Martin Fowler · 2026-04-14 · 7 min read
"The problem is that LLMs inherently lack the virtue of laziness. Work costs nothing to an LLM. LLMs do not feel a need to optimize for their own (or anyone's) future time, and will happily dump more and more onto a layercake of garbage. Left unchecked, LLMs will make systems larger, not better."
"If we want AI systems that can operate safely without constant human oversight, we need to teach them not just how to decide, but when not to. In a world of increasing autonomy, restraint isn't a limitation, it's a capability. And in many cases, it may be the most important one we build."
Simon Willison · 2026-04-05 · 1 min read
"CLI tool now streams results as they are found rather than waiting until the end, which is better for large directories."
Kent Beck · 2026-04-03 · 5 min read
"The bottleneck isn't engineering. It's narrative."
"That's why all three moved together. The same investor class, the same stage, the same moment when 'trust us, it'll work out' stops being enough."
"Uncomfortable question: are the limits temporary—a bridge while supply catches up—or are they the beginning of a new equilibrium where heavy AI usage is a premium product, not a default?"
Martin Fowler · 2026-04-02 · 7 min read
"Cognitive debt lives in people. It accumulates when shared understanding of the system erodes faster than it is replenished. It limits how teams can reason about change."
"A consequence of System 3 is the introduction of cognitive surrender, characterized by uncritical reliance on externally generated artificial reasoning, bypassing System 2."
"If agents handle execution, the human job becomes designing verification systems, defining quality, and handling the ambiguous cases agents can't resolve. Your org chart should reflect this."
Martin Fowler · 2026-04-02 · 14 min read
"A well-built outer harness serves two goals: it increases the probability that the agent gets it right in the first place, and it provides a feedback loop that self-corrects as many issues as possible before they even reach human eyes."
"The human's job in this is to steer the agent by iterating on the harness. Whenever an issue happens multiple times, the feedforward and feedback controls should be improved to make the issue less probable to occur in the future, or even prevent it."
Kent Beck · 2026-04-01 · 1 min read
"TCR is a TDD variant where, if the tests fail, you reset automatically to the last known good state. Poof. If the tests pass, you commit."
"Can we use Skills to get the genie to work TCR-style?"
Martin Fowler · 2026-03-31 · 14 min read
"The standards that shape how an AI generates code, refactors existing systems, checks for security vulnerabilities, or reviews a pull request should not be tips shared on Slack or tribal knowledge carried in a senior's head. They should be versioned artifacts that encode 'how we do things here' into a form that executes consistently for everyone."
"A team standard encoded as an AI instruction does not depend on someone remembering to apply it. The instruction is the application."
"This is a systems problem, not a skills problem. And it requires a systems solution."
Kent Beck · 2026-03-30 · 1 min read
"The worse things get, the worse things get. Every time you try to intervene, it gets even worse."
"This is the most powerful tool I have for making big changes with small efforts."
Martin Fowler · 2026-03-26 · 4 min read
"But what we actually found were people organized around what they value—financial security, learning, human connection—watching advancing AI capabilities while managing both hope and fear at once."
"The spec document is the blueprint. The safety net is the test suite that catches the moment your code drifts away from it."
"The current administration has decimated the national security elements of both agencies through firings and forced resignations. People with decades of experience in building interagency and critical source relationships around the world, handling high-pressure, complicated investigations straddling classified and unclassified spaces, and acting in time to prevent violence and preserve evidence have been pushed out the door."
Martin Fowler · 2026-03-24 · 5 min read
"An Architecture Decision Record (ADR) is a short document that captures and explains a single decision relevant to a product or ecosystem. Documents should be short, just a couple of pages, and contain the decision, the context for making it, and significant ramifications. They should not be modified if the decision is changed, but linked to a superseding decision."
"But perhaps even more valuable, the act of writing them helps to clarify thinking, particularly with groups of people. Writing a document of consequence often surfaces different points of view - forcing those differences to be discussed, and hopefully resolved."