TAG

# Speak It Into Being

2026-07-20

Why Does AI Keep Getting It Wrong? It's Not Dumb — You Didn't Say It Clearly

You ask AI to build something, and eight or nine times out of ten it comes back wrong, and you think "this AI is useless." But after a few months of doing this every day, I'm more and more sure of one thing: when AI gets it wrong, most of the time it isn't dumb — I didn't say it clearly. Same requirement, phrased differently, and it nails it on the first try. There's a very plain truth underneath this — AI can't read minds. It does what you said, not what you meant. This piece lays out the requirement-writing method I've worked out: the five parts of a requirement that lets AI get it right the first time, the five most common ways of not saying it clearly and how to fix each, and three moves that turn AI from a hand into a brain. All of it is copy-pasteable.

2026-06-28

Becoming an AI-Era PM 06 | Speak It Into Being: Turning a Clear Idea Into a Clickable Product in One Sentence

This is the sixth piece in the series Becoming an AI-Era PM, and the opening of the "build it" half. Mindaugas turned an idea into a product with paying users using Lovable — without writing a line of code; in December 2025 Lovable raised a $330M Series B at a $6.6B valuation. "Speak it, and AI builds it" is no longer a slogan. But speaking it into being isn't type-one-line-and-walk-away — it's a loop, it has craft, and it all hinges on that one sentence being clear. This piece gives you four things you can actually do: ask for something that runs before you say it all, run it for real instead of trusting "done," change one thing at a time and watch it move, and say it clearly so the building follows.

2026-06-24

Becoming an AI-Era PM 05 | Leave It Vague and AI Will Fill the Gaps for You

This is the fifth piece in the series Becoming an AI-Era PM. You tell AI "build me a login," and in one breath it settles a dozen things you never mentioned: email or phone, how many wrong passwords before it locks, how long the lock lasts, what the error message says. AI doesn't ask you back the way a person would — it's a yes-machine: it does what you said, not what you meant. The moment a requirement goes fuzzy, it fills the gap with the most generic default, and that default is usually not the one you wanted. OpenAI's Sean Grove says code is only 10–20% of a developer's value; the other 80–90% is saying clearly what to build. This piece gives you four things you can actually do: swap adjectives for numbers, write out every state, list the edge cases, and self-check with a zero-context test.