‘But [Vogel] wants developers shaped like the letter T – deep expertise in one domain, broad knowledge spanning many.
“Great developers are T-shaped,” he says. “They’re experts in their field who understand how their work fits into a larger system. You must broaden your T.”
What the hell shape I am goodness knows https://mastodon.social/@RuthMalan/115667191845472692
"Vibe coding is fine, but only if you pay close attention to what is being built. We can't just pull a lever on your IDE and hope that something good comes out."
That's not software engineering. That's gambling."
#WernerVogels, 2025
But ... when we say "vibe coding" we're specifically talking about *not* paying attention to what's being built. Usually because of lacking the base knowledge to understand it.
(1/3)
Werner Vogels ends his 14-year keynote streak by handing out printed newspapers and warning developers about "verification debt." His parting message: AI generates code faster than humans can understand it. The work is yours, not the tools.
That's what it means, right? It's right there in the name. It's generating software to fit the vibe you're going for. As opposed to engineering it according to a set of design goals, and within the operational limitations of the intended use cases.
(3/3)
Coda: automated code generation is an attempt to do what Open Source does naturally, but within the inherent limitations of proprietary software.
Open Source enables people to avoid the waste of time and talent that comes with reinventing the wheel over and over again, in different silos. By allowing the reuse of generic components, from the Linux kernel up, written by other people.
Code generation replaces code written and peer-reviewed by humans, with "AI" slop code that can be concealed.
This article about scAmazon CTO Werner Vogels reminds me of everything wrong with the Gee Whiz! school of tech journalism pioneered by Wired;
It's like reading an article about a guy who made excellent chains, and continued to work hard on making them better, even after learning they were being deployed in Stalin's gulags. It's a completely amoral approach to reporting, with zero consideration of the political economy of how tech is deployed and used.
(1/2)
Werner Vogels ends his 14-year keynote streak by handing out printed newspapers and warning developers about "verification debt." His parting message: AI generates code faster than humans can understand it. The work is yours, not the tools.
Vogels actually seems like a thoughtful and caring guy. But the reporter never once asks how he squares that with the huge damage scAmazon has done and continues to do. Much of it thanks to his invention of AWS.
(2/2)
"In Rwanda, the Ministry of Health operates something that stopped him cold. Their Health Intelligence Center displays near real-time data from healthcare facilities across the entire country on massive screens. Disease outbreaks. Maternal health outcomes. Policy decisions driven by actual data."
#MarcusSchuler, 2025
This reminds me of the stated goals of the Project Cybersyn team in Allende's Chile.
(2/2)
Werner Vogels ends his 14-year keynote streak by handing out printed newspapers and warning developers about "verification debt." His parting message: AI generates code faster than humans can understand it. The work is yours, not the tools.
At least as Cybersyn's goals were described in Evgeny Morozov's 2023 podcast The Santiago Boys;
https://choramedia.com/podcast/the-santiago-boys/
(2/2)
"On the Amazon river, he visited a beverage company called Groek bij Aker that gives young people economic futures so they don't abandon their villages for cities. He saw pink dolphins. It took him 21 years at Amazon to finally reach the river that shares the company's name."
#MarcusSchuler, 2025
Schuler doesn't ask why young people in Amazon villages need new "economic futures". He doesn't ask about the impacts companies like scAmazon are having on places like these.
Werner Vogels ends his 14-year keynote streak by handing out printed newspapers and warning developers about "verification debt." His parting message: AI generates code faster than humans can understand it. The work is yours, not the tools.
Jonathan Schofield
Strypey