Writing
Essays on AI products, the business of building them, and solving problems worth solving. Each one makes an argument.
- Google wins consumer AI on distribution Intelligence is commoditising, so the model stops being the moat. What's left is distribution and a business that profits from giving intelligence away, and Google is the only company with both - fighting Nvidia, Apple, Amazon, Microsoft and Meta each on one front while it works all five.
- Context pruning is a bet on the future When an agent's window fills, the obvious move is to drop the oldest, biggest tool results. That's a cache-eviction bet you can't make optimally without seeing the future, and the right one depends entirely on your workload.
- Mechanistic interpretability as generative art If a network has learned a concept, that concept is a location in its embedding space. Steer a generator toward that location and you don't get a diagram of what the model knows - you get a picture of it.
- AI is an interface The interface was always the tax you paid to use a system: its menus, its query language, its API. AI's most durable job is to take that tax off, by turning plain language into the task you meant and the systems that carry it out.
- Three gaps: coverage, synthesis, intent Most AI insights requests get treated as a synthesis problem. That's the wrong reframe. There are three stacked gaps - coverage, synthesis, intent - and you can't skip a layer without trust collapsing underneath you.
- Live billing as a forcing function I turned billing on at hey anna later than the principle deserved, and learned something about my own discipline in the process. 'I'll charge later' is almost always the wrong call, and here's why.
- The unit economics of a one-person AI product hey anna's variable cost is about 40% Claude API. That single fact rewrites the pricing, the acquisition maths, and whether viral growth is load-bearing. The SaaS playbook assumed COGS was a rounding error; it isn't anymore.
- Make every AI claim clickable Customers don't want AI that hands them the answer. They want to verify it. Deep-research-style citations are the line between an AI feature people adopt and one they paste into a doc and never open again.
- Measuring the platonic representation The platonic representation hypothesis says capable models converge toward one shared picture of reality. A language-anchored multimodal encoder lets you test a sharp version of it: encode one concept through four senses and measure what they agree on.
- The Wayback Machine is an investigative tool The sharpest cold open isn't you versus a competitor. It's their own public X against their own public Y - pricing page against blog, hero copy against last year's hero copy. The Wayback Machine and a diff are the cheapest research nobody runs.
- Cancel the meeting When the data isn't ready ten minutes before a stakeholder readout, the move is to cancel and send it async later. Forcing the numbers into a calendar slot anchors everyone on figures you'll want back tomorrow.
- Stated versus revealed preference, in booking data Travel marketing assumes people book what they say they want. The booking data says otherwise, especially on luxury tier and trip length. The gap between stated and revealed preference is where the useful product decisions live.
- Metacognition is the unlock Model progress has moved through paradigms: reactive, then reasoning, then agentic. The next is metacognition - thinking about its own thinking - and it's what separates a model that repeats mistakes from one that compounds on them.
- The trust-calibration tax The hard part of AI UX isn't generating answers. It's teaching users when to trust the system, when not to, and how to recover when it's wrong. That work is the real cost, and most teams never budget for it.
- Not every AI feature should be a chat Enterprises trust AI for invisible categorisation and distrust it for reversible work behind a chat box. 'Chat, move this five pixels' is all-or-nothing and risky. Sometimes the right surface for an AI feature is a button.
- Context engineering is the design surface Prompt engineering treats the words as the lever. Context engineering treats the whole context window as the design surface, with a stage and an artifact for each step. It's the better abstraction for production work.
- AI should be a dumb renderer The default pattern for AI insights dumps data into a model and asks it to count, compare, and conclude. That's the wrong order. Precompute the numbers deterministically; let the model render and narrate them.
- Why I still write code as a product leader Writing production code as a product leader compresses decision latency: faster iteration, feasibility you can check yourself, AI behaviour you can debug directly, and no organisational telephone.
- The four-mode product manager Strategy, market analysis, solution architecture, implementation. The old role split these across people and handoffs. The job now is to move between all four in a single conversation.
- Production AI is mostly workflow design The model-intelligence obsession misreads where production AI actually succeeds or fails. Across government, enterprise, and consumer, the wins came from orchestration, retrieval, evaluation, and fallback handling, not a smarter model.
- AGI won't be one big brain The monolithic-superintelligence story is the wrong mental model. Real general intelligence looks more like an orchestra of specialists than a single giant model.
- Stop micromanaging your AI Most people prompting models have quietly become managers - and they're managing badly. The fix is to build systems for context, not to craft better one-off prompts.
- Generalists are a startup's secret weapon The work system optimises for specialists - a hangover from the assembly line. Early-stage startups need the opposite, and most hiring processes can't see it.
- Bullshit jobs and the missing 15-hour week Keynes predicted a 15-hour work week by 2000. The technology arrived; the leisure didn't. The gap is filled with work that even the people doing it suspect is pointless.
- Happiness is a verb Aristotle's eudaimonia is not a state you arrive at. It's an activity you perform. That distinction changes how you should spend a life.
- The intelligence illusion AI keeps hitting milestones that used to sound terrifying, and they keep landing as boring. That reaction says something specific about what intelligence actually is.
- AI welfare: foresight or premature? Anthropic hired an AI-welfare researcher. The question is real, the uncertainty is genuine, and the honest position sits between dismissal and panic.
- Most problems are information problems A specific claim: nearly every failure - in product, in business, in life - traces back to someone missing a piece of the puzzle. Fix the information flow and most of the rest follows.
- Chain of thought, and where it breaks Asking a model to reason step by step reliably improves its answers. The catch most advice skips: in a production app, that visible reasoning is often the last thing you want.
- Prompt chaining: split the work, raise the floor Asking a model to do one complex thing in a single call invites failure. Breaking it into a chain of focused calls makes each step more reliable and easier to debug.
- Simple prompting: less magic, more method After thousands of prompts in production, the lesson is that prompt engineering isn't about magic words. It's clear thinking and structured communication, and it reduces to three principles.
- Doing less to get more done The product manager who has a hand in everything is the bottleneck. Stepping back isn't abdication - it's leverage, and it compounds.
- Prompting with frameworks the model already knows Structure your prompts the way a consultant structures a brief, then borrow a framework the model was trained on. You give it a running start instead of describing every step.