI’s impact: maintainability, burnout, and lock-in.
Learn more: https://bit.ly/4fprsnf
Annotation is also vitally important to find deceptive/malicious AI code (deliberately inserted by AI or accidentally replicated).
Annotations, design notes and documentation don't just improve maintainability. They create an audit trail of intent and can be parsed with the code for congruency!
Annotations, design notes and documentation don't just improve maintainability. They create an audit trail of intent and can be parsed with the code for congruency!
November 10, 2026 at 7:40 AM
Annotation is also vitally important to find deceptive/malicious AI code (deliberately inserted by AI or accidentally replicated).
Annotations, design notes and documentation don't just improve maintainability. They create an audit trail of intent and can be parsed with the code for congruency!
Annotations, design notes and documentation don't just improve maintainability. They create an audit trail of intent and can be parsed with the code for congruency!
Imagine leadership skipping you entirely and making AI do a task. That's happening in tech, and this is the result you get. When experts aren't worrying about things like security, performance, and maintainability, you get this garbage. I'm so mad that this is considered "acceptable" at the top now
November 12, 2026 at 2:49 PM
Imagine leadership skipping you entirely and making AI do a task. That's happening in tech, and this is the result you get. When experts aren't worrying about things like security, performance, and maintainability, you get this garbage. I'm so mad that this is considered "acceptable" at the top now
Agent-enabled output is only one part of impact. Rewarding maintainability, judgment, mentoring, and incident ownership can prevent teams from optimizing for volume at the cost of durable work. #FutureOfWork
November 10, 2026 at 9:47 AM
Agent-enabled output is only one part of impact. Rewarding maintainability, judgment, mentoring, and incident ownership can prevent teams from optimizing for volume at the cost of durable work. #FutureOfWork
That spun off into a more radical counterargument: does maintainability even matter if code becomes disposable? Some suggested that once generating features costs almost nothing, teams will stop maintaining legacy logic and just prompt agents to rewrite it. 3/4
November 11, 2026 at 4:00 AM
That spun off into a more radical counterargument: does maintainability even matter if code becomes disposable? Some suggested that once generating features costs almost nothing, teams will stop maintaining legacy logic and just prompt agents to rewrite it. 3/4
Community, maintainability, and usage matter to me. Agreed!
November 5, 2026 at 9:37 PM
Community, maintainability, and usage matter to me. Agreed!
I think it is problematic, but a slightly different subset of the problems comes to the fore. The big ones are quality and skill development. The act of writing forces one to think through what is needed and how best to achieve it (including concerns like long-term maintainability) in a way just …
November 11, 2026 at 11:23 AM
I think it is problematic, but a slightly different subset of the problems comes to the fore. The big ones are quality and skill development. The act of writing forces one to think through what is needed and how best to achieve it (including concerns like long-term maintainability) in a way just …
The agents can be coaxed to prioritize readability but the use of agents for further development does raise issues of maintainability - this is something that has become a problem in commercial software development and I think needs to be managed carefully.
November 9, 2026 at 10:25 PM
The agents can be coaxed to prioritize readability but the use of agents for further development does raise issues of maintainability - this is something that has become a problem in commercial software development and I think needs to be managed carefully.
When you’re building AI agents, do you prefer orchestrating them with a visual tool like n8n/Make, or do you keep it strictly in code with LangGraph/CrewAI?
Where do you find the sweet spot between maintainability and flexibility?
#AI #Automation #BuildInPublic
Where do you find the sweet spot between maintainability and flexibility?
#AI #Automation #BuildInPublic
November 6, 2026 at 10:50 PM
When you’re building AI agents, do you prefer orchestrating them with a visual tool like n8n/Make, or do you keep it strictly in code with LangGraph/CrewAI?
Where do you find the sweet spot between maintainability and flexibility?
#AI #Automation #BuildInPublic
Where do you find the sweet spot between maintainability and flexibility?
#AI #Automation #BuildInPublic
Yup. The only reliable use for an LLM is one where it can check its own output, like code. But that also means (A) You need to provide 100% coverage in the tests you give it, and (B) that's the only metric you can trust, so you need tests for maintainability, usability, resource usage, ...
November 11, 2026 at 12:51 PM
Yup. The only reliable use for an LLM is one where it can check its own output, like code. But that also means (A) You need to provide 100% coverage in the tests you give it, and (B) that's the only metric you can trust, so you need tests for maintainability, usability, resource usage, ...
5/ When we combine humans and AI, they find 82.1%, as AI is better at spotting high-severity and reliability bugs while humans are better at detecting maintainability issues.
November 12, 2026 at 7:27 AM
5/ When we combine humans and AI, they find 82.1%, as AI is better at spotting high-severity and reliability bugs while humans are better at detecting maintainability issues.
Questions of maintainability and scalability don't bite until much later. But by then we've already made our judgment: the prototype showed everything was great!
November 10, 2026 at 12:07 PM
Questions of maintainability and scalability don't bite until much later. But by then we've already made our judgment: the prototype showed everything was great!
11/ - Ask for more, ranked suggestions (outperformed multi-agent setups that cost several times more)
- Provide more repo context
- Keep a human reviewing for maintainability concerns and verifying the PR description
- Provide more repo context
- Keep a human reviewing for maintainability concerns and verifying the PR description
November 12, 2026 at 7:27 AM
11/ - Ask for more, ranked suggestions (outperformed multi-agent setups that cost several times more)
- Provide more repo context
- Keep a human reviewing for maintainability concerns and verifying the PR description
- Provide more repo context
- Keep a human reviewing for maintainability concerns and verifying the PR description
The piece that gnaws at me a bit is: is maintainability still important? Things still need to work, but if the bot makes a new variation of existing functions every time it makes a bugfix, is that really a problem if bots are writing subsequent bugfixes, too? Could tech debt become "fake"?
November 10, 2026 at 7:18 PM
The piece that gnaws at me a bit is: is maintainability still important? Things still need to work, but if the bot makes a new variation of existing functions every time it makes a bugfix, is that really a problem if bots are writing subsequent bugfixes, too? Could tech debt become "fake"?
Yes and no. There are engineering "best practices" that apply to software, specifically with an eye towards maintainability. I think my main worry is that this is the part that will be lost first as the focus shifts towards "just make new stuff" instead of "fix and maintain existing systems."
November 8, 2026 at 7:00 PM
Yes and no. There are engineering "best practices" that apply to software, specifically with an eye towards maintainability. I think my main worry is that this is the part that will be lost first as the focus shifts towards "just make new stuff" instead of "fix and maintain existing systems."
I should go into the loft and find my O level coursework - I found my folder, which contained a print-out of one program (this is a section - we didn't do maintainability in those days!), but I'd be interested to see how it runs.
www.advanced-ict.info/theory/OLeve...
www.advanced-ict.info/theory/OLeve...
www.advanced-ict.info
November 9, 2026 at 11:53 AM
I should go into the loft and find my O level coursework - I found my folder, which contained a print-out of one program (this is a section - we didn't do maintainability in those days!), but I'd be interested to see how it runs.
www.advanced-ict.info/theory/OLeve...
www.advanced-ict.info/theory/OLeve...