Typing faster was never your edge. Taste is.

There was a brief cultural moment when typing fast looked like competence. Fingers flying. Slack green. Documents multiplying. I have never been convinced. Speed at the keyboard was always a thin proxy for value-useful for deadlines, useless as a strategy. Now that models can generate drafts in seconds, the proxy has finally collapsed in public.

Your competitive advantage was never typing faster. It was knowing what deserved to be typed at all. As writing becomes easier, thinking becomes more valuable-not less. The market is flooding with articulate output. The scarce resource is judgement about direction, trade-offs, and taste.

This piece is about why generation cheapens words without cheapening decisions, how to redesign work around thinking leverage, and what founders should stop measuring as “productivity.” Related: AI can write your content; it can’t replace your perspective.

If you are hoping for a list of keyboard shortcuts that restore advantage, stop here. Advantage now looks like clearer briefs, fewer public contradictions, and teams that can explain the product without opening a deck. That is slower to brag about than token counts. It is also harder to copy.

When output is cheap, selection is expensive

Economics still apply. When the cost of producing text, images, and variations falls, the bottleneck moves upstream. What should we build? Who is it for? Which claim do we defend? Which channel earns trust rather than noise? Those questions were always the work. Generation simply removed the comforting delay that used to look like progress.

Founders who respond to cheaper writing with more writing often feel busy and look interchangeable. Founders who respond with better questions look slower for a week and clearer for a year. Clarity compounds. Volume without selection does not. See what founders actually need isn’t more marketing.

• Cheaper drafts increase the cost of choosing which draft matters.
• Faster publishing increases the cost of being wrong in public repeatedly.
• More variations increase the need for a coherent spine.
• Typing speed was never the moat; decision quality was.

Thinking as leverage when generation is free

Leverage used to mean “I can produce more artefacts per hour.” That definition is obsolete. Leverage now means “I can improve the quality of the decision that every artefact depends on.” One sharp positioning choice can make fifty pieces coherent. Fifty pieces without a positioning choice make a content graveyard with good grammar.

If your calendar is full of drafting and empty of deciding, AI did not make you productive. It made your avoidance more eloquent.

A framework: decide → distil → delegate → defend

1. Decide: protect time for irreversible choices

Not every decision needs a war room. The ones that set direction do: audience, offer, positioning, pricing logic, proof standards, what you will not say. Block thinking time as if it were a client meeting. Because it is-future-you is the client. Strategy before screens still applies; see why I start every project without opening Figma.

2. Distil: turn thinking into a reusable spine

Thinking that lives only in your head does not scale. Distil it into a short spine: who you serve, what you believe, what proof you require, what voice rules you keep. A spine lets generation stay on-brand without turning the brand into the model’s median personality. For authenticity under automation, read don’t let AI become your personality.

3. Delegate: give repetition to tools and process

Once the spine exists, delegate the labour: outlines from brief, variations from master, summaries from research notes, first-pass FAQs from support themes. Delegation without a spine is abdication. Delegation with a spine is craft. The division of labour is the whole point of AI should handle the repetition; you should keep the judgement.

1. List weekly tasks that are repetitive but necessary.
2. Mark which ones require taste versus which require pattern matching.
3. Automate or assist only the pattern-matching column.
4. Keep a human gate on anything that commits the brand publicly.
5. Review monthly whether the gate is too loose or too precious.

4. Defend: maintain standards when speed tempts you

The temptation of cheap writing is to lower the bar because “we can always publish more.” More is not a standard. Defend a proof bar, a voice bar, and a usefulness bar. Consistency of standard beats consistency of volume. Related: the internet rewards consistency more than brilliance.

A fictional contrast: two product marketers

Marketer A treats AI as a volume machine: twenty posts a week, endless product updates rewritten as tips, a calendar that looks industrious. Engagement is soft. Sales says the site still cannot explain the product. Marketer B spends Monday on one positioning decision and one proof standard, then uses AI to expand a single essay into three formats. Fewer posts. Clearer pipeline conversations. Same tools. Different definition of work.

Strategy and judgement as the real throughput

Throughput used to be measured in pages shipped. Useful throughput is measured in reduced ambiguity. Did this week make the next campaign cheaper? Did this decision make the website less contradictory? Did this stake make the team stop arguing about adjectives? Those are thinking returns.

This is why product strategy before designing a screen still matters-and why marketing that feels like product design outperforms marketing that feels like content factory work. See the best marketing feels like product design and a product strategy framework before designing a screen.

• Ambiguity reduced is a form of shipped work.
• A clear no is often more valuable than ten drafted yeses.
• Judgement improves with feedback loops, not with more tokens.
• Teams that celebrate only output volume will hire for typing, not thinking.

The temptation to confuse motion with leverage

AI makes motion cheap. You can fill a content calendar before lunch. You can produce six homepage variants before the strategy call ends. Motion soothes anxiety. Leverage reduces future anxiety. The difference shows up two weeks later: either the team is still arguing about who the product is for, or the argument is closed and every artefact inherits the decision.

I have seen founders celebrate “we shipped forty assets this sprint” while the sales team still cannot explain the offer in one breath. That is not a content problem. That is a thinking debt problem, paid in public with fluent prose. The invoice arrives as rising CAC, soft replies, and a brand that feels busy rather than inevitable.

There is also a hiring implication. If you celebrate volume, you will recruit people who produce volume. If you celebrate decisions that reduce rework, you will recruit people who can hold ambiguity long enough to resolve it. Tools amplify whichever culture you already reward. Choose the reward carefully before you scale the stack.

1. Before generating assets, write the decision each asset depends on.
2. If the decision is missing, stop generating and decide.
3. Cap weekly public artefacts until the spine is stable for fourteen days.
4. Use AI to explore options inside the decision-not to postpone it.
5. Review whether last week’s output made this week’s decisions easier.

What research is useful for

For decision quality under uncertainty, Harvard Business Review remains a practical source on judgement and strategy. Nielsen Norman Group is useful when “thinking” must become usable interfaces and content. Daniel Kahneman’s work on Thinking, Fast and Slow is a reminder that fluency is not the same as correctness-especially relevant when models make fluency effortless.

How this shows up in my work at nau

I build systems where founders spend more time on stakes and proof, and less time pretending that drafting speed equals progress. Tools are welcome. Confusion dressed as output is not. If your growth feels like activity without advantage, start with why I care more about systems than campaigns.

The practical ask is simple: change what you celebrate in the weekly review. Celebrate retired ambiguities, approved stakes, and proof captured. Treat word count as an input cost, not a scoreboard. When the room laughs nervously at how little “content” you shipped while how much clarity you gained, you are doing it right.

Typing will keep getting cheaper. Thinking will not. Build the company around that fact, or watch competitors who still can decide outpace teams that only generate.

Stop racing a machine at its best sport

You will not out-type a model. That was never the contest. Out-think the week. Decide what matters. Distil it so tools can help. Defend the standard when speed whispers that anything will do. Competitive advantage was never in the fingers. It was in the judgement that told the fingers what was worth saying.