Make reversible calls fast, take your time on the rest

Most product decisions can be undone, so make them quickly and change course if you got it wrong. Bezos calls these two-way doors. Save the slow thinking for the few you can't take back. On the hard ones I ask what I'm not doing because I'm doing this.

If it goes wrong, it's your fault.

It might have been an engineering bug or a design miss. As the PM, the blame still lands on you. Don't take it personally. You're taking the fall for your team, the same way you're the one who makes the last call when marketing, engineering and design all want different things.

Product skills are the easy part.

When a candidate asks me what it takes to be a PM, I leave product skills off my list on purpose. Those can be learned. Being OK with discomfort and knowing how to get stuff done are harder to teach.

Be the person whose time others fight for.

Engineers and researchers guard their calendars. If they schedule around you and tag you in everything, you're excelling at the job.

Trade off between superstar and rockstar mode on purpose.

A mentor of mine once told me there's either rockstar mode or superstar mode. Superstar is steep growth: uplevel, go-go-go. Rockstar is doing the job really well on a gradual trajectory, deepening the craft and not gunning for the next title. Kim Scott makes the same split in Radical Candor. Stay all superstar and you'll burn out. Stay all rockstar when you want to climb and you'll get bored.

Do every job nobody owns yet.

My first year as a founding PM covered model bring-ups, reliability war rooms, support, developer docs, partnerships and the website. This is what gave me full visibility, context and made me influential at the company early on.

Your support replies are the roadmap.

In December 2023 I answered developers on X and email one by one. Feedback loops were quick. This helped me get to know the customer better.

Think in systems, not just features.

In AI infrastructure especially, small product decisions compound very quickly. They affect reliability, margins, customer behavior, and whether a platform actually becomes something teams build on long term or just try once and move on.

Know what not to build.

Knowing what not to build, when to pause an initiative, or when to sequence work differently often matters more than shipping another feature.

Context before details.

Make sure people understand the context before going into the details. Maybe the details are not that important. Focus on story and not features.

Some early decisions stick.

Shipping early and iterating fast is incredibly valuable, but some early decisions stick. The choices you make at the start can end up shaping velocity, scalability, and the kinds of problems you face months down the line, for better or worse.

The most impactful work isn't flashy.

Some of the most impactful work I've been part of wasn't flashy at all. It was about making the platform more complete and cohesive so customers could grow without friction.

Try the job before you pick the career.

In college I did consulting and venture internships and came away unexcited. One summer at a small startup did more for me than both, and that's how I ended up in product.

Run toward something, not from something.

A mentor told me this years ago. When work gets hard, try to make it work, look at other options seriously, and take your time.

Stay curious without the snowball.

Try new things, but not every spark should become a project. The ones that do turn into a second full life.

Friends, travel and weekends are the life.

I don't save them as prizes or the end goal, they are the reason why I work so hard and I save time for them.

Treat mistakes as a compass.

As Dan Koe puts it, "Mistakes are nature's compass. If happiness can't exist without sadness, success can't exist without failure." If you haven't failed, you probably haven't really tried.

Start the community you would want as friends.

A few friends and I started Babel, a dinner series for people working in AI in San Francisco. A year in, I wrote that the best part was that "they have become some of my closest friends along the way." Another friend and I started a dinner series for women. Another group of friends and I kicked off a book club. I also practice Yoga at Baptiste and I’ve met some incredible people there. These communities have been the best thing to have happened to me in early adulthood.

People are precious.

Times can be hard. But there's one thing that kept me going: people. Someone I respect deeply once told me that what you do with your life matters less than who you do it with. I get that now.

How loudly I love is my task.

I borrowed this from The Courage to Be Disliked: whether someone loves me back was never my task. The task of how loudly I love is mine. The task of reciprocating and loving me back is theirs. Living freely means some people won't respond the way I want, and that has to be ok.

The only person who had me stuck was myself.

In the past, I’ve been so focused on the injustice of a hard situation that I get stuck in a negative thought pattern. The shift came when I stopped caring about that and started asking how to get unstuck. I realized the only person who had me stuck was myself.

Write letters.

What's left of you is some memories, eventually a photo album. If you have letters you're giving someone something physical that stays. Knowing it might help someone, or mean something to them, is important.

Don't run from hard people.

Working with someone difficult is uncomfortable, and the instinct is to avoid it and run. Hard people show up at every stage of life, and it only gets more likely the more senior you get. Running from every one means missing something beautiful.

Go to the right store for the milk.

There's an old recovery saying: don't go to the hardware store for milk. I don't have to get everything from the same few people. They're like hardware stores, amazing for some things, but they don't always have the item I need. Learning where to go for what I need helped me stop feeling hurt for going to the wrong places for support.

Let it shake out.

When something goes wrong I'm quick to react. The better move is patience: give it some time for things to shake out.

Know the five things that feel like you.

Mine right now: morning walks and reading in the park, sending letters, calls with loved ones, catching up with friends, and having a routine but leaving room for serendipity. I can’t make time for all of them every week, but I make it a goal to do at least a couple to keep me sane.