The room I almost did not get to stand in

In 2023 I stood at a podium and gave the keynote at the Hide In Plain Sight gala. I was also the second student the program had ever supported. The talk was the same story told from the other end of it, which is a strange thing to do in front of a room, and the only reason I said yes.

What the program is

Hide In Plain Sight supports homeless and at-risk students. The name is the argument: the students it exists for are not visibly in crisis. They are in your classroom, turning work in, and nobody in the building knows where they sleep. That is what makes the need easy to miss and easy to underfund.

I was the second student it ever took on, and the first to finish a bachelor’s and then a master’s through it. Both degrees are on the experience page with the GPAs beside them, and neither number is the part that matters. What matters is that the support was targeted rather than general, and it went to someone who could compound it.

The board years

I served on the board from 2020 to 2023, doing there what I do professionally: sharpening the brand, improving the digital fundraising, and getting the student-impact stories told in a way that actually moved donors. It is the only board seat I have held, and the work was the same work. A nonprofit that cannot explain itself raises less money than one that can, and the students do not care whose fault that is.

Why the talk changed how I work

A keynote is the purest version of the job I now do every week. Take something complicated and personal, and make a room understand it in twenty minutes, once, with no second attempt and no chance to hand them a document afterwards.

It is also where I worked out the thing I believe hardest about building software for people: I would rather build someone a tool than do the work for them. Doing the work for someone is help that ends when you leave. A tool they can run is help that compounds, which is the same shape as what the program did for me. Every system on this site is designed to be handed over and left, and the test is always month three, once the novelty is gone and nobody is asking me anything.

The unglamorous sequel

The speaking I do most now has no stage at all. It is a weekly company-wide AI standup at a 16-lab manufacturer: what shipped, what changed, what to try, who got stuck. It is the least impressive-sounding thing in my calendar and one of the most important, because rollouts do not die of bad technology. They die of silence. The standup is where questions surface while they are still small, where wins get seen by people who were not in the room, and where the 20+ people I have onboarded stay users instead of quietly lapsing.

If you are running an AI rollout and there is no recurring room where people talk about it, you do not have a rollout. You have an announcement.