Leadership in Tech
Leadership in Tech
#298 – March 29, 2026
CEO wanted to cut R&D spending to match a low-cost competitor using a wrong metric
How to measure R&D investment
12 minutes by Phil McKinney
A chance meeting with HP's CEO in 2005 sparked a long debate over how to measure R&D investment. The CEO wanted to cut R&D spending to match a low-cost competitor, using a metric that measures input but not value. Pushing back was right, but the argument was framed in the wrong language. Phil explores the decisions that shape whether innovation survives inside large organizations.
Improve accessibility and speed up your development
sponsored by Level Access
How can development teams deliver inclusive, compliant products without slowing down? Testing earlier is just table stakes. Check out our on-demand webinar for tools and strategies you can use to accelerate progress by bringing accessibility into your existing development workflows.
Slow down to speed up
9 minutes by James Stanier
AI makes it tempting to build and ship faster than ever, but speed applied in the wrong direction just creates bigger problems faster. James argues the real leverage now lies in the decisions that come before writing any code. He suggests to slow down to clarify the problem, run pre-mortems, and validate assumptions first, then unleash AI's speed once the direction is clear. The teams that ship fastest over time are often the ones that know when to pause.
Governance: Documentation as a knowledge network
11 minutes by Frederick Vanbrabant
According to Frederick most company wikis fail not because of the tools, but because of how they are set up. The fix is to treat documentation like a web of connected pages rather than a folder dump. He suggests to keep each page focused on one concept, link everything together, and use landing pages to help people find their way in. This makes information easier to find, easier to maintain, and harder to let go stale.
How to build your own executive assistant with Claude Code
10 minutes by Obie Fernandez
Obie built a personal AI operating system by pointing Claude Code at an empty folder and asking it to organize everything. It now tracks team context, logs decisions with full rationale, preps meeting notes, and posts to Slack, all through plain conversation. Three weeks in, it covers 23 people and 9 strategy documents with no manual upkeep.
Twitching before you sprint
7 minutes by Mike Fisher
Sleeping rats twitch during rest to map their own bodies before performing any real movement. Mike points out that organizations could learn from this. Most companies launch, reorganize, and scale before truly understanding how they work, then pay a steep price when reality pushes back. Mike says that small, low-stakes experiments run early reveal how a system actually behaves, building real understanding before full commitment locks in the cost of being wrong.
CI was designed for a different era. Depot CI is Fast By Design.
sponsored by Depot
AI agents are shifting the bottleneck from writing code to integrating it. Your team can generate more code than ever — but it all flows through CI. And legacy CI wasn't built for this volume. Depot CI is a new CI engine: 2-3 second job starts, parallel execution, per-second billing at $0.0001/sec. Migrate your existing GitHub Actions workflows in minutes. Don't let CI kill your team's momentum.
And the most popular article from the last issue was: