On chapter two and already have neck muscle soreness from vigorously and frequently agreeing with Mik. I’ll have to switch up my response for the rest of the book.
https://itrevolution.com/product/output-to-outcome/
Remote
Matt Weagle
@mweagle@hachyderm.io
mastodon
4.6.6
1️⃣ People 👥, 2️⃣ Ideas 📚, 3️⃣ Technology 💻.
0
Followers
0
Following
41
Posts
Joined November 06, 2022
Blog:
Github:
Posts
“The single clearest strategic warning from the conference, repeated in different words
across security, governance and regulated-industry sessions: agentic AI amplifies
discipline and habits — positive or negative — that already exist in an organization.
Teams and organizations with weak testing culture, unclear risk ownership or poor
documentation hygiene will get worse, faster, not better.”
https://www.thoughtworks.com/content/dam/thoughtworks/documents/report/tw_future_of_software_engineering_europe_2026.pdf
Replying to
@adrianco@mastodon.social
@adrianco@mastodon.social Yes. Also measuring unplanned time allocated to keeping the system from more catastrophic failures. “What’s happening when nothing is happening” is that people are keeping it operational, often at the expense of declared commitments.
Replying to
@jeffmcneill@hachyderm.io
@jeffmcneill@hachyderm.io Yep. Herd immunity works when the herd follows the protocol and the protocol administrators are funded.
“This is what I call the Resilience Amnesia Cycle. It is a predictable pattern where organizations systematically forget why they invested in prevention, precisely because that prevention worked.”
https://newsletter.resiliumlabs.com/p/the-prevention-paradox
This is true for all types of organizational communication. It’s especially critical if that communication is likely to affect delivery timelines.
“The real power lies in having a shared language at the company, not in the table’s definitions.”
https://www.theengineeringmanager.com/growth/code-yellow-code-red/
Flying Blind is an excellent book about the rapid erosion of the engineering culture that sustained Boeing for decades. One of the tangible examples they highlighted was that teams were rewarded for assembly line velocity and individuals who called out quality issues were targeted.
https://www.museumofflightstore.org/flying-blind-the-737-max-tragedy.html
“I sat in all of those exit conversations thinking I’d hear something different each time, such as a better job offer, a life change, or something I could file away as circumstance. However, all three said a version of the same thing. One of them put it plainly:
“I cannot handle this pressure of making things right without any credit for it.””
https://leaddev.com/ai/ai-productivity-is-burning-out-your-best-engineers
Today is a day when I need to follow my own "go for a walk" advice.
The problem is that the organization didn’t create an environment where those departed people chose to stay and were rewarded for mentoring the next generation. What is the average job tenure for new entrants? Do they see huge comp bumps for “understanding” or for dive-and-save offers?
“The problem is that nobody left remembers why the batch job must run in that exact order on the last business day of the month. What died was not the language. It was the context. The decisions. The scar tissue.”
https://www.ivanturkovic.com/2026/07/07/ai-native-developers-gatekeeper-generation/
I am not the first, and I know I won't be the last person to make this observation, but code at rest has absolutely no value. The only benefit to code is the effect it produces. There is no functional difference between no code and code that doesn't work.
There's the product adage "people don't want a drill, they want a hole". When I reach for a drill, I expect it to work. If it doesn't, I now add the problem of how to responsibly dispose of my oddly-shaped paperweight to my original problem of wanting a hole.
Using an LLM to understand JIRAs implicit domain model and hidden constraints. AMA.
""I wish I had kept the message the CTO sent me. It approximately read, “Stop all releases immediately. If we can’t keep the service up, we shouldn’t be checking in any code.””
https://randsinrepose.com/archives/sometimes-your-job-is-to-get-in-the-way
Columbia Gorge in the Spring. #bloomscrolling
“A nuclear weapons program lost the ability to make a material it invented. The knowledge existed only in people, and the people were gone.”
https://techtrenches.dev/p/the-west-forgot-how-to-make-things
“Jevons paradox applies to software. When the cost of producing software collapses, demand explodes. Internal tools that were never worth building get built. Bespoke apps for teams of six get built. The demand has no ceiling I can see.”
https://www.ivanturkovic.com/2026/04/18/engineering-age-ending-ai-software-future/
“Much better is: Would you learn from this person? Would others at the company learn? In my experience this is not only true of every great executive, but it’s so clearly true that you can tell from the interview.”
https://longform.asmartbear.com/hire-better-than-you/
“Doing it better would be easy. Simply taking a little step back, thinking for a moment instead of just running after someone else’s goal, and adding a tiny bit of empathy to the interaction with our peers would make things a lot better. Instead of running after every “groundbreaking innovation of the day”, we could ask ourselves what we actually need and then ponder how AI can support us best in achieving these goals.”
https://www.ufried.com/blog/worst_enemies/
JIRA: not a fan
Mik's new book will be released on July 14, 2026 & I'm confident it will be a must read.
"In a world of abundant knowledge production, becoming AI native means shifting our organizations’ operating model and our professional focus from building outputs to delivering outcomes."
https://itrevolution.com/articles/sneak-peek-at-mik-kerstens-new-book-output-to-outcome
"On one hand, I think it’s absolutely incredible to think that users with all kinds of different disabilities could create their own assistive technology by describing to AI what they need."
"I’m using this enormously complex system, an LLM, to help me solve a problem that was created by software complexity in the first place."
https://surfingcomplexity.blog/2026/01/31/ashby-taught-us-we-have-to-fight-fire-with-fire
Many want to plant.
Fewer want to weed.
"The model appears optimistic, which is enough to make the whole thing sound temporarily reasonable. Each addition looks manageable in isolation. That is how the Kraken works. It does not attack all at once."
"The researchers call this the offloading paradox. Offloading can lead to better quality thinking and deeper learning — but only under two conditions..."
https://drphilippahardman.substack.com/p/the-cognitive-offloading-paradox
"The whole industry is caught in this loop now. You need the fear to justify the funding. You need the funding to build the thing. The thing then justifies more fear. Try telling the story in plain, simple language.”
https://om.co/2026/04/16/eat-your-words/?__readwiseLocation=
Replying to
@davemangot@hachyderm.io
@davemangot Yeah - it's a great article. Thank you Pete!
There's been a lot of technological change in 400 years, but power gradients, sunk cost, and production pressure persist.
“The lesson is what it took to make that learning happen: a catastrophic, public, entirely avoidable disaster, witnessed by thousands. Jacobsson already knew what needed to change. He simply hadn’t felt able to say so with sufficient authority until the failure was so large and so visible that a change of approach was essential.”
https://psychsafety.com/the-vasa-disaster/
Replying to
@brianleroux@indieweb.social
@brianleroux Moral support: "The other thing I have come to realize is how much of a productivity loss flaky tests can yield in a large SW project: the entire enterprise slows down as engineers fight with CI. Thus, de-flaking tests is one of the most beneficial things a project can invest in."
http://www.twtext.com.cdn.cloudflare.net/article/1291899271885623296
Two quotes that I'm reminded of reading testimonials about AI-fueled increases in code volume without regard to outcomes:
Drucker: "There is nothing so useless as doing efficiently something that should not have been done at all"
Deming: "A bad system will beat a good person every time"
"Every manufacturer optimizes for margin; the distinction lies in which variables are treated as fixed and which are allowed to move. Brands like Häagen-Dazs appear to treat ingredient integrity and sensory appeal as hard constraints. Others have treated those variables as adjustable, relying on incremental changes that remain below the threshold of consumer detection. One approach preserves product identity and adjusts price transparently. The other is optimization by dilution."
“The actual path to seniority isn’t learning more tools and patterns, but learning when not to use them. Anyone can add complexity. It takes experience and confidence to leave it out.”
https://terriblesoftware.org/2026/03/03/nobody-gets-promoted-for-simplicity/
"But distributed bugs don’t respect ceremony. They show up after retries, batching, and unlucky orderings nobody pictured when the code was written. They can sleep in stored state for months and then wake up on a Tuesday afternoon. LLMs are excellent at happy paths; the corners of distributed behavior are not happy places.
We’re about to try and automate the hardest part of our craft."
Replying to
@mattmay@mstdn.social
@mattmay "Not a good culture fit" does so much improper heavy lifting.
Replying to
@codinghorror@infosec.exchange
@codinghorror “I could code that in a weekend. How hard could it be?”
There are two problems in computer science:
1. We could do this better
2. We should not do this at all
Replying to
@justinsheehy@hachyderm.io
@justinsheehy @rstephensme Agree & thanks for sharing. A bias towards point-in-time, individualistic & measurable artifacts (PRs, lines, batch size) could make code-level AI seem like a productivity boost. This is an output lens.
Those with a bias towards reliable systems, where teams of people collectively understand how to evolve a system to meet new demands and exploit new capabilities, seem more likely to view code-level AI as actively detrimental. This is more an outcome lens.
Replying to
@shiftingedges@hachyderm.io
@shiftingedges I can imagine ICs experiencing less context-switching, feeling more productive, and in aggregate throughput is reduced.
“In regards to AI in particular, as some of these metrics contradict the industry narrative about the technology, I revisit the theory of constraints and hypothesize that writing code isn’t the bottleneck to deploying reliable applications.”
Replying to
@mikebroberts@hachyderm.io
@mikebroberts@hachyderm.io On point. The atomization of people and work is very disappointing.