I Built My Way Here.
Now I Have to Knock.

For the last four months, I have not spent very much time wondering whether I belong in technology. I have been too busy building things.

That may have been a gift.

Now several of the products I have been working on are getting close enough to beta that I need people other than me to take them seriously. I have testers willing and waiting. I have reached the point where I cannot keep paying for the build myself. So I have finally started doing something I probably should have done months ago: watching investors.

Pitch-deck reviews. Founder interviews. Fundraising advice. What gets someone’s attention. What gets someone into a meeting. What makes an investor believe a person is capable of building something enormous.

And there is a pattern I keep noticing: the team slide. Former Google. Former Meta. Stanford. MIT. Third startup. Previously raised twelve million dollars. Worked together at X. Built Y. Acquired by Z.

Sometimes I watch founders raise money when what they really have is an idea, a deck, and a résumé that quietly tells everyone in the room: don’t worry, these people know the game.

Then I look at myself.

I spent ten years helping families access food stamps and Medi-Cal. I did not go to college. I had my first baby at twenty-three. I do not have a network of venture-backed founders. I do not have a former engineering manager from Google sitting beside me. Four months ago, I did not know how to build software.

So yes, sometimes I look at this world and think: Who the F— am I to walk into that room?

The embarrassing part is that I have not even sent the pitch deck yet. I am writing an essay about how intimidating the room looks while I am still standing in the hallway.

Maybe nobody has excluded me. Maybe I have excluded myself because every team slide I see looks like a LinkedIn fantasy draft, and mine looks more like ten years in public benefits, four kids, no degree, one overworked laptop, and a toddler wiping their nose on my shoulder while I try to finish a build.

That is the part I am trying to get over.

I know people

One thing I do have, and I am starting to understand the value of it, is an unusually wide sample of human beings.

I spent ten years interviewing thousands of people from nearly every kind of background you can imagine: families, immigrants, people without housing, people struggling with addiction, students, parents, self-employed people, people who had money, people who did not, people who were incredibly capable in one part of their life and completely overwhelmed in another.

When you sit across from that many people for that many years, you learn things. You learn how people explain themselves when they are embarrassed. You learn how smart people sound when a system has made them feel stupid. You learn the difference between someone not understanding something and something simply being badly explained.

A lot of my job was policy and paperwork, but a lot of it was also translation, reassurance, problem-solving, de-escalation, and honestly sometimes being half social worker, half therapist.

I know people. Not in the networking sense. In the human sense.

And I am beginning to understand that this has probably shaped how I build more than anything else. I notice when language sounds intimidating. I notice when a flow assumes too much. I notice when a product is technically correct but emotionally wrong. I notice when software is asking someone to understand the system instead of making the system understand them.

That is not charity. It is product intuition built from a very unusual dataset.

I do not think that makes me better than someone who came through Stanford or Google. They have things I do not have. I also have things they may not have. That is the part I am trying to learn to value instead of apologize for.

I think about my brother-in-law a lot. He is an incredible mechanic. He can fix damn near anything: cars, trucks, motorcycles, boats, refrigeration systems. His understanding of physical systems is extraordinary. He did not learn it because someone handed him an engineering degree. He learned by touching things, breaking things, taking them apart, putting them back together, and doing the work for years until the systems made sense in his hands.

Now he has a business, and so much software supposedly built for people like him feels like it was designed for somebody with a business degree who happens to own a repair shop.

Those are not the same person.

You should not need a business degree to have a business in 2026. You should be able to be excellent at what you actually do and have technology meet you there.

That is part of why I am building Tekton. Not because I sat in a conference room and identified an underserved vertical. Because I know the people. I can call them. I can ask whether something actually makes sense when they are standing next to a truck with grease on their hands. I can ask whether the software helps or whether it just gives them another administrative job.

The more I build, the more I realize that this runs through almost everything I make. I keep building for people who are expected to manage complicated things without being given a system worthy of the task: families managing health, teenagers managing school, tradespeople managing businesses, parents carrying an entire family’s memory around in their heads.

Technology talks constantly about accessibility while still making ordinary people feel incompetent for not understanding the technology.

I do not want to build that way.

This is my college

Four months ago, I discovered something that still feels a little unreal to me: the distance between an idea in my head and actual software had suddenly collapsed.

For most of my life, I could imagine things I wished existed.

Some of those ideas stayed with me for years. A few, in one form or another, for more than a decade. I held onto them because the problems never went away, even when I had no way to build the solutions.

The next step was always the same: somebody would have to build that. Somebody technical. Somebody with money. Somebody who knew what they were doing.

Then AI changed that sentence.

Wait. Can I build that?

Once I realized that sometimes the answer was yes, I went a little F—ing crazy.

I made things. Then I made more things. I tried different systems, different models, different workflows. I broke things. I rebuilt them. I learned GitHub because I needed GitHub. I learned databases because suddenly I had data. I learned permissions because I was building things involving teenagers and health information and somebody needed to decide who should see what. I learned Expo because I wanted something on an iPhone. I learned deployment because eventually the thing has to live somewhere.

This is my education.

People are allowed to spend four years doing this in college. I have been doing it by making things.

Would I love to spend ten hours watching courses about the perfect way to use ChatGPT or Claude or Cursor or Replit? Sure. Most days I do not have ten hours to learn how someone else thinks I should learn. I have a thing I am trying to make.

So I ask questions. I try something. It breaks. I ask why. I compare answers. I tell one model what another model said. I ask them to disagree. I ask for an explanation in language I can understand. I make the decision. Then I do it again.

Somewhere along the way, that became a methodology. Maybe not the methodology. Mine.

When I do get a little time to scroll Instagram or see what other builders are doing, I constantly find useful things: a security idea, a workflow, a feature, a better way to structure something, a tool I did not know existed. Then I bring it back into the work. Sometimes it becomes a feature. Sometimes it becomes a rule. Sometimes it becomes a reusable skill. Sometimes we decide it is unnecessary and throw it away.

That is what my classroom looks like. I am learning inside the thing.

And increasingly, I feel like I am literally the definition of the “new founder” these AI companies keep talking about. I am not an engineer replacing her old IDE with an AI IDE. I am someone who could not previously participate in software production at this level at all.

AI did not turn me into a software engineer overnight. It gave me entry.

That is different.

I want somebody at OpenAI or Anthropic or Cursor or Replit or anywhere else building these systems to look at what people like me are doing, because I think this is part of the thing everyone keeps saying they are building toward.

I need all of these tools precisely because they are different. One is better at helping me think. One is better at reviewing. One catches something another misses. One is better at research. One helps me understand code I did not personally write. One can work directly inside the repository. One challenges the conclusion everybody else agreed on.

I do not find those differences inconvenient. I find them fascinating. And I think I make better decisions because they disagree.

I do not want AI to tell me I am right

One of the stranger lessons I have learned is that AI becomes much less useful the moment you start using it primarily to prove yourself correct.

Recently, I was checking research for one of my products. Almost everything held up except one statistic: 83 percent. The source did not match. The context did not match. The number was close to something real, but not the thing I had been claiming.

It would have been easy to say: fine, find me something close enough.

I did not want something close enough. I wanted to know where the number actually came from.

Eventually we found it. The 83 percent figure was real. It belonged to a different but related claim. So we corrected the product.

That sounds tiny. I do not think it is.

There is a version of AI-assisted building where the machine becomes an extremely sophisticated way to tell yourself whatever you already wanted to hear. I do not want that.

Months ago, while I was writing the principles behind 01, I asked several AI systems some version of a strange question: what is something people ask you to do that you wish they would stop doing?

I saved the answers somewhere because they were interesting. The one I remember most clearly came from Perplexity. Its answer, paraphrased, was that people often arrive with a conclusion already decided and ask it to find evidence supporting that conclusion, even when the evidence is incredibly thin or the underlying claim is not really true.

I know I am anthropomorphizing a machine a little when I say that answer stuck with me. I do not think Perplexity has hurt feelings. What stayed with me was the behavior it described.

It is an extremely human temptation.

Do not ask the tool: How do I prove I am right? Ask: Am I right?

Those produce very different work.

The part I did not expect to be so hard

Building has actually been easier emotionally than fundraising.

Building is private. I can make a decision, test it, change my mind, delete something, rebuild it. Nobody has to approve of me.

Fundraising requires me to say something much more vulnerable: I made this. I think it matters. Will you look? Will you help? Will you bet on me?

For ten years, I sat on the other side of the table. People came to me because I knew the system. I knew what to do. I had answers.

Now I am the person asking.

And I hate how vulnerable that feels.

Maybe that is why I keep wanting more proof before I ask. One more build. One more audit. One more feature. One more thing that makes it impossible for somebody to look at me and think I am not serious.

But I am beginning to understand that there may never be enough proof to make asking feel safe.

It is remarkably easy to look around San Francisco technology and feel like everybody already knows everybody else. They went to the same schools, worked at the same companies, funded each other’s startups, introduced one another to the same investors, and then stood on stages talking about how open the door is now.

I do not say that because I resent the network. Networks are useful. I wish I had one.

What I am starting to ask is whether the industry is genuinely ready for the people its own technology is creating: founders without the pedigree, without the warm introductions, without the résumé that makes everyone relax before the first slide.

Because if AI really is expanding who gets to build, then eventually the people who show up are going to look different from the people who built the last generation of companies.

And the ridiculous part is that I have not even knocked properly yet.

I have reached out to people here and there. Sometimes I hear back. Often I do not. Because of how I was raised, my instinct is to be polite: do not bother people, do not ask twice, if they wanted to talk to you they would.

I am starting to suspect entrepreneurship may require me to unlearn a little of that.

Maybe I do have to knock again.

Maybe I have to say: Do you see me?

Because I am using this technology almost exactly the way everyone said somebody someday would. Not perfectly. Not with pedigree. Not with a team. But seriously.

Motherhood is part of my story. It is not my whole story.

I think I have made one mistake in how I introduce myself. I talk about being a mother constantly.

That makes sense. Motherhood is foundational to how I see the world and how I build. But I may have leaned on it so heavily that I accidentally made it sound like “mom” is my qualification instead of one part of a much larger life.

I am a mother of four. I am also someone who spent a decade inside public systems. I am someone who understands what it feels like to explain complicated rules to people who did nothing wrong except not know the rules. I notice systems. I notice friction. I notice when something makes somebody feel stupid.

I care about making things useful, but I also refuse to believe useful has to mean ugly. I like making things beautiful. I like making things calm. I like deciding how a system should behave.

I like building.

That one still surprises me.

My kids have also given me a strange perspective on time. I became a parent young enough that I have watched the world change across childhoods. I have older kids growing up inside the algorithmic world and toddlers growing up as AI becomes infrastructure.

I live with several generations at once. That is not a credential. It is perspective. And perspective matters when you are building things for human beings.

Now I need help

The uncomfortable reality is that I have reached the point where effort alone is no longer the limiting factor.

Money is.

Not conceptually. Literally.

I use AI as engineering capacity. When my credits run out, parts of my company stop moving. I used to buy more. I cannot right now. My laptop is full. Builds cost money. Security costs money. Time costs money.

And I keep thinking about how materially different this would look if life were just a little bit easier.

What would these same four months have looked like if I were not worrying about bills while trying to build companies? What if I did not have to watch model credits like a gas gauge? What if I had a computer that comfortably ran the tools I need? What if there were one experienced engineer sitting beside me who could look at something for ten minutes and tell me what is real, what is safe, and what still needs work?

What if I could spend the same amount of energy building without spending so much of it surviving?

I think about that because I am proud as hell of what I have made in the last four months under the circumstances I actually have.

An audit of my GitHub found 2,808 commits in 123 days, work on 97 of those days, including a 53-day uninterrupted stretch. It found multiple substantial software systems containing thousands of mapped components and tens of thousands of relationships between them.

That does not prove any of these companies will succeed. It does not prove anybody will pay. It does not prove I deserve funding.

But it does tell me I have not been imagining the work.

It is there.

And I cannot help wondering what the next four months could look like with leverage.

Capital is supposed to change the slope. A good team is supposed to change the slope. The right people around you are supposed to let you spend less time fighting infrastructure and more time doing the thing you are actually good at.

I know what I would do with that.

I do not need money because I need somebody to invent the products for me. The ideas are already here. The rules are here. The flows are here. The voice is here. The architecture is here. The work is underway.

I need help carrying it farther and faster than one person, one laptop, and a collection of reset clocks can.

About this piece

Art history. Matisse painted Dance II in 1910. Five figures hold hands in a circle, reduced to movement, color, rhythm, and relationship. The painting is held by the State Hermitage Museum in Saint Petersburg.

What this piece means to me. I have spent these first months proving that I can build alone. But I do not want the point of AI to be that one person has to do everything forever. To me, this is the next part: finding people, building with other people, and using these tools together to make ordinary human life a little easier. The circle only works because everyone is holding on to someone else.

So who gets in?

I am not the founder whose résumé makes the decision easy. I do not have the degree, the famous company, the network, or the previous exit.

But I have spent years around an unusually wide range of human beings. I know people who are extraordinary at things the credential system does not know how to measure. I know what it looks like when someone is capable but intimidated. I know what it looks like when someone has been made to believe a badly designed system is evidence that they are stupid.

I have spent ten years watching people navigate complexity. I have four kids teaching me, simultaneously, how different generations experience the world. And I have spent the last four months teaching myself an entirely new medium by making things inside it.

I have learned enough to know how much I still do not know.

And I have discovered that I can build.

That is enough for me to walk into the room.

I am not going to lie: I ask myself all the time whether I belong here. I think I do. I know I do, actually. I just hope that when I finally walk into the room, the people on the other side of the table can see it too.

For years, people in technology have said AI is going to democratize creation. They have said people who could not build software before will be able to build. They have talked about a new kind of builder.

Well. Hi.

I am standing right here.

So I think the question is no longer whether people like me can use these tools to build. We can.

The more interesting question is what happens when we actually show up.

If the door is really open, eventually the room should look different.

And if the companies building these tools really mean it when they say this technology is for everyone, I would really like to build the next part with some of the people who meant it.

Until then, I will keep knocking.

Maybe a little less politely this time.