You're Using AI Wrong. Here's How to Tell.
The productivity flex is real, it's everywhere, and it's making you look busy while your judgment quietly erodes.
Everyone is talking about AI. Everyone is posting about AI. Everyone is generating 7,000 agents on the weekend and building websites by Monday and showing up to investor meetings with beautiful product specs and financial models and go-to-market strategies that took forty-five minutes to produce. And Barry O'Reilly has a name for all of it: the productivity flex. And he's not impressed.
Why This Conversation Matters
Barry O'Reilly has been advising boards and executive teams on enterprise transformation long enough to have watched the same mistake play out across every technology wave — and AI is just the latest, loudest version of it. He's the co-founder of Nobody Studios, a global top 10 AI venture studio building 100 companies in five years, and the author of three books including Unlearn and Artificial Organizations. He's also dyslexic, which turns out to be directly relevant to the best thing he said in this conversation. Barry doesn't talk about AI the way most people do. He talks about behavior, judgment, and what it actually means to do your best work — and then backs in from there.
The Signal
Start with yourself. The tools come last.
Barry's framework for AI adoption is deceptively simple: first, understand how you do your best work. Then, identify the tasks where you create the most value. Only then — once you've answered both of those questions honestly — do you start thinking about which tools might help. Most people do this completely backwards. They grab the tool first, poke around with it, and then wonder why nothing changed. Barry's stat on this is staggering: 83% of transformations fail precisely because they start with tools instead of behavior. The tool is never the transformation. Behavior change is the transformation.
The productivity flex is real, and it's everywhere.
Nobody Studios sees it constantly. Founders show up with pitch decks, product specs, go-to-market strategies, full data rooms — the kind of documentation that used to signal serious preparation. And then the conversation starts, and within twenty minutes it becomes clear that none of it was actually thought through. The ideal customer profile doesn't match the go-to-market. The financial model doesn't work. The documents exist, but the thinking doesn't. That's the productivity flex: high output, low quality of thinking. Barry's waiting for LinkedIn to add a button that lets you flag it. So am I.
AI is exposing leadership systems, not replacing leaders.
One of the most useful reframes in this episode: AI isn't replacing leaders — it's revealing whether they had a real leadership system to begin with. The people who benefit most from these tools are the ones who already had rigorous ways of challenging their own thinking, looking for blind spots, pressure-testing assumptions. The ones who ask "summarize this document" and take the output as gospel are quietly giving up all of their decision-making agency, one prompt at a time. The gap between those two groups is going to widen fast.
Thinking is an activity. Schedule it.
Barry made a confession in this episode that I think a lot of leaders will recognize: he got so excited about AI productivity gains that he hit what he called "productivity fatigue." He was moving fast, generating output, executing — and realized he was missing the whole point. He's paid to make decisions, and some decisions need slow thinking. He now deliberately blocks time to walk in the park, go for a run, sit quietly before high-stakes meetings. Not as a wellness practice — as a decision infrastructure practice. The best decisions don't happen when you're running from meeting to meeting with a head full of AI docs. They happen in the calm.
The job is not the tasks.
A job is a container for a bunch of tasks. AI is automating the low-level tasks. If you've tied your entire professional identity to tasks that are now automatable, that's a real problem — not because you'll be replaced, but because you're measuring your excellence in the wrong thing. The project manager who built a beautiful task-tracking system should now be focused on solving project problems, not tracking tasks. The analyst who created the best spreadsheets should be interrogating the data and making decisions, not building the spreadsheet. The elevation of the work is the point.
Standout Quotes
"The last place you start is the tools. The first place you start is with yourself — how you do your best work. And then what are the tasks that you create value from? That's how you start figuring out what tools you might need."
The inverse of what almost every AI conversation leads with. Paste this somewhere visible.
"AI isn't replacing leaders. It's just exposing their leadership systems."
Eight words that should make every executive who's been dodging the AI conversation sit up a little straighter. It's not about whether you're using the tools. It's about whether you had a real system before the tools showed up.
"Keep telling authentic stories. In the age of AI, staying as authentic as possible is actually the ultimate flex — just being yourself while everyone else is being a machine."
Barry literally has this written on his desk. I'm considering laminating it.
Steve's Take
I've had a lot of conversations about AI at this point, and most of them follow a predictable arc: here's the fear, here's the tool, here's the outcome, here's the framework. Barry's conversation doesn't follow that arc. What he's really talking about is the relationship between self-knowledge and technology — and the idea that if you don't know how you do your best work, no tool on earth is going to help you do it better. He told a story about being dyslexic and spending years trying to write the way he thought writers were supposed to write — sitting at a desk, staring at a page. Nothing came. The unlock was realizing that talking was his natural mode. Once he knew that, the "tool" was obvious: a journalist who could interview him and turn the conversation into a chapter. He was producing 10,000 words in forty minutes. The tool didn't change the output. The self-knowledge did.
That's the thing I keep coming back to. We're so focused on what AI can do that we're not spending nearly enough time asking what we do — what makes us irreplaceable, what our highest-leverage moments actually are, where our judgment is genuinely valuable. Barry's answer to the AI conversation is fundamentally a human conversation. And that's exactly why it landed.
Listen to the Full Episode
Everything above is the skim. The full conversation goes deeper on decision velocity, decision advantage, reversal rate as real AI metrics, and why the executives who are furthest behind are often too embarrassed to admit it — which means they're the least likely to ask for help. Don't miss it.
Listen on Apple, Spotify, or wherever you get your podcasts.
The Question
Barry said the biggest tell on whether someone is actually using AI well is whether they're talking about output or outcome. Activity versus impact.
So here's mine: What's one outcome — not output — that you can point to in the last 90 days and say AI actually made that decision or result better?
Drop it in the comments. I'm not asking what you generated. I'm asking what changed.