I Asked Claude the Question I Could Never Ask My Boss

Katie -- Again, thank you for a great, introspective piece. And I agree overall with Mr. Greene's feedback (below... including his P.S. for a better commentary tool!). I'm only gonna add that the tough honesty with which you wield your self-criticism is both unusual and refreshing -- not everyone can or will share and air their self-view and self-assessments to the degree that you've done here and before. If I was your manager &/or personal friend, I'd certainly councel dialing your level of self-doubts back a few notches... but not to zero, certainly not by too much, as this is one of your personal super-powers -- the ability to introspect with honesty, and to use that as a personal motivator. In other words, we like you just the way you are! Keep up the great work, Katie, and I certainly look forward to your Every contributions in 2026! Cheers, Happy New Year!

This has got me thinking about where I can get this kind of data for my job, which is vastly different. I'm a programming librarian by day, so the best metric I can think of is attendance at programs, but I think I need baseline numbers to compare them to, the way you can to other writers, or to the resulting action people take after reading your articles.

I really love this one! One of the few Every articles I saved to Reader and highlighted. Your conclusion feels like it misses something that it seemed like you were clear on earlier though (and which I personally am very excited about). You ended with:
"But my relationship to the doubt has shifted because now I have the data.
I can look at the numbers instead of believing the narrative that my mind tells me."

I don't think that's what made the difference. You said it yourself several times throughout, here's a key example (emphasis mine):
"At one point, the AI asked me point blank: Why are you asking this? What is it about it that’s so hard to believe? That’s when I said the thing I’d never said to a manager or a therapist or even, really, to myself:
“I don’t know. I guess just pathological self-hatred and an inability to trust that I do good work.”
**I could say it to a chatbot because there was no social cost, and I didn’t have to be worried about burdening someone on the other end with my baggage**. If I believed the data, the next step was to believe myself."

You *had* the data before, you probably even had some data on your performance in past jobs to some degree, but you never *believed it* or allowed it to influence your underlying opinion of yourself and your capabilities. And having more and more data didn't really seem to help you believe it more! What *does* seem to have helped is having "someone"/something repeatedly use the data to demonstrate to you how your self-perception was wrong. The key aspect there is not, IMO, the data, but the use of it, consistently, to reinforce a more realistic and (by fortunate happenstance) positive view of yourself. Not only that but there is clearly an element here of it working because you were interacting with more of a "something" than a *someone*, i.e. you could relate to and utilize the AIs in a way that you could not with a human. This fact in itself led to some transformation, a shift that arguably would not have been possible with traditional human resources (your boss, a good friend, a therapist; none of them are fully equipped to navigate you through this to where you ended up!).

This right here is the actual key and IMO is the stronger end (or a version of it) to this whole article and the case you're making (emphasis mine):
"Kate had already told me that we want to publish more of my column in 2026. But I wasn’t prepared to accept it. Coming from a person, I would have brushed it off as encouragement, the kind of thing a good manager says to keep you motivated. **Coming from a system that had just walked me through the data and played out scenarios for what each choice would mean, it landed differently.** "

This shifts it significantly from a fairly familiar "I just needed the data to show why I was right/wrong" message to a much more interesting and new "the social dynamics of artificial intelligence allowed me to ask a question and get an answer that simply wouldn't have convinced me coming from a person". That's fascinating and powerful because it's a *new* capability, a new paradigm, that we haven't had before. There are potential versions of it, perhaps, ranging from anonymous feedback to certain forms of therapy, certain interpersonal relationships that you really trust, etc. But for this to be a *service* that almost anyone in the developed world can access (setting aside the data, I'm talking about the mechanism of how the shift in *your* mindset worked here), that's an incredible thing.

There are a lot of people wringing their hands about AI being used in place of therapists and whether it will be helpful, or concern about people "befriending" AI, etc. And while I think there are genuinely concerning aspects of it all (not least to do with privacy and confidentiality, training data, etc.), it's stories like this - and some of my own experiences - that I think are at least as worthy of exploring and discussing. This is not a case of you doing something with an AI that a therapist could have done for you (most likely): this is a meaningfully new, different, and potentially more powerful method, if only by dint of its sheer omnipresent availability. My former coach would constantly say he wished he could be there in the very moment I would procrastinate or otherwise struggle with some challenge, and now AI *can* be in a way no human assistant ever could be. So I would love to see more in-depth explorations of this kind of thing!

P.S. I'm sure improving the comment system here is low on the list of Every.to priorities, but I would love if someone on the team could expend some AI cycles to give us some basic formatting here (italic, at least, bold) and quote ability, Being able to highlight on-page and then quote into the reply box would be awesome. Maybe there don't seem to be enough chat interactions to justify it, but that could also be chicken-and-egg. Maybe if you built a much nicer commenting/interaction mechanism you'd get more interaction...

Keep up the good work!

I am always up for reading an article with your byline. You have a distinctive voice, and you are using the tech you write about in novel ways. You are definitely one of the reasons I open my emails from Every. Glad to hear we'll be seeing more articles from you in 2026

I really appreciated this piece. Your openness about self‑doubt, and the way you used AI to turn raw numbers into a story about impact, resonated deeply. The distinctions you drew between conversion, brand building, and reader satisfaction were clarifying—and the takeaway that vulnerability can be a strategic advantage felt both brave and practical. Thanks for sharing the prompts and process; they made the analysis feel replicable. I’m rooting for more Working Overtime in 2026 and grateful for the reminder to build our own case with data.

Loved this—clear, vulnerable, and genuinely useful for self‑review.

I’d add tracking effort/energy alongside outcomes so you can sustain the mix that drives loyalty and trials without burning out.

BTW, tiny ops tip: for quick specs on merch or shoots, this Ring Size Chart is handy.