Sam Corcos' Advice on Writing: How a CEO Uses Writing to Build a $300M Company

Sam Corcos is the co-founder and CEO of Levels, a health tech company valued at roughly $300 million. Levels is fully remote. It runs on writing. Corcos spends less than five percent of his time on strategy, and most of that is writing long-form memos. He believes this is more strategic thinking than most CEOs do. He has a pool of twelve executive assistants shared across a forty-person company. He does think weeks modeled on Bill Gates. He keeps his phone permanently on airplane mode because "technological tools are more powerful than my will."

His writing advice comes from David Perell's How I Write podcast and Levels' published internal culture documents.

Writing Is Thought

"If you can't write out your ideas, you don't have coherent thoughts." Corcos does not treat this as a platitude. He treats it as a diagnostic. When someone at Levels proposes a project and cannot articulate the expected outcome in writing, the project does not move forward. Not because the idea is bad. Because the idea has not been thought through.

Most people, he finds, have ideas in their heads and believe they know the answer. The writing reveals that they do not. Sentences require precision. Paragraphs require logic. "When you actually force yourself to put these ideas on paper, you realize all these gaps that you have in your thinking."

This is not about good prose. It is about clear thinking made visible. A memo riddled with hedging and vague claims is not a writing problem. It is a thinking problem. Fix the thinking and the writing fixes itself.

The Hypothesis

Every project at Levels starts with a written hypothesis: what do you believe will happen, and what is the return on investment? "A hypothesis, even more simply, is what you think is going to happen."

The specificity does not need to be high. Order of magnitude is enough. One project might project $50,000 a month in business value for very little effort. Another might project $8 in ROI. "On paper they look the same, but this one is basically completely worthless, and you wouldn't know that unless you did the math."

The hypothesis serves a second purpose: it prevents rationalization. Without a written prediction, your brain retroactively edits your expectations. The feature ships. It does less than expected. "Oh yeah, I mean, we all knew that was going to happen." But if your original prediction is written down, you cannot run from it. "If you go back to what your prediction was and you realize you were two orders of magnitude off, that's actually really bad. How did that happen?"

Writing the hypothesis in advance creates a feedback loop. Without it, you are guessing and then forgetting that you guessed.

Memos Beat Meetings

The median tech worker cannot go more than six minutes without checking a communication tool. At no point in their day do they get more than thirty minutes of uninterrupted time. Corcos finds these statistics haunting. "Could you imagine trying to write something if every 30 minutes you got distracted? You couldn't do it."

Levels' answer: turn deep work up to 11. Most employees have three or four meetings per week. Some have a single meeting day - Monday - and the rest of the week is open. No-meeting Wednesday, the celebrated policy at other companies, is the default state at Levels.

The writing culture replaces what meetings would do. Strategy memos. Hypotheses. Retrospectives. Loom videos instead of synchronous calls. An internal podcast instead of all-hands meetings. "A lack of communication is a lack of performance" - but communication does not require everyone in a room at the same time.

The Async Week during onboarding forces new hires to break the Slack reflex. For an entire week, you communicate only asynchronously. You send Loom videos instead of scheduling calls. You send voice memos instead of typing quick messages. "We try to push people to the point where it's intentionally past the point where it's useful." A three-second Loom to say "looks good" is inefficient. That is the point. It shows you how far down the async path you can go. The boundary is always further than you think.

Frame the Problem Before Writing the Solution

Before diving into a solution, Corcos insists on framing the problem correctly. "A lot of the time people don't get to the right answer because they're not even framing the problem correctly in the first place."

This echoes Peter Thiel: whatever board you think you are looking at, whatever you think the rules of the game are, your frame might be off. A brilliantly written memo that solves the wrong problem is worse than no memo at all. It consumes resources and creates false confidence.

At Levels, this means spending disproportionate time on problem statements before writing the proposed solution. The hypothesis structure forces this naturally. If you cannot articulate what you think will happen and why it matters, you have not defined the problem yet.

When to Write a Memo and When Not To

Corcos borrows Bezos's one-way door / two-way door framework. A one-way door decision is a tattoo - permanent. A two-way door decision is a hat - you can take it off. Write a memo for one-way doors. Ship without one for two-way doors.

He adds a resource dimension. "If a thing only takes a week or a couple days or even a couple hours, you don't need to write a long memo." But even for small decisions, he recommends writing a one-sentence hypothesis. "I think in two hours I can do this and it's going to have this outcome." That is all. Just write it down. Then do it.

The value of the one-sentence version is not the memo itself. It is the habit. You train your ability to predict outcomes. Over time, your hypotheses get better. You learn where your judgment is miscalibrated. You cannot develop that feedback loop without writing things down.

Think Weeks

Once a quarter, Corcos blocks off a full week to think and write. No email. No team communications. Writing in Notion. Walking with a physical notebook.

The first two or three days produce only tactical thinking. "When you're in the flow of day-to-day problems, it's really hard to see the whole forest." The narrow stuff has to empty out before the strategic thinking can begin. By day four or five, the frame shifts. "Wait, why are we doing any of this at all? Winning is going over there. How do we get from here to there?"

Most of his published strategy memos cluster around the same dates. They come from think weeks. He starts with something tactical - a memo on the product development process - and by mid-week he is writing about company direction, market shifts, fundamental questions about what Levels should be.

He does think weeks with other people, not alone. "There's an energy you get when you're around other people in the same mental space." He can see them across the room in deep flow. It motivates. When he does it alone, he gets distracted and loses momentum.

Write Good Content

Levels' blog gets millions of visitors. People spend three or four minutes per post - meaning they actually read. The editorial director's motto, borrowed from somewhere: "Add value to the internet."

Before writing a post, the team asks: has somebody already written the definitive piece on this topic? If yes, leave it alone. If not, write it. Make it deeply researched and rigorous. "It turns out people like good content."

Corcos has no patience for SEO tricks. "In our experience, the correct answer is just make the best content that people want to read." This maps to Y Combinator's founding principle: make something people want. "Sometimes these simple things just get completely overlooked."

Writing Builds Trust

"Writing a long-form memo is the proof to me that you have done the work to understand your industry, your field, that you know what you're doing." Corcos calls this proof of work. You can only delegate to someone you trust, and trust comes from seeing that they have thought deeply about their domain.

This scales beyond companies. He gave a talk. Afterward, someone asked if he had a book. He said no. "He almost looked at me like he didn't have the same trust in me." A book - any book - signals that the author has thought more about their subject than the average person. Society feels this intuitively.

At Levels, people who communicate their ideas effectively tend to get promoted. "It usually involves some artifact that convinces me that they know what they're doing. And it never happens in just a free-form conversation."

Key Takeaways

• Writing is thinking made visible. If you cannot write out an idea clearly, you have not thought it through.
• Start every project with a written hypothesis. Even one sentence. It prevents rationalization.
• Memos beat meetings. Async communication gives people the deep work time they need.
• Frame the problem before writing the solution. A memo that solves the wrong problem is worse than no memo.
• Write long memos for one-way doors. Write one-sentence hypotheses for everything else.
• Block a full week per quarter for strategic thinking. The first three days will be tactical. Push through.
• Writing is proof of work. It builds the trust that makes delegation possible.

Corcos runs a $300 million company on memos instead of meetings. The writing does not need to be beautiful. It needs to be clear, specific, and honest about what it does not know. Athens is built for that kind of writing - business prose that thinks on the page, with AI editing that tightens your logic without replacing your voice.

This post draws from Corcos' appearance on How I Write and Levels' published culture documents. For another CEO who uses writing as a business tool, see Jason Fried's writing advice.