9 Comments
User's avatar
Travis Wright's avatar

"Find the step where a wrong answer becomes permanent and put a person there" is the most underrated sentence in the agent economy. Everybody is racing to pick a model while the real moat turned out to be the approve button. Also love that Sequoia wrote the seed check with zero deck while the rest of us were still kerning slide 7.

Justin Lassiter's avatar

the friday thing stuck with me the most. like the founder of a 5 billion dollar company still blocks a whole day just to mess around with tools and see whats actually changing, and meanwhile most of us cant even protect an hour without meetings eating it. i started doing something like that a few weeks ago, not a full day, just a morning, and its already changed what i recommend to clients cause im actually touching the stuff instead of just reading about it

Qualuation's avatar

Lesson 1 is close to how I built my valuation tool. The model argues the story, fixed code does every calculation, and I sign off before any value appears.

Your test, find the step where a wrong answer becomes permanent, surprised me. It isn't the arithmetic; code handles that. It's the story: once I sign off, every number downstream inherits it. This week a reader's comment made me rethink a story I'd already approved, and the code had no way to catch that.

White Fang's avatar

The n8n story is a strong example of how solving a specific problem can scale into a massive business. The interesting part isn't just the $5.2B valuation, but the product-led growth and profitability behind it.

Daniel Stein's avatar

Stealing the blocked Friday idea, though honestly I doubt I would defend it for more than two weeks before meetings eat it.

Vinta's avatar

Lesson 2 is the one that stuck with me — Sequoia writing a $1.5M seed cheque in 2020 without ever seeing a deck, and Jan not making his first one until the Series B at $270M. “Whether I reach out to them or they reach out to me changes the conversation” is such a clean way to put it. And the Friday rule is a keeper too — one day a week actually using the tools, with that three-question filter for what earns his time. The lean-by-design cap at around 1,000 people, at that scale, is the part I'd love to see more founders copy.

Roger Chua | The Long Game ♟️'s avatar

Fine till it sends money. Then I want a human 😅

Harry Dang's avatar

Two things arrived late in this story, and they arrived for the same reason.

The first is the pitch deck. Jan wrote his first one for the Series B, at a $270 million valuation — after Sequoia had already backed the seed and Felicis had already led the Series A, both without seeing a slide. The second is the category. Jan says it plainly in Lesson 4: 'I think we are already very well known, but very often people don't put us in the right bucket.' A company worth $5.2 billion, past $100 million in ARR and profitable, and the market still does not have a word for what it is.

Both lags have the same cause: the value came first and the language came second. The deck is the language investors need; the category is the language buyers need. Neither is a prerequisite for the value — both are consequences of it. The deck got written when the number got big enough to need a story: $1.5 million is a bet on a person and needs no slides, but $270 million is a bet on a number and needs a document. The category will get written the same way, when a number forces it — and May 2026 already showed how: SAP bought in at $5.2 billion and put n8n inside Joule Studio, its tool for building agents. The bucket is being chosen, and it is being chosen by the buyer, not by n8n.

That is the lesson underneath Lesson 2 and Lesson 4, and it is the one founders should actually steal. You do not need the word before you have the thing. You need the thing, and the word arrives when someone with money needs to say it. The deck was the first time someone else's money wrote n8n's story. Joule Studio is the second. The open question is whether n8n gets to write the third one itself.

Aaron Laam's avatar

Lesson 1 is the one I'd tape to the wall. The right person for that checkpoint is usually whoever does the step today, since they already know which mistakes get expensive.