A $2B Company's AI Strategy Was Written by Someone Who Never Used ChatGPT. The Backlash Has Begun.
Three communities — corporate critics, cognitive scientists, and content readers — arrived at the same conclusion this week: uncritical AI use is producing bad decisions, eroding judgment, and generating detectable slop. Here's what to do about it.
By Patin Team · Examples are illustrative composites
The professional immune response to AI hype arrived this week — and it came with receipts, data, and a free web tool. If you produce content with AI assistance, or make decisions about where AI belongs in your work, the bar just shifted.
What happened
On July 19, Simon Willison amplified Nik Suresh's "AI Mania" essay documenting corporate AI hype at scale: a C-suite leader built a $2B company's AI strategy without ever opening ChatGPT. Engineers rewrote working codebases to hit token leaderboards. Vendors pitched "100x productivity" numbers that nobody in the room could verify — and nobody pushed back because the social cost of scepticism felt higher than the cost of a bad decision.
The same week, an essay titled "Are we offloading too much of our thinking to AI?" hit 523 points on Hacker News (July 14). Commenters described using LLMs to draft responses to family conflicts, design products without forming their own opinions first, and process grief. A separate Ask HN thread proposing mandatory flags for AI-generated content cleared 1,000 points (July 13); site moderator dang confirmed HN already bans AI-generated text but acknowledged enforcement is difficult.
Willison responded with something concrete: the LLM Cliche Highlighter (July 17), a free web tool that colour-codes AI writing patterns in any text. "No fluff, no filler." "Let's unpack." "Comprehensive." "It's worth noting." Paste in a document and the tells light up.
What to do differently Monday morning
Three things shifted this week:
Critical evaluation of AI claims is now a professional skill. Knowing when AI is the wrong tool saves more money than knowing the perfect prompt. The executive who never used ChatGPT before building the AI strategy didn't make an AI mistake — they made a judgment mistake.
Delegating tasks and delegating thinking are not the same thing. AI writing a first draft of a meeting summary keeps your judgment intact. AI deciding what the meeting meant erodes it. The cognitive offloading debate isn't about whether to use AI — it's about where your reasoning ends and the model's begins.
AI-generated content is now detectable, and the audience has tools. The 1,000-point HN thread represents a readership that has seen enough AI copy to recognise it on sight. Willison just gave them a highlighter.
A content manager at a 60-person B2B SaaS company
She drafts a quarterly newsletter using ChatGPT as a starting point — brief in, 600 words out in under a minute. Before it goes to her editor, she runs the draft through the LLM Cliche Highlighter. Eleven phrases light up: three variations on "in today's fast-paced," two instances of "it's worth noting," and a closing paragraph that opens "As we look ahead."
She rewrites those eleven sentences from scratch. The newsletter reaches 8,000 subscribers. Nobody flags it. Her editor asks if she had a good week because the copy sounds like her.
An operations director at a 180-person professional services firm
Her CEO asks for an AI adoption plan by end of quarter. She has used AI tools three times in her career.
The AI Mania essay describes this situation directly: a plan built on vendor decks by someone who has never run a real task through the tools. She blocks three hours before she writes a single slide. She runs seven real work tasks through Claude and ChatGPT — meeting summaries, a client briefing, competitive research, a policy draft. Four come back usable with minor edits. Two need significant rework. One is worse than doing it herself.
Her plan is built on that data. Every recommendation in it maps to a task category she has personally tested.
The one thing
The backlash isn't against AI. It's against AI used as a substitute for judgment — in strategy, in content, in decisions that needed a person thinking. Willison's free highlighter is the bluntest version of the message: if your words don't survive a paste test, they might not be your words.
<BlogPracticeSection />Put this into practice
Reading is a start — but skill comes from doing. Try these drills now.
Reading about it only gets you so far
Patin turns this into five-minute drills that score what you write and tell you why. It's in closed beta — join the waitlist and we'll email you when your cohort opens.
Just want the writing? .
Keep reading on this
Reviewing AI Output: What to Actually Check
Reading it and approving it isn't a review — it's the exact filter AI output is best at passing. Interrogating is a different activity, it takes about ninety seconds, and it's most of what separates the people getting value from AI.
5 min readThree AI Experts Said the Same Thing This Week. It Changes What 'AI Skills' Actually Means.
Ethan Mollick, Charity Majors, and Simon Willison reached the same conclusion independently within five days. The professionals who get the most from AI are not the best prompters — they are the best judges.
5 min readPeople Are Tired of Talking to AI. The Quality Backlash Just Went Mainstream.
Three AI failure modes got named publicly in the same week — forwarding raw AI answers without reading them, optimising for speed over quality, and letting the model think while you sign off on the result. Each one has a skill that fixes it.
5 min read