Suppose that 95% of enterprise AI transformations fail. How does that compare to the failure rate of normal enterprise IT projects?
This might seem like a silly question for those unfamiliar with enterprise AI projects - whatever the failure rate, surely it can’t be close to 95%! Well. In 2016, Forbes interviewed the author of another study very much like the NANDA report, except about IT transformations in general, and found an 84% failure rate.
Programmers irrationally chose based on brand identity
As many of us say, technology is fashion and passion. Programmers identify with the language they use just people do with brands (beer, shoes, cars, political parties).
🔗 Why engineers can’t be rational about programming languages
Man-razors aren't the best
Use women’s razors for a better, more comfortable shave, he says.
🔗 How Gillette Used Gender to Make Face Shaving Needlessly Worse for Men
Organizations are measuring DevEx metrics more, JetBrains survey
JetBrain’s The State of Developer Experience and Developer Productivity survey is out. It’s an oddly, narrowly focused survey. Mostly on metrics used to measure developer experience.
Above, the metrics most used for tracking/rating/metric’ing development..As the conclusion says, measuring DevEx is getting better:
The good news is that companies are getting more serious about tracking developer experience, not just productivity. The drop in irregular DevEx assessments – from 53% in 2024 to 29% in 2025 – shows a positive shift toward more structured and thoughtful measurement.
I’m hear in the AMS KLM and lounge, and because the business people are blaring their calls on the speaker, I realize that we’re all talking about AIs. Even if you have a VoIP app. I mean, including me. Just not VoIP.
They all have a take on “agentic.” “Hey, guys, here’s the thing about agentic…”
I just heard the phrase “break out economics” and “offshore agent.”
Loving it…?
Links and original content
Original contentThe Enterprise Turing Test, Software Defined Talk #544: “This week, we discuss Claude’s new Excel skills, whether AI is augmenting or automating humans, and the latest developer surveys. Plus, AI making the command line cool again!”
The World Wide Web, Content, Work, Blogging Adventures, Short Videos, etc., with Russell Davies, Software Defined Interviews #112: we talk with Russell Davies about…all sorts of things, very content-y, advertising, being interesting. Just, you know, lots of delightful “and stuff” that results from someone who “mucks about on the internet.
There’s still enterprise marketing writing that beings “In today’s fast-evolving technology landscape…”
How Tanzu Platform Fixes the AI Deployment Problem
You need a proper AI platform:
Interested? Try a 90 day trial of the Tanzu Platform: https://trytanzu.ai.
3 + 1 Cloud Foundry Stories
We don’t tell Cloud Foundry stories enough. Here are four:
This talk was also a request to make more “case studies” and “customer references,” that is stories. So, I try to go over how to do that as well. At the last Cloud Foundry Day, I spoke more generally about the need for more Cloud Foundry marketing. Maybe I’ll become “that one guy who talks about Cloud Foundry marketing at CF Day”…if they let me.
AI video generation is very good, but its current prudishness will limit art and general use
AI-generated videos and images could be significantly “better” without current technological and content restrictions.