AI Enthusiasts Race Against Time. AI Skeptics Race Against Entropy. Both Are Right.
Agentic AIMost conversations about AI adoption treat skeptics as obstacles and enthusiasts as the future. That framing misses the actual problem — both groups are right, about different timescales.
What Happened
Simon Willison surfaced a framing from Charity Majors that reframes the AI adoption debate with unusual precision. Enthusiasts face competitive pressure: teams that sit this out while competitors are hustling “could be out of business before the dust settles.” The concern is organizational survival across a multi-year competitive shift.
Skeptics face a different existential threat: shipping code faster than comprehension allows leads to “systems nobody understands, products burbling into incoherence, and on-call rotations that grind people up.” The concern is organizational survival across a multi-year accumulation of invisible technical and operational debt.
Both are legitimate. Both are backed by real-world evidence. The productive observation is that there is no natural feedback loop connecting the two groups — which means organizations rarely achieve coherent AI strategy by letting the two camps argue. They achieve it through deliberate leadership choices about which timescale to prioritize, and under what conditions.
Why It Matters
The typical enterprise AI conversation tries to resolve the enthusiasm-skepticism tension by finding a compromise position — “move fast but be responsible.” That framing is useless in practice because it doesn’t tell anyone what to do when the two timescales come into direct conflict on a specific decision.
The Majors/Willison framing is more useful because it separates the timescales explicitly. Enthusiasts are not wrong about competitive velocity. Skeptics are not wrong about system coherence. They’re optimizing for different failure modes, and both failure modes are real.
This connects to the agentic AI deployment gap between what companies promise and what they operate: many companies have enthusiast-driven roadmaps and skeptic-driven governance. Neither group wins, and the savings projections miss accordingly.
What This Means for Builders
For engineering leaders, the practical value of this framing is diagnostic: it predicts where AI initiative friction will appear and why. A team that can’t ship any AI feature without an extended security review has a skeptic-dominated governance structure operating on an entropy timescale. A team that has shipped fifteen AI features without any of them in reliable production has an enthusiast-dominated roadmap operating on a velocity timescale. The AI ROI problem in both cases is structural, not technical.
The resolution isn’t compromise — it’s explicit negotiation between the timescales. What’s the minimum viable governance process that doesn’t destroy velocity? What’s the minimum viable autonomy limit that doesn’t accumulate invisible entropy? These are design questions, not debate questions.
The So What
The AI enthusiasm-skepticism divide is a leadership problem with a concrete fix: get the two groups in the same room, name their timescales explicitly, and negotiate the specific policies that honor both constraints. Organizations that treat this as a culture war between optimists and pessimists will keep fighting it indefinitely. Organizations that treat it as a conflicting-requirements problem will solve it — and those are the ones that ship reliable AI systems at competitive speed.
Source: Simon Willison's Weblog
Content created with AI assistance and reviewed for accuracy.
Join the conversation
Stack Insiders is our free community for readers who want to go deeper — share resources, ask questions, and connect with others across every vertical we cover.
Join Stack Insiders →