I'm tired of the AI tone.
\nNot AI writing itself. AI is incredibly useful. I've probably used ChatGPT more this week than I'd like to admit.
\nI'm talking about the weird, increasingly recognizable way everything on the internet now sounds like it was written by the same extremely articulate 27-year-old who has never had a bad day.
\nYou know the tone.
\nEverything is a wedge.
\nEvery idea is not just X, but Y.
\nEvery paragraph contains an em dash — like this — because apparently the comma has been deprecated.
\nEverything is "a powerful shift."
\nSomething doesn't just solve a problem. It "unlocks a new way to think about the problem."
\nA product isn't useful. It "sits at the intersection of X and Y."
\nA company doesn't build something. It "reimagines how we..."
\nAnd every post somehow ends with:
\nThe future is already here.
\nThanks, Claude.
\nThe AI vocabulary
\nThere are certain words that have become suspicious.
\n"Wedge" is probably the biggest one.
\nEvery startup now has a wedge.
\n"We're starting with X as our wedge into Y."
\nNobody talks like this in real life.
\nI've been in startups for a long time. I've sat in hundreds of product meetings. I've talked to founders, engineers, customers, and investors.
\nI have almost never heard someone say:
\n«"This is our wedge."»
\nBut put the same person in front of ChatGPT and suddenly they're building a "land-and-expand wedge into the enterprise."
\nOther suspicious words include:
\nNone of these words are bad.
\nThat's what makes it worse.
\nThey're perfectly normal words that AI has managed to make feel like corporate spam.
\nAnd then there are the em dashes
\nWhy are there so many fucking em dashes?
\nSeriously.
\nAI loves an em dash.
\nIt uses them to connect thoughts that a normal human would separate into two sentences.
\nOr a comma.
\nOr just leave disconnected because humans don't actually write with perfect transitions all the time.
\nAI writing has this strange obsession with making every sentence flow beautifully into the next one.
\nReal people don't do that.
\nSometimes we start a sentence badly.
\nSometimes we use parentheses.
\nSometimes we repeat ourselves.
\nSometimes we say "and" four times.
\nSometimes we write:
\n«"I don't know. Maybe I'm wrong."»
\nAnd move on.
\nThat's writing.
\nThe perfect structure
\nAI also loves structure.
\nYou can usually recognize an AI-generated post from about 50 feet away.
\nIt starts with a provocative sentence.
\nThen:
\nHere's the thing:
\nThen three or five perfectly balanced points.
\nThen a counterargument.
\nThen a thoughtful conclusion.
\nThen a sentence about how "this isn't about X. It's about Y."
\nThen the inspirational ending.
\nIt's like every LinkedIn post has gone through the same screenwriting course.
\nAnd God forbid you make a list.
\nAI will give you exactly five bullets.
\nNot four.
\nNot six.
\nFive.
\nBecause five feels comprehensive.
\nThe fake vulnerability
\nMy favorite is the AI version of vulnerability.
\n"I used to think X.
\nI was wrong.
\nAfter spending years working on Y, I've come to realize Z."
\nThat's not how people talk.
\nThat's how a machine imagines someone talks after analyzing 40,000 LinkedIn posts.
\nReal vulnerability is messier.
\n"I completely screwed this up."
\n"I thought this was going to work. It didn't."
\n"I don't really know what I'm doing here."
\n"I changed my mind."
\n"I have no idea if this is right."
\nThose sentences are interesting because they're actually human.
\nThe problem isn't that AI writes well
\nIronically, the writing is often technically better.
\nThat's the problem.
\nAI removes all the little imperfections that tell you a human wrote something.
\nThe weird phrasing.
\nThe unnecessary joke.
\nThe sentence that goes on too long.
\nThe opinion that doesn't quite fit the rest of the argument.
\nThe typo.
\nThe "I don't know how to explain this, but..."
\nThose things give writing a personality.
\nAI optimizes them away.
\nAnd when everyone uses the same model, everyone optimizes toward the same mean.
\nWe are slowly converging on one universal Internet Voice™.
\nPolite.
\nClear.
\nStructured.
\nNuanced.
\nOptimistic.
\nSlightly corporate.
\nOccasionally profound.
\nAnd completely fucking interchangeable.
\nPlease sound like yourself
\nI'm not arguing that we should stop using AI to write.
\nQuite the opposite.
\nUse it.
\nHave it rewrite your terrible first draft.
\nHave it fix your grammar.
\nHave it make your argument clearer.
\nHave it turn your 2,000-word ramble into something people can actually read.
\nBut then put some of yourself back into it.
\nDelete the em dashes.
\nRemove "unlock."
\nKill the "not X, but Y."
\nDelete "wedge" unless you are actually talking about a wedge.
\nSay what you mean.
\nUse shorter sentences.
\nHave an opinion.
\nBe slightly weird.
\nDon't make every paragraph perfectly symmetrical.
\nAnd for the love of God, occasionally write a sentence that sounds like something you would actually say out loud.
\nAI is getting incredibly good at sounding intelligent.
\nThe competitive advantage now might be sounding human.
\nPost draft is ready, do you want me to make it shorter?
","headings":[]}}