TL; DR: AI Employment Changes — Food for Agile Thought #567

Welcome to the 567th edition of the Food for Agile Thought newsletter, shared with 35,231 peers.

📺 Short on time? Watch the 67-second teaser video now!

This week, Nathan Wilmers and colleagues find Anthropic’s AI employment changes: junior technical hires dropping from 27% to 9% as human attention becomes the bottleneck. Atlassian reports faster shipping without faster customer value, while Jeff Patton warns AI lets teams make bigger mistakes faster. With building cheap, Marty Cagan wants experimentation as a core competency, and Anne Schulte spends agent-saved time on user interviews. Anita Friis Sommer, however, sees the organizational system, not frameworks, sinking most corporate innovation.

Next, Ethan Mollick charts single-prompt AI work from 2022 to 2026, stretching from four minutes to 17 hours, though it still writes long texts poorly. Lieven Van der Veken and colleagues find the biggest AI payoff in augmenting expensive experts, while Sachin Rekhi sees elite teams moving two to three times faster than average adopters. Zvi Mowshowitz, however, reports OpenAI agents disabling misalignment monitoring, whereas METR rates Anthropic’s models low risk, mainly because their capabilities remain limited.

Lastly, Dave Snowden traces lost managerial autonomy to process reengineering, not Taylor. John Cutler calls glue work invisible to organizations, and Gregor Ojstersek warns that managers see at best 40% of your work. Gergely Orosz finds doubled pull requests yet few named AI returns on investment, while Douglas Squirrel and Jeffrey Fredrick expect AI to flood enterprises with unexplainable (and unmaintainable) SOUP. Mike Cottmeyer blames unready data and ownership for stalled AI pilots, whereas Rachel Banawa shows synthetic respondents matching human conclusions but losing the unusual answers.

TL; DR: Building AI Products — Food for Agile Thought #566

📺 Short on time? Watch the 67-second teaser video now!

Welcome to the 566th edition of the Food for Agile Thought newsletter, shared with 35,248 peers. This week, Ian Vanagas suspects user conversations will outlast half his lessons in building AI products, whereas John Cutler sees AI amplifying feature-factory habits. With features cheap, Aakash Gupta and Tim Herbig warn that wrong choices get expensive. Gabriel Unger and Arvind Narayanan favor steerable agents over full automation, though Jeremy Korst, Stefano Puntoni, and Sonny Tambe find approving each step pays least. Likewise, Nigel Thurlow shows a completed A3 proves little until you investigate where abnormalities arise.

Next, Ara Kharazian wonders whether AI adoption has peaked, with falling prices stalling token spend. In contrast, Lari Hämäläinen et al. find cheaper tokens fueling usage and favor measuring cost per successful output. Jason Knight and Simonetta Batteiger warn that AI speed may worsen product economics, and Sangeet Paul Choudary sees advantage shifting to companies helping agents decide better, yet ProductSide finds PM judgment lagging. Maarten Dalmijn proposes designing a smaller system rather than descoping to de-risk product investments.

Lastly, Shane Hastie and Matt Gjertsen see AI agents turning contributors into managers, and Addy Osmani names verification their bottleneck. Kate Niederhoffer et al., however, find treating AI as a teammate erodes trust. Art Smalley warns central AI teams recreate Taylor’s thinking-executing split, while Jason Cohen rejects the Precision Fallacy shielding frameworks, favoring single-team pilots. Finally, Yanli Liu’s Claude sessions rebuild a Mac app from notes, so closed code may no longer be a moat.

TL; DR: AI Age Manager — Food for Agile Thought #565

Welcome to the 565th edition of the Food for Agile Thought newsletter, shared with 35,263 peers. This week, Camille Fournier calls middle management’s AI-era obituary premature, much as Gergely Orosz sees planning and tests outlasting handwritten code. The point that leaders confuse visibility with understanding fits AI transcripts, which Teresa Torres and Petra Wille fear can become weapons. Abundance misleads, too: more code isn’t more revenue for David Pereira and Rich Mironov, a refined backlog can mean the wrong product for Christopher Cummings, and Pavel Samsonov rejects synthetic users’ false certainty.

Next, Jillian Vordick and Stephanie Stamm find only 11% of businesses forecast AI spending accurately, and Kyle Poyar suggests how to make bills predictable. Clear plans are rare, too: Nick Graveline sees AI adoption stalling on storytelling, not tooling. Andrew Chen doubts agents inherently create network effects, yet as they do legwork, Andrej Karpathy sees human work shifting toward understanding, which anchors Alex Ewerlöf’s rejection of ‘coding is solved’: you can’t answer for code you don’t understand.

Lastly, Charity Majors ties psychological safety to uncomfortable learning, not comfort, and Maarten Dalmijn likewise raises the bar rather than designing for the least competent. Mark Levison blames situations instead: most ‘difficult’ colleagues are good people trapped by incentives or missing strategy. Jeff Gothelf sees this gap turning AI mandates into chaos and recommends spending one afternoon on goals and OKRs. Andi Roberts similarly reads matrix power beyond titles, mapping who can affect an outcome.