Midlife MavericksIt’s 7pm on Sunday. I’m at my desk. Zoom open with Orel, the founder of WriteStack, whose MCP I’d been trying to connect for half the day. Younger than me. Deep in tech, and a fabulous guy too, otherwise he would not be giving up his precious time to help me.
He’d been venting for ten minutes. He’d spent the entire day battling N8N, trying to build a bunch of agents for his product, getting nowhere.
I told him to save his money and his headache because he didn’t need N8N. Almost all of it could be done inside Claude Cowork in a fraction of the time. I showed him two or three things I’d worked out over the last few months.
He had no idea any of it existed.
Quote of the Week
“The trouble with the world is that the stupid are cocksure and the intelligent are full of doubt.”
Bertrand Russell
The Two-Day Course That Taught Me Nothing
The day before that call with Orel, I’d done something else strange.
A few weeks ago I bought a two-day AI event. Two of the sessions on the agenda looked properly useful. I couldn’t make the live dates, so this weekend I sat down to study the recordings. I take notes, make implementation plans, rewatch the bits I want to test. (Yes, I am that person.)
The teacher is a woman who runs her own business. Doing the work. Teaching what she has built.
Two full days of “studying” later, I closed my laptop and realised something uncomfortable.
I had not learned a single new technique.
A few new alternative approaches to look at, yes. Techniques? Nothing. Every single thing she was teaching, I was already doing in my own work.
A year ago I had zero technical skills. Zero. I am 49. I started learning AI properly twelve months ago because nobody else was going to do it for me. And on Sunday at 7pm I was the one explaining the workflow to the tech founder.
The Workshop in the Background
Two weeks before all this, I ran an AI workshop for 67 women in finance. Twenty-seven of them filled in the feedback form when it ended.
This was not the AI-curious crowd. The room had a CIO of a family office, an SVP M&A lawyer at a large consulting firm, a senior portfolio manager at a UK fund, a Head of Sustainability, the GC of a fund, a director at a major bank. Senior women, running real work at real institutions.
The feedback came back, and it took me a while to fully take it all in. Half of them walked in saying they used AI but without confidence. By the end of the 60 minutes, only one person was still unsure. The single most repeated surprise across the whole room? “I had no idea there were this many tools.” Ten of them wrote some version of that line.
The second most repeated surprise: you can train AI to sound like you. The third: AI is sycophantic and will agree with you forever unless you tell it not to. Linda put it cleanly in her feedback: “AI is always agreeing if you don’t tell it to be critical.” (That sycophancy thing deserves its own newsletter. Soon.)
A few of the women had ticked “I use AI regularly” coming in. After the session, they ticked “starting to see how it fits.” On paper that reads like a step backwards. What actually happened was that they realised they had been using AI badly all along, and that moment is worth more than five years of casual dabbling.
The Confidence Gap
The Bank for International Settlements ran the numbers last year. Half of men have used generative AI in the past 12 months. Just over a third of women. That gap is real.
Three quarters of that gap is just self-assessed confidence. Women rating themselves as less capable than they actually are.
We have been told for so long that the technical world belongs to someone else that we walk into rooms full of people who know less than we do, and we shrink. The man on the stage with the brand and the LinkedIn following, the tech bros with the funding, the 26-year-old in the hoodie at the conference. We assume all of them know more. Sometimes they do. A lot of the time, they don’t.
This Isn’t Just About AI
The voice in your head telling you that you don’t know enough to use AI properly is the same voice that talks you out of raising your hand for the promotion. You’d only ticked seven of the ten boxes on the job spec, and you wanted to be sure before you went for it.
The colleague down the corridor ticked five of the ten boxes and raised his hand last Tuesday. He is interviewing this week.
In nature, the bolder animal eats first, not the smarter one. We have spent decades training women to wait until they’re sure, and “sure” never quite arrives. While we wait, somebody less qualified walks up to the bowl and eats.
That voice doesn’t only cost you AI fluency. It costs you the title, money, the room, the recognition, and if you’ll allow me, it also costs us a ton of wrinkles, grey hair and bloody mental space.
OWN IT
This is the part I keep coming back to.
I bought the two-day course because I assumed the teacher would have a body of knowledge I lacked. She did not. The same has been true at almost every other point I’ve assumed someone knew more than me on this topic.
Most of the time, the thing standing between you and competent AI use is a voice in your head that decided, before you ever opened the tool, that you didn’t know enough to use it properly. That voice is wrong more often than it’s right.
This week, pick one thing about AI you have been telling yourself you “don’t really understand.” Open Claude or ChatGPT. Explain that thing out loud, as if you are teaching it to someone who knows nothing. Time yourself. Notice how much actually comes out, and how much of what you thought you didn’t know is already sitting somewhere in your head. Then, while you’re at it, ask Claude or ChatGPT to find your real knowledge gaps and tell you what to do about them.
The problem is that, us women, always assume we don’t yet know enough. But you might just be starting from a much higher floor than you have been giving yourself credit for. And that, is a hill I’ll die on.
Reply and tell me what you tried to explain, and what came out when you did. The thing you thought you did not understand and what you actually did. I read all of these.
See you Wednesday
