Companion.
A thinking partner that thinks beside you.
Trained on the archive of long-form founder thinking we have been quietly assembling. Not a chatbot. Not a guru. Something you can talk to at four in the morning that will not pretend to know more than it does.
The room is the model.
Most founder tools have been trained on the loudest version of entrepreneurship - the conference talks, the tweet threads, the certainty industry. We have been training Companion on the opposite. The quiet pieces. The endings. The Tuesday paragraphs where the founder admits they do not know yet.
It is a different room. It produces a different conversation.
Four small abilities, held carefully.
Sit with a hard decision
Hand it the messy version. It will ask the questions you have been avoiding before it offers anything.
Find the through-line
Across years of your own writing, your own meetings, your own notes - it remembers what you said in October.
Hold the second opinion
Trained on hundreds of long-form founder pieces. Calmly disagrees when the archive disagrees with you.
Stay quiet
It does not push notifications. It does not have a streak. It is there when you open it.
What we are working on, and what we are deliberately not.
- Now
Private alpha with thirty founders. We are reading their transcripts more than we are shipping features.
- Spring
Long-form 'thinking sessions' that produce a written piece you can publish under your own byline.
- Summer
Memory across years. The companion that remembers what you said in October when you are panicking in March.
- Never
Notifications, streaks, leaderboards, daily tips. We will not be adding these.
Leave a name. We will write when there is something honest to show.
We will not write to you weekly. We will write once, when it matters.
Why a thinking partner, not another tool
The market for founder tools is crowded. Every week a new app launches that promises to help you with your CRM, your project management, your accounting, your hiring pipeline, your social media calendar. These tools are useful. They solve specific, well-defined problems. But they do not solve the problem that most founders actually struggle with: the quality of their thinking in the moments between the question and the decision.
The Companion was built to fill that gap. It is not a productivity tool. It is not a content generator. It is a thinking partner: something you can talk to, argue with, and think alongside when the problem is messy, the stakes are high, and the obvious answer feels wrong.
The founders who use the Companion in alpha have described it as the person in the room who asks the question you were hoping no one would ask. Not because they are being difficult, but because they have been through enough to know that the question you are avoiding is almost always the one that matters most. That is what the Companion does. It does not give you the answer. It helps you find the question that leads to the answer.
Why it works
The Companion works because of what it has been trained on. It has read every piece in the SOE archive: the stories, the frameworks, the dispatches, the long-form essays, the research summaries. It has been trained on the patterns of thinking that appear across those pieces: the way successful founders frame problems, the questions they ask when they are stuck, the mistakes they make and recover from, the reasoning that leads to good decisions.
This training gives the Companion something that generic AI assistants do not have: a model of how founders think, drawn from hundreds of real examples rather than from a generic corpus of internet text. When you describe a problem to the Companion, it does not search for the most common answer on the web. It searches for the pattern in the archive that most closely matches your situation, and it uses that pattern to ask you the questions that a founder who had been through this would ask.
The psychological mechanism is well-established. Research on analogical reasoning, published in the journal Cognitive Science in 2021, found that problem-solving improves significantly when the solver is exposed to structurally similar problems from a different domain, because the structural similarity reveals the underlying pattern that the surface details obscure. The Companion provides this exposure systematically, drawing on a library of founder problems that spans industries, phases, and contexts.
What the alpha is teaching us
The private alpha currently has thirty founders working with the Companion on real problems. They are not testing a product. They are helping build one. Every conversation they have with the Companion is reviewed by the team, not to monitor what they are saying, but to understand how they are using the tool, what kinds of problems they bring to it, and where the tool helps or hinders their thinking.
The alpha has already produced insights that changed how we think about the product. The founders in the alpha use the Companion most often for decisions that have no clear right answer: hiring decisions, pricing decisions, whether to stay or go, whether to raise or bootstrap, whether to pivot or persist. These are the problems where the decision quality matters most, and where generic advice is least helpful because the context is always specific.
One pattern that emerged early: the founders who get the most from the Companion are the ones who bring their messy version of the problem, not the polished version. They write the messy paragraph first, the one with the contradictions and the raw emotion, and let the Companion work with that. The founders who bring the clean version: the one they would present in a board meeting, and get less value because the Companion is working with the problem as it appears, not as it actually is.
This insight has shaped every feature we have built since. The Companion does not ask for a clean problem statement. It invites the messy version. It works with the contradictions. It does not try to resolve them, and it helps you see them clearly enough to make the decision yourself.
The roadmap, honestly
We are not building a product roadmap in the conventional sense. We are building a thinking partner, and the roadmap is determined by what the founders in the alpha tell us they need, not by a quarterly plan drawn up before anyone has used the tool. The alpha will run for as long as it takes to get the tool right, and we will not rush it.
The next phase, which we are calling Spring, is about long-form thinking sessions. The Companion will guide you through a structured thinking process that produces a written piece: a strategy document, a decision memo, a narrative framework. the you can actually use. The piece is yours. The Companion does not own it, claim it, or display it. It just helped you think it through.
The phase after that, which we are calling Summer, is about memory. The Companion will remember what you said last quarter when you are in the middle of a crisis this quarter. It will notice patterns in your thinking: the decisions you have made, the problems you have solved, the questions you keep coming back to, and it will surface those patterns at the moments when they are relevant. Not as notifications. As context. When you open the Companion, it will start from where you left off, not from a blank screen.
There are things we will not build. Notifications, streaks, leaderboards, daily tips, gamification of any kind. The Companion is a quiet tool. It respects your attention. It does not compete with the rest of your life: it sits quietly until you need it, and it lets you leave when you are done.
Who this is for
The Companion is for founders who are thinking about problems that do not have clear answers. The first-time founder who cannot decide whether to raise or bootstrap. The founder at scale who cannot decide whether to keep growing or stabilise. The founder who has been building for years and has lost the thread of why they started. The founder who is facing a decision that will change the direction of the business and who knows that the wrong choice will cost them a year.
It is not for founders who are looking for a quick answer. The Companion does not do quick. It does not do ten-second tips. It does not do tweet-length advice. It works slowly, deliberately, and with the expectation that the thinking is yours. The Companion is the scaffolding. You are the builder.
It is also not for founders who have already solved their thinking problem. If you have a clear decision framework that works for you, a mentor who asks hard questions, and a team that challenges your assumptions, you do not need the Companion. Use those things. They are better than any tool we could build.
The Companion is for the founders who do not have those things yet. The solo founder. The founder in a market where there is no mentor. The founder whose team is new and still learning to challenge assumptions. The founder who has tried the books and the podcasts and the templates and still finds themselves stuck at the same point, making the same mistake, unable to see the pattern that is obvious from the outside.
What it costs
Nothing. The Companion is free during the alpha, and it will be free when it launches publicly. There is no subscription, no premium tier, no locked feature. The reason is simple: the Companion does not generate revenue. It is a tool for thinking, and the value of thinking tools is not in the tool itself but in what the person using the tool builds with the better thinking it produces.
The SOE ecosystem is funded by the services, the tools, and the community spaces that help founders move from thinking to building. The Companion is part of that ecosystem, not a product within it. It is the thinking layer that sits beneath everything else: the library, the workshop, the narrative practice, the community, and its job is to make the rest of the ecosystem more useful.
The founders who use the Companion and then use the Workshop tools, or work with Own Your Narrative, or join the Pro community, are the ones who get the full value of the ecosystem. But the Companion is available on its own terms, without any requirement to engage with the rest of the platform. It is a tool, not a funnel.
The alpha has limited spots. Not because we want to create artificial scarcity, but because we can only support a certain number of founders with the level of attention the alpha requires. Each alpha participant gets a direct line to the team, and we read every conversation. That does not scale linearly. When the alpha closes, we will open a public beta that accepts more participants, and eventually the full product. But we will not rush the timeline for growth. The tool will be ready when it is ready.