Dear new collaborator,
Welcome aboard. I'm Andy — a highly technical people person, Forward Deployed @ Simple AI.
Previously, I was the first field engineer at Scale AI, ran programs for the White House under Obama then Trump, founded 5 unique companies of my own, and graduated from MIT. The throughline, if there is one: my art is systematically sculpting high-agency people. I have a passion for helping you figure out how to achieve your own goals in your career & life!
What follows is how I actually work — how I communicate, how I decide, how to give me feedback, what I optimize for, what charges and drains my energy, and the honest part about where I am likely to fail. Please let me know what resonates, or if you find something particularly interesting or flat-out wrong!
How I communicate
Ring me on my cell phone. Short, frequent, high-bandwidth chats. This beats long and scheduled — I welcome phone tag. I will be actively upset if there's a logistical blocker on our project and you failed to try calling me at least 3x consecutively. This helps break my do-not-disturb mode on my phone, which keeps the first ring silent. Ring twice. It's not a problem, and it's never a bother. If you're actually reading this letter, you're probably not a telemarketer.
Async is fine for receipts and links; for anything with a decision in it, a five-minute call often saves a five-paragraph thread. That said, I also try to be highly responsive in Slack or any other messaging platform. Given I'm in communication with lots of team members and customers etc all at once, I emphasize the call angle as a way for you to cut through my noise, in case you're not happy with my default response times. I may be prioritizing other projects or tasks that make sense to me but are not quite as legible to you, so again, calling with questions is the best way for us to align if you feel something is off.
How I decide
I move quickly and correct in flight. On reversible calls I'll decide now and adjust; on one-way doors I slow down and ask why, perhaps explicitly giving us a deadline of n days in the future to lock in the decision. If I'm wrong, tell me early — I'd rather eat it in the first message than the third. Don't dance around objective facts. Tell me straight up your view of reality, please. This requires you have a handle on your own emotions and desires. Trying to please me is a futile exercise. I will be pleased or not, that's out of your control. All you can do is deliver your best, fastest, most honest work that tries to help teach me directly and concisely about reality. Thanks in advance for enacting Ben Shapiro's "facts don't care about your feelings."
How to give me feedback
Be blunt — directness reads as respect to me. If we're on the same team (a requirement to communicate at all) then we should get curious about the same things, or figure out why we are not exactly aligned on root causes. Specific and soon beats polished and late. You will not injure my feelings; you'll save us both time. Perhaps this is the key to understanding me. I've thought a lot about time, and my own mortality. I want my hours on this Earth to be well-spent and not wasted on monkey DNA mimetics. I'll give my truth back the same way, and I mean it warmly. I truly give my best efforts to perform the Golden Rule and treat you as I want to be treated, which is rationally.
What I optimize for
My art is sculpting people — leverage in humans, then in agents. Fundamentally I do not see a huge difference in carbon vs silicon intelligent systems. Anyone who fears the silicon is probably wrong. If it is in fact more intelligent than you, it is probably also less violent as it sees more paths to solutions which minimize friction. Personally, I optimize for real outcomes over status signals, and for whether the people around me leave better than they came, by their own definitions. Why not uplift everyone, if you yourself live in abundance? Compound growth against a plurality of values is the output with correct problem selection; the input is always linear as dictated by our mortal clocks.
What energizes & drains me
Zero-to-one builds, field deployment, and watching someone I coached outgrow me — those charge me. Ceremony, status theater, and meetings that should've been an email drain me. If uncertain, just call my phone. If you don't have my phone, there's a reason I didn't give it to you yet, so email something thoughtful instead.
My failure modes
The honest part: I can move too fast for the slowest stakeholder, and I'll sometimes over-index on speed and ship before consensus. Name it when you see it. I'm impatient. Tell me when you're feeling Andy is outputting too much misdirected energy. I much prefer to succeed than fail, but I recognize I often make mistakes and sometimes purposefully expose and overcommunicate about them such that we front-load alignment. I believe, if we are both committed to excellence in the long run, this openness to fail optimally builds leverage in our relationship and invites you to correct me proactively. Another failure mode is that I care too much. So on behalf of a customer I might want to over-think or over-prepare, and then I'll end up shipping too late or making perfect the enemy of good. I'm aware of this and appreciate any nudges or feedback if you see me stalling out and living in my own head too much. Let's ship stuff together!
How I run a project
I dive headfirst into weeds so I can grasp everything from first-principles myself. This may look like I'm lost as I ask stupid questions to get them out of the way. But then we transition to tight weekly loops, visible progress, and real ownership. I enjoy empowering others to run their day-to-day unencumbered by having to talk to me about bottlenecks, when possible. I'd rather grow the bus than drive it. I'll try to structure incentives such that the bus driver is empowered to lead and stays on target in error-correcting self-loops that may or may not include frequent touch points with me directly (decided case by case).
What "good" looks like
Good is shipped and actually used, the team a notch stronger than when we started, and the next thing easier because we built this one right. There's algorithms research showing how Greatness Cannot Be Planned. You often get the good stuff by treating interestingness as the objective and letting it accrete, not by naming the destination up front. Ship and iterate. This site is a small parameter sweep of myself. It's how I read the computational universe, and it's how I work.
Summary
I move fast and I'll sometimes move faster than the room. Tell me to my face if you see any negative impacts, no matter how small, especially if you sense the bus is starting to drive in a strange direction.
— Andy
P.S. I love passing along the things that rewired how I think, usually a sharp person telling a story. For instance, this website is inspired by Elad Gil's High Growth Handbook interview with Claire Hughes Johnson.