Leadership in Tech
Leadership in Tech
#310 – May 10, 2026
individual AI productivity gains don't automatically become organizational ones
When everyone has AI and the company still learns nothing
12 minutes by Robert Glaser
Individual AI productivity gains don't automatically become organizational ones. Companies are past the "buy seats and run training" phase and into a messier stage where adoption is uneven, partially hidden, and disconnected from real learning. The key question isn't whether people use AI, but whether the organization captures what changed because of it. That requires linking operational control, loop-level feedback, and capability sharing into something that builds learning velocity over time.
Unblocked: The context layer for modern engineering teams
sponsored by Unblocked
Whether it's written by an agent or an engineer, the difference between compilable and mergeable code has always been context. Unblocked turns code, docs, tickets, and conversations into actionable context, so engineers move faster and agents stay on track.
Build the right thing
7 minutes by Mike Fisher
Mike argues that successful product teams focus on customer outcomes before financial returns. Using the Wright brothers and Samuel Langley as examples, he shows how innovation suffers when organizations demand ROI too early. Experiments should prioritize learning, problem-solving, and customer value, while financial metrics should guide later-stage optimization. Great businesses emerge from discovering what truly matters to customers, not from optimizing spreadsheets too soon.
Early and late-stage hypergrowth
3 minutes by Will Larson
Will argues that early and late stage hypergrowth require different leadership strategies. Early on, teams focus on solving one big problem at a time, so expanding an existing leader's scope works well. Later, everything needs solving at once, so dedicated leaders for each area matter more. AI can speed up the early stage, but handling the late stage complexity remains an open and important question.
The brag document
4 minutes by Ben Balter
Good work doesn't promote itself. Keeping a running log of your wins, with specific outcomes and links, fights recency bias and makes self-assessments easy. Update it every Friday. Share highlights with your manager often, long before any promotion talk begins.
The Vasa disaster
11 minutes by Tom Geraghty
The Vasa disaster shows how organisations can fail when pressure, poor communication, and fear of authority prevent problems from being raised. Built in 1628 for the Swedish king, the warship was repeatedly redesigned, overloaded, and known to be unstable before launch. Despite warning signs, nobody stopped the project. The ship sank on its maiden voyage, becoming a lasting example of how power gradients and lack of psychological safety can turn known risks into disaster.
Paging is just the starting gun. Don’t ignore the rest of the race.
sponsored by incident.io
Don't just trade one UI for another while keeping the same alert fatigue. Most migrations fix the 10% (the alert) but ignore the 90% (the workflow). incident.io uses your tool move to solve the ownership and inventory gaps legacy tools miss. Build a real service catalog and a scalable model that stops burnout before it starts.
And the most popular article from the last issue was: