Software Development

The last post left the engineering station with a one-line test: if the reasoning behind a decision would be useful to the engineer who touches this code in six months, write it down. That test tells you what deserves an entry, but it does not tell you how to recognize the moment, in the middle of a session, when you have crossed from accepting a suggestion into making a decision. Those are different skills, and the second one is a skill nobody ever had to develop before. ...

I ended the last post with a question: the next challenge is not building the Judgment Log. It is whether anyone writes in it once the deadline is two hours away. That question only has a useful answer if the artifact is light enough to actually use. So instead of arguing for it further, I want to show it. Take a fictional but familiar scenario. A checkout flow. A promotional window. A promo code validation feature built using AI-assisted development. It shipped. Three weeks later, it broke when the campaign introduced expired codes. In the post-mortem, nobody could answer the three questions that mattered: what did the PM cut and why, what did the designer choose between, and what did the engineer override. ...

In April, Meta employees burned through 73.7 trillion tokens in roughly thirty days. The company found out not because spending crossed some alarming threshold, but because an internal leaderboard, nicknamed Claudeonomics, had turned token consumption into a competition. Employees and teams were ranked by how much they used. The system did exactly what it was built to do: usage went up. What it could never show anyone was whether any of that usage produced something worth the cost. Meta is now dismantling the leaderboard in favor of a centralized monitoring platform called AI Gateway, built to track spending in real time and flag unusual spikes. ...

In 2012, I was given a project at Dalet that I did not fully understand at the time: move the entire company from waterfall to Agile. We were a mid-size product company with engineering teams spread across time zones, and waterfall was doing what waterfall always does at that scale, creating the illusion of control while real problems accumulated in the gaps between phases. The transition worked. And the productivity gains were not incremental. Engineers who had spent months waiting for approval to write code were suddenly shipping every two weeks. The business could see working software, give real feedback, and change direction without blowing up a six-month plan. It felt like the industry had discovered something fundamental. ...

In a post I wrote last December, I discussed the production triangle: the constraint that governs every production system, including software. You can optimize for two of three dimensions, time, quantity, and quality, but never all three. Push velocity while adding features, and quality absorbs the cost. Every time, without exception. This constraint is not new. Kahneman mapped the cognitive mechanism in Thinking, Fast and Slow: under time pressure, we disengage System 2, the slow, deliberate reasoning that catches the things that will hurt us later, and default to fast, intuitive pattern-matching. Kuutila et al.’s 2020 systematic review of the software engineering literature confirmed the outcome: increased throughput, decreased quality, and a root cause that almost always traces back to the pressure itself being a product of bad estimation. The triangle is not a theory. It is physics. ...

Anyone Can Prompt. Not Everyone Can Engineer. What the AI coding revolution actually changes, and what it doesn’t. The Translation Layer Is Gone I recently watched an IBM video that made a point I’ve been thinking about ever since. The speaker walked through the entire history of programming languages, from machine code and assembler, through COBOL and FORTRAN, to object-oriented, web, and scripting languages, and made a simple observation: every generation moved a little closer to the way humans actually think and speak. ...

In the previous post, I argued that defining decision boundaries is necessary if we want judgment to survive in an AI-accelerated development environment. When execution becomes cheap, the number of decisions explodes. Without clarity around who decides what, speed simply amplifies risk. But defining boundaries alone is not enough. The next question is more practical: how does work actually flow through those boundaries? How does an idea move from a conversation with a customer to something running in production, while AI accelerates the process without collapsing responsibility across roles? ...

Early in my career, coming from industrial engineering, I learned a simple rule that has shaped the way I think about software development ever since. Software may look abstract, creative, fluid. But underneath, it behaves like a production system. The best articulation of this idea appears in The Goal by Eliyahu M. Goldratt, one of the most influential books I have ever read. In production, every system revolves around three outputs you can influence: ...