how to figure out what you want
_**The summer cohort of How to Be a Supermanager with AI starts in one week.**Over 300 leaders have taken it to sharpen how they manage with AI, and at 4.8 stars it’s one of the highest-rated courses on Maven. Enroll today._
I graduated from college with a literature degree and an understanding of the job landscape that was largely based on the career tracks you read about in Richard Scarry books.
Everyone in my family worked in architecture. My older brother showed early talent for drafting; he landed in a prestigious firm right out of school and has been there ever since, an impressive 15-year run with one employer. This talent did not trickle down; when I played The Sims, designing the house always felt like unnecessary tedium standing between me and my ultimate goal: creating a series of bizarre situations for my Sims to navigate. So with architecture off the table, I set out to develop a concept of “jobs.”
My research consisted of exactly two coffee chats.
The first was with a family friend who worked in publishing, a line of work that seemed fitting for a young English major. He sat down and looked over my resume.
“It doesn’t look like you have done any internships,” he observed, correctly. “Because you were…teaching scuba diving in the Caribbean?”
“Yes,” I chirped. “I learned many transferrable leadership skills. Did you see I was also a bartender? I am good under pressure.”
He gazed into the distance, as if looking into the future and finding it grim.
“Listen, Hilary. This industry is tough, and getting tougher. There are maybe ten good jobs available to new graduates, they are very competitive, and they do not pay well. You will not make any money for the first decade of your career, and you will not have job stability. Perhaps you should reconsider.”
So reconsider I did.
The second coffee chat was with a senior executive at National Geographic, a meeting generously arranged by my brother’s then-girlfriend, who used to babysit the executive’s kids. Despite recognizing my good fortune at the connection, I did not rise to the occasion. He asked me why I was interested in working in TV, and at that moment it crossed my mind that, come to think of it, I was not particularly interested in working in TV at all. The conversation stalled out from there.
I graduated on the heels of a recession, in 2010, so I was not alone in struggling to find a job. Maybe, I decided, I would be better off traveling around the world for a year, and then giving employment another go when the economy picked up. Then, by a stroke of luck, my dad told me that someone at his office had received a mailer for an organization called Ocean Conservancy, and he suggested I email them about an internship. How clever! If my burgeoning scuba career did not qualify me for a glamorous role in book publishing, here’s something it _did_ qualify me for. I sent the webmaster a cold email, and soon enough, I was the new webmaster. I made $40,000 a year, and as long as I fired off five tweets about the ocean every day, nobody gave me much trouble.
When I got engaged, I set out to find a wedding dress. Prior to my engagement, I had not thought much about what I would wear to my own wedding. I browsed Pinterest a bit, located a variety of local bridal ateliers, grabbed my best friend, and tried on a couple dozen dresses over the course of a weekend. I ended up deciding between two wildly different dresses, and ultimately went with the very first one I tried on.
I loved my wedding dress, and I loved my first office job, but when I made those choices, it felt a little like spinning a globe and picking my destination with my eyes closed. This is destabilizing for me, as someone who likes to have a strong point of view, and enjoys the work that goes into developing one. So over time, I started studying how people develop a strong point of view about what they consider good.
What I have found is that their approach is remarkably consistent, even across domains.
I am often asked for advice on getting out of a rut. People come to me and say some variant of: _I feel like it is time for a change. What should I do next?_
And I say: _Well, what do you want?_
Rarely does anyone have an answer to this question. It’s not for lack of trying. _I don’t know,_ they say, _but I’ve been trying to figure that out. I’ve journaled, I’ve talked to people, I’ve meditated, I’ve thought a lot about it._
These efforts have usually driven them in circles, or worse, into the warm glow of some Tik-Toker. Either way, they lose confidence in their own judgment, because none of these activities actually help you develop a clear point of view.
The people who have good judgment and a clear sense of what they care about all share one thing:
**They have engaged with at least 10x — often 100x or 1000x — more data points in their domain than the average person. They are constantly pattern-matching across all that data to work out what separates good from bad. And they evaluate anything new against those patterns, almost automatically.**
This is quite simple, but it takes effort. Creatives often describe it as “learning to notice.”
When I started looking at wedding dresses, I was learning about and evaluating, all at once: silhouette, necklines, proportion, fabric, drape, texture, color, embellishment, bustles, comfort, how it photographs, cost, trendiness, and so much more.
You cannot build any real mental model across this many criteria from twenty examples.
Compare this to someone who has dreamed about her wedding dress her whole life. As a teenager she pored over pictures in magazines, taking in thousands of dresses — and if she has a good eye, she wasn’t just looking, she was engaging critically: what am I drawn to? why? — noticing the patterns as they emerged.
Eventually, this is what I learned to do any time I lacked the ability to discern what I want. I have, for example, developed a strong point of view around how to build a career that is aligned with your values. When I talk to people about job options, they’re often amazed by how quickly I can find dozens of listings that might suit them. But I am kind of a freak about reading job listings. I am constantly looking at them, even when I am not looking for a job, because they are by far the best way to understand the evolving economic value of various skills, something I want to have good judgment about.
Sometimes your life arranges itself so that you’re in repeated contact with data and have no choice but to engage with it — and the point of view develops on its own. At work, for example, you might read a ton of grant proposals, and over time you develop a clear sense of what makes one good. But if you want to develop a point of view on something where you are not actively engaging with a large data set in this way, you have to proactively build one.
**I call this the Rule of 100.** Any time you feel like you don’t know what you want — whether you don’t know what job you want, or what haircut you want, or what wedding dress you want — resist the urge to jump into “shopping” mode before you have completed “learning” mode. Learning mode starts with collecting a dataset of 100 examples of that thing.
If you are stuck on what job to get next, go find 100 job descriptions that sound even a little bit interesting to you. Ignore any and all constraints — the location, the seniority, whether you’re qualified — and focus only on the feeling of “I like something about this.” Grab some from industries totally different from your own. (If you are doing this for, say, a wedding dress, same rules: ignore the price, the practicality, whether it’s even available, whether it’ll look good on you.)
Collect the job descriptions, and **only after you have 100, should you sit down to actually read through them.** Then sort them into buckets — what attributes are you most drawn to? — and stack-rank your top 20 across those attributes. Stack ranking is by far the best way to clarify what you care about; as a result it is very hard to do!
The point is not to end up with a list of target companies, or any usable artifact. You can drag your entire moodboard or kanban board into the trash when you are done. In fact, you _should_ drag it into the trash, because the purpose of this exercise is entirely about the clarity it creates in your brain, and 0% about the resulting product you end up with at the end. If at any point you start thinking, _hmm, do I know anyone who could refer me to this job?_ you have entered “shopping” mode and you need to get yourself back into “learning” mode.
Resist, also, the urge to use AI here (at least, for anything other than building yourself a simple workspace to sort your 100 data points into different buckets; I’m including a prompt to build such a tool after the paywall.) Do not allow the AI to research jobs for you. Do not allow the AI to find patterns to bucket the jobs into, or ask it to look at the list and share observations with you. Again, the whole point of this is training your brain. The AI can help later, once you have clarity. Letting AI collect and sort for you is like reading a summary of a study instead of running it yourself. How many summaries have you really, truly internalized after reading them?
Once I learned this technique, I started noticing all the times I’d gone from confusion to clarity the same way — and realized I’d been using this approach. I once had to come up with messaging for how Dropbox helps product teams, sales teams, marketing teams, finance teams, etc. After about a dozen conversations, I felt completely lost in the sauce. It wasn’t until I pushed through, ultimately having about 100 conversations, that clarity emerged, and I ended up doing some of the best work I’ve done.
I think about my early career years when I hear about the malaise taking over today’s new grads. In some ways I was in a similar position — unable to shake the feeling that good jobs no longer existed, or at least weren’t available to me — but in other ways my post-college years would be unrecognizable to them.
I was operating with essentially no data (_Two_ coffee chats! I have to laugh). Today’s new grads have the opposite problem: too much data. Anyone can find and apply to any job posting, look up people who work at any company on LinkedIn or X, and see what anyone they’ve ever met is doing for work — and what that lifestyle affords them — on Instagram.
Cruelly, the downsides of all this data reach everyone automatically, while capturing the upsides takes a deliberate, thoughtful approach.
You can see what everyone around you is doing, you can find virtually all job listings that exist (and so can your 10,000 closest friends who, by the way, already applied), and you have way too much access to the destructive internal thoughts of the people attempting to emotionally regulate by posting on X about the permanent underclass. This can all make you insane — especially, as I wrote in monkey grapes, the specific insanity that comes with seeing your neighbors get grapes while you get cucumbers — but it does not have to be this way.
The key is not to evaluate some random person’s job, or life, against your own, but to use what you know about them to inform your mental model of the many ways to make a living. Read job listings, not because you need to apply to them all, but because evaluating them in the aggregate can bring you clarity. And yes, pay attention to what people say online — but remember they’re making choices consistent with their own internal worlds, not with any objective truth.
There is a sort of data paranoia in the air recently. Do you like a dress because it expresses something about you or because you saw an influencer wearing it? Do you like the music you listen to because it speaks to you, or is it just what the algorithm served up? Is that beautifully decorated home even real, or is it AI-generated? As you practice the art of noticing, it can feel disorienting not to know whether you're swimming on your own or being dragged out to sea by the undertow of digital marketing.
“Too much data” is a problem we all share, but any fix that tries to navigate the data for you — an algorithm, an AI — only warps your judgment further, putting one more layer between you and the clarity that comes from synthesizing the information yourself.
The only way through is to touch the data directly.
xoxo,
hils
_**Behind the paywall:** The hardest part of the Rule of 100 is keeping all that information organized, especially because the best way to organize it depends on the nature of the data._
_Paste the below prompt into Codex or Claude Code to create a workspace that will allow you to drop in 100 data points, zoom in and out, move things into buckets, and stack rank your options. Crucially, it will not help you with the analysis! But it will make your life a bit easier as you do the analysis._