#342 – August 30, 2026
slow, bloated decision-making is now the main bottleneck in product development
The art of simplifying decisions
7 minutes by David Pereira
Slow, bloated decision-making is now the main bottleneck in product development. Most teams default to endless meetings or consensus-building, which wastes time and produces weak results. A better approach is to first identify whether a decision is reversible or not, then pick the right method: democratic for low-stakes speed, directive in a crisis, consensus for major commitments, and collaborative for most product decisions.
Choosing the right integrations platform for AI agents
sponsored by Nango
Conversations with 300+ engineering teams reveal a consistent evaluation checklist: beyond catalog size, teams weigh agent readiness, control, auth, reliability, observability, security, and predictable pricing. From the team at Nango, the open-source agent integrations platform with a pre-built catalog of 900+ APIs.
Selling out
20 minutes by Sean Goedecke
Adjusting your work personality is not the same as losing your authentic self. Drawing on Marx, the Situationists, Sartre, and American sociologists, the piece explores whether acting professional causes psychological harm. The conclusion is that it mostly does not, as long as you keep some mental distance from the role. Everyone chooses their own level of compromise between pure integrity and full corporate absorption.
AI-pilling our company: Lessons learned
7 minutes by Neil Rahilly
The best software engineers have always vastly outperformed their peers, but AI agents may change that dynamic for everyone. Sierra built a single internal agent called Pinecone, giving all 600+ employees access to tools across engineering, sales, and operations through one interface. The key insight was that business context, not raw intelligence, is now the main bottleneck. Usage is high, but the real goal is measuring actual outcomes, not just activity.
The management pretzel
4 minutes by Molly Graham
Keeping a team member can push managers into making decisions that only make sense for that one person, not the business. It's worth asking: would you make this change if they weren't here? If the answer is no, you may be contorting yourself around one person instead of leading clearly.
How HP quit and Apple won
8 minutes by Phil McKinney
Most failures repeat the same patterns. A smartwatch built years before the Apple Watch failed due to four classic ones: protecting existing profits over new ideas, ignoring warnings without changing decisions, leaving hard problems without a clear owner, and moving too slowly for the market. Spotting these patterns early is only half the skill. The other half is naming them out loud, writing them down, assigning an owner, and setting a deadline.
[Webinar] A defence of technical excellence for lovers and haters of LLMs
sponsored by IcePanel
The discipline of software development is changing faster than ever, while management expectations, already excessive, are reaching dizzying new heights of absurdity. Learn how to think rigorously, critically, and systematically to cut through the noise. Join us for a FREE talk on September 15 with Chris Simon.
And the most popular article from the last issue was: