IT Revolution
@itrevolution.com
Helping technology leaders succeed through books (The Phoenix Project, Team Topologies), events (DevOps Enterprise Summit), research, podcasts (The Idealcast), and more.
Install safety rails before someone drives off the loading dock: extra validation for AI code, more testing than before, clear ownership standards, rigorous engineering when on-call. The alternative headline? "Junior Dev and Chatbot Erase $40M in Revenue."
itrev.io/3LpQkPN
itrev.io/3LpQkPN
November 16, 2025 at 6:01 PM
Install safety rails before someone drives off the loading dock: extra validation for AI code, more testing than before, clear ownership standards, rigorous engineering when on-call. The alternative headline? "Junior Dev and Chatbot Erase $40M in Revenue."
itrev.io/3LpQkPN
itrev.io/3LpQkPN
Token burn per developer is the simplest AI adoption metric. Developers can only experience the upside if they learn it, and they can only learn by using it. Set targets, publicize leaderboards, give silly trophies. Friendly competition beats mandatory training.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
November 16, 2025 at 2:01 PM
Token burn per developer is the simplest AI adoption metric. Developers can only experience the upside if they learn it, and they can only learn by using it. Set targets, publicize leaderboards, give silly trophies. Friendly competition beats mandatory training.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
Every future outage will be blamed on AI. Lean into it. Host public retrospectives, document what happened, capture the new guardrail. Over time your org learns that accidents are opportunities for growth. Without psychological safety, engineers avoid AI entirely.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
November 15, 2025 at 9:01 PM
Every future outage will be blamed on AI. Lean into it. Host public retrospectives, document what happened, capture the new guardrail. Over time your org learns that accidents are opportunities for growth. Without psychological safety, engineers avoid AI entirely.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
Start with low-risk areas where the cost of mistakes is manageable. Gradually expand autonomy as teams demonstrate competence and alignment.
Organizations that balance autonomy with alignment move faster than centrally controlled competitors.
itrev.io/48PDd4d
Organizations that balance autonomy with alignment move faster than centrally controlled competitors.
itrev.io/48PDd4d
November 15, 2025 at 8:02 PM
Start with low-risk areas where the cost of mistakes is manageable. Gradually expand autonomy as teams demonstrate competence and alignment.
Organizations that balance autonomy with alignment move faster than centrally controlled competitors.
itrev.io/48PDd4d
Organizations that balance autonomy with alignment move faster than centrally controlled competitors.
itrev.io/48PDd4d
Kim and Yegge recommend giving developers access to at least 2 different AI models. Like learning a second language, it helps you understand the first one better. They started with one model for their book, expanded to five, then over twenty. Each had distinctive characteristics.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
November 15, 2025 at 6:02 PM
Kim and Yegge recommend giving developers access to at least 2 different AI models. Like learning a second language, it helps you understand the first one better. They started with one model for their book, expanded to five, then over twenty. Each had distinctive characteristics.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
Your approval process isn't providing control. It's providing the illusion of control while slowing innovation.
The teams moving fastest aren't the ones with the fewest constraints—they're the ones with constraints in the right places.
itrev.io/48PDd4d
The teams moving fastest aren't the ones with the fewest constraints—they're the ones with constraints in the right places.
itrev.io/48PDd4d
November 15, 2025 at 4:01 PM
Your approval process isn't providing control. It's providing the illusion of control while slowing innovation.
The teams moving fastest aren't the ones with the fewest constraints—they're the ones with constraints in the right places.
itrev.io/48PDd4d
The teams moving fastest aren't the ones with the fewest constraints—they're the ones with constraints in the right places.
itrev.io/48PDd4d
If engineers are silently using AI to complete 8-hour tasks in 5 minutes but reporting 8 hours, they're capturing productivity surplus for themselves. The organization never learns what's possible. Solution? Celebrate AI usage publicly, don't hide it.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
November 15, 2025 at 3:02 PM
If engineers are silently using AI to complete 8-hour tasks in 5 minutes but reporting 8 hours, they're capturing productivity surplus for themselves. The organization never learns what's possible. Solution? Celebrate AI usage publicly, don't hide it.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
Malcolm Gladwell's tipping point framework maps perfectly to AI adoption. Mavens who burn tokens before coffee. Connectors who spread tips between projects. Salespeople who turn weekend hacks into demos. Often the same people who accelerated DevOps adoption.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
November 14, 2025 at 10:01 PM
Malcolm Gladwell's tipping point framework maps perfectly to AI adoption. Mavens who burn tokens before coffee. Connectors who spread tips between projects. Salespeople who turn weekend hacks into demos. Often the same people who accelerated DevOps adoption.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
Individual autonomy: single contributors make progress without waiting. Team autonomy: small groups operate independently. User autonomy: consumers control their experience.
All three matter for Progressive Delivery.
itrev.io/48PDd4d
All three matter for Progressive Delivery.
itrev.io/48PDd4d
November 14, 2025 at 8:07 PM
Individual autonomy: single contributors make progress without waiting. Team autonomy: small groups operate independently. User autonomy: consumers control their experience.
All three matter for Progressive Delivery.
itrev.io/48PDd4d
All three matter for Progressive Delivery.
itrev.io/48PDd4d
One betting company built identity verification in Python (a language they barely knew) using AI-assisted development. The demo was so good, cautious production leadership approved deployment. Then their vendor hiked prices. The experiment became a critical production service.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
November 14, 2025 at 5:04 PM
One betting company built identity verification in Python (a language they barely knew) using AI-assisted development. The demo was so good, cautious production leadership approved deployment. Then their vendor hiked prices. The experiment became a critical production service.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
GitHub gives developers deployment autonomy but maintains responsibility for platform reliability. That's delegation, not abdication.
You can trust autonomy when you back it with engineering practices that minimize the cost of mistakes.
itrev.io/48PDd4d
You can trust autonomy when you back it with engineering practices that minimize the cost of mistakes.
itrev.io/48PDd4d
November 14, 2025 at 3:03 PM
GitHub gives developers deployment autonomy but maintains responsibility for platform reliability. That's delegation, not abdication.
You can trust autonomy when you back it with engineering practices that minimize the cost of mistakes.
itrev.io/48PDd4d
You can trust autonomy when you back it with engineering practices that minimize the cost of mistakes.
itrev.io/48PDd4d
Picture handing a chainsaw to someone who's only used a hatchet. Without guidance, they'll ruin it or accidentally destroy something, then tell everyone the tool is dangerous. This is exactly what happens with bad AI first experiences in engineering teams.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
November 14, 2025 at 2:02 PM
Picture handing a chainsaw to someone who's only used a hatchet. Without guidance, they'll ruin it or accidentally destroy something, then tell everyone the tool is dangerous. This is exactly what happens with bad AI first experiences in engineering teams.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
Your engineers are watching how YOU experiment with AI. Post snippets, brag about wins, share failures. When you hide behind policies, they sense fear. Visible optimism from leadership matters more than any training program.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
November 14, 2025 at 12:01 AM
Your engineers are watching how YOU experiment with AI. Post snippets, brag about wins, share failures. When you hide behind policies, they sense fear. Visible optimism from leadership matters more than any training program.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
Amazon's two-pizza team rule: teams should be small enough that two pizzas feed everyone and have everything needed to build, deploy, and maintain services independently.
It's all about where you place constraints.
itrev.io/48PDd4d
It's all about where you place constraints.
itrev.io/48PDd4d
November 13, 2025 at 11:01 PM
Amazon's two-pizza team rule: teams should be small enough that two pizzas feed everyone and have everything needed to build, deploy, and maintain services independently.
It's all about where you place constraints.
itrev.io/48PDd4d
It's all about where you place constraints.
itrev.io/48PDd4d
Quinn Slack at Sourcegraph built a leaderboard tracking token usage across the company. Their VP of Finance topped the charts one week. Engineers saw non-technical colleagues succeeding and felt inspired to dive deeper. Visibility sparks curiosity, curiosity sparks competition.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
November 13, 2025 at 9:02 PM
Quinn Slack at Sourcegraph built a leaderboard tracking token usage across the company. Their VP of Finance topped the charts one week. Engineers saw non-technical colleagues succeeding and felt inspired to dive deeper. Visibility sparks curiosity, curiosity sparks competition.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
Most organizations try to control risk through approval processes instead of through engineering practices.
This creates false confidence while slowing everything down.
itrev.io/48PDd4d
This creates false confidence while slowing everything down.
itrev.io/48PDd4d
November 13, 2025 at 7:02 PM
Most organizations try to control risk through approval processes instead of through engineering practices.
This creates false confidence while slowing everything down.
itrev.io/48PDd4d
This creates false confidence while slowing everything down.
itrev.io/48PDd4d
Before you schedule brown-bag lunches or commission analyst reports about AI adoption, open a chatbot and spend 10 hours coding with it yourself. Kim and Yegge are right—hands-on play will inform your strategy better than 100 pages of research.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
November 13, 2025 at 5:05 PM
Before you schedule brown-bag lunches or commission analyst reports about AI adoption, open a chatbot and spend 10 hours coding with it yourself. Kim and Yegge are right—hands-on play will inform your strategy better than 100 pages of research.
itrev.io/3LpQkPN
itrev.io/3LpQkPN
Autonomy only works with high psychological safety. Without it, you get either paralysis (people afraid to act) or recklessness (people avoiding scrutiny).
You can't mandate psychological safety, but you can build it through consistent responses to failure.
itrev.io/48PDd4d
You can't mandate psychological safety, but you can build it through consistent responses to failure.
itrev.io/48PDd4d
November 13, 2025 at 4:06 PM
Autonomy only works with high psychological safety. Without it, you get either paralysis (people afraid to act) or recklessness (people avoiding scrutiny).
You can't mandate psychological safety, but you can build it through consistent responses to failure.
itrev.io/48PDd4d
You can't mandate psychological safety, but you can build it through consistent responses to failure.
itrev.io/48PDd4d
Adobe lets users choose which AI model version to use. Developers can experiment without company-wide approvals. Product teams decide exposure timing.
Innovation at the speed of development without forcing disruptive changes on users.
itrev.io/48PDd4d
Innovation at the speed of development without forcing disruptive changes on users.
itrev.io/48PDd4d
November 12, 2025 at 8:02 PM
Adobe lets users choose which AI model version to use. Developers can experiment without company-wide approvals. Product teams decide exposure timing.
Innovation at the speed of development without forcing disruptive changes on users.
itrev.io/48PDd4d
Innovation at the speed of development without forcing disruptive changes on users.
itrev.io/48PDd4d
Get even more out of Progressive Delivery with bonus content designed to enhance your reading experience!
▪️Quick Start Guide
▪️The Four A's of Progressive Delivery
▪️Reader's Guide
▪️Quick Start Guide
▪️The Four A's of Progressive Delivery
▪️Reader's Guide
Progressive Delivery Bonus Content
PROGRESSIVE DELIVERY BONUS MATERIALS Access the bonus material and sign up for updates from IT Revolution. Get Access Enter your email address and share your proof of purchase to access the bonus…
buff.ly
November 12, 2025 at 6:02 PM
Get even more out of Progressive Delivery with bonus content designed to enhance your reading experience!
▪️Quick Start Guide
▪️The Four A's of Progressive Delivery
▪️Reader's Guide
▪️Quick Start Guide
▪️The Four A's of Progressive Delivery
▪️Reader's Guide
Delegation maintains accountability while transferring authority. Abdication transfers both.
Progressive Delivery advocates radical delegation, not abdication. You give teams power but retain responsibility for outcomes.
itrev.io/48PDd4d
Progressive Delivery advocates radical delegation, not abdication. You give teams power but retain responsibility for outcomes.
itrev.io/48PDd4d
November 12, 2025 at 3:04 PM
Delegation maintains accountability while transferring authority. Abdication transfers both.
Progressive Delivery advocates radical delegation, not abdication. You give teams power but retain responsibility for outcomes.
itrev.io/48PDd4d
Progressive Delivery advocates radical delegation, not abdication. You give teams power but retain responsibility for outcomes.
itrev.io/48PDd4d
Trust = (Competence + Reliability + Alignment) / Risk
This equation determines how much autonomy teams can handle. Build competence, demonstrate reliability, maintain alignment, and reduce risk through engineering.
itrev.io/48PDd4d
This equation determines how much autonomy teams can handle. Build competence, demonstrate reliability, maintain alignment, and reduce risk through engineering.
itrev.io/48PDd4d
November 12, 2025 at 1:02 PM
Trust = (Competence + Reliability + Alignment) / Risk
This equation determines how much autonomy teams can handle. Build competence, demonstrate reliability, maintain alignment, and reduce risk through engineering.
itrev.io/48PDd4d
This equation determines how much autonomy teams can handle. Build competence, demonstrate reliability, maintain alignment, and reduce risk through engineering.
itrev.io/48PDd4d
To all who served: Thank you for your dedication, sacrifice, and unwavering commitment to protecting our nation. We salute you today and always. 🇺🇸 #VeteransDay
November 11, 2025 at 8:03 PM
To all who served: Thank you for your dedication, sacrifice, and unwavering commitment to protecting our nation. We salute you today and always. 🇺🇸 #VeteransDay
Before Git developers worked serially, only one person could modify code at a time. Then, Git enabled parallel autonomous work. That wasn't just a technical improvement—it changed how software work could be organized.
itrev.io/48PDd4d
itrev.io/48PDd4d
November 10, 2025 at 10:01 PM
Before Git developers worked serially, only one person could modify code at a time. Then, Git enabled parallel autonomous work. That wasn't just a technical improvement—it changed how software work could be organized.
itrev.io/48PDd4d
itrev.io/48PDd4d
Autonomy without alignment = chaos. Alignment without autonomy = stagnation.
Most organizations lean heavily toward alignment and wonder why innovation moves so slowly.
itrev.io/48PDd4d
Most organizations lean heavily toward alignment and wonder why innovation moves so slowly.
itrev.io/48PDd4d
November 10, 2025 at 7:01 PM
Autonomy without alignment = chaos. Alignment without autonomy = stagnation.
Most organizations lean heavily toward alignment and wonder why innovation moves so slowly.
itrev.io/48PDd4d
Most organizations lean heavily toward alignment and wonder why innovation moves so slowly.
itrev.io/48PDd4d