Founder Mode Is Real So Is the Trap Inside It
The essay that told founders to stop delegating and get back into the weeds diagnosed something true. It also handed out the most convenient permission slip in recent business writing – because it never quite said what, exactly, it was permission for.
By Navoch Mohanayak
A talk that was meant to run half an hour at a Y Combinator event stretched to two, and the room reportedly didn’t want it to stop. A little over a week later, a co-founder of that same accelerator turned the talk into a roughly thousand-word essay, gave the idea in it a name, and within twenty-four hours it had its own newspaper trend pieces, its own merchandise, and its own small civil war on social media. The essay was called “Founder Mode.” The talk was Airbnb’s Brian Chesky, describing why the conventional advice he’d been given about scaling a company – hire strong people, hand them a piece of the org chart, stay out of the details – had, in his own account, nearly wrecked the thing he’d built. The essay’s author, Paul Graham, gave the alternative a catchy label and let the internet do the rest.
It’s worth being honest about how much of the underlying diagnosis is actually correct, because a lot of the backlash it triggered treated it as obviously false rather than partially true and badly generalised. There is a genuine failure mode where a founder, told often enough that hands-on involvement is a maturity problem they need to outgrow, delegates the parts of the business that were never actually generic to begin with – the specific texture of the product, the instinct for which customer complaint is a signal and which is noise – to executives who are competent at running a function but have no way of inheriting judgment that was never written down anywhere. That’s not a strawman. It happens often enough, in enough well-known companies, that an essay naming it was always going to find an audience desperate for the vocabulary.
Which is exactly why it landed as hard as it did.
For years, the entire apparatus around growing companies – business schools, board decks, seasoned operators brought in to “professionalise” things – had quietly implied that a founder still deep in the details three years post-launch was a founder who hadn’t grown up yet. Plenty of founders who stayed hands-on did so with a nagging, unexamined guilt about it, because every piece of conventional wisdom told them this was a personal failure to delegate rather than possibly the correct call for their specific business at its specific stage. An essay that said, in effect, no, actually, some of you were right to stay in the weeds, and the advice telling you otherwise was often just cover for hiring people who look competent on a resume and quietly run things into the ground, was never going to be received calmly. It was going to be received as relief.
Here’s the trap, and it’s a genuinely interesting one, not just a contrarian’s footnote: the essay never actually defines founder mode.
It gestures at examples – a founder attending skip-level meetings, staying obsessively close to product detail, refusing to treat the org chart as a set of sealed boxes – but it declines, almost proudly, to draw a boundary around what counts and what doesn’t. That’s not sloppy writing. It’s arguably the single biggest reason the idea spread as fast as it did, because an undefined concept can absorb absolutely any behaviour a reader already wanted validated. A founder who’s genuinely irreplaceable on three specific calls and delegates everything else can call that founder mode. A founder who simply doesn’t trust anyone else to make any decision, about anything, ever, can call that founder mode too – and until someone draws the line, both get to cite the same viral essay as their operating philosophy. Vagueness that generous isn’t a bug in a piece of writing designed to spread. It’s practically the design spec.
The distinction that actually matters, and that the discourse mostly skipped past, is between founder mode as a targeted exception and founder mode as a permanent identity.
The defensible version is narrow: stay personally involved in the small number of decisions where your judgment is genuinely, currently irreplaceable, and mean it – not as a general management style, but as an honest, continuously updated list. Early on, that list is often long, because almost nobody else yet understands the product or the customer as well as the person who built the first version by hand. The trap version treats founder mode as a permanent personality rather than a shrinking, specific exception – which quietly re-imports the exact bottleneck the company was trying to outgrow in the first place, just with better branding and a viral essay to cite when someone asks why nothing moves without your sign-off. A company where the founder’s irreplaceable list gets smaller and more specific over time is maturing. A company where that list stays the same size, or grows, because “founder mode” became an identity rather than an audit, is just delegation avoidance with a Y Combinator citation attached.
This is also where the essay’s own famous examples quietly work against its popularised version. The account of Chesky’s own practice – obsessive product review, skip-level conversations, an insistence on understanding the work at a level of detail most CEOs skip – is a specific, narrow discipline applied to a specific, narrow set of decisions. It is not the same thing as a founder personally approving every hire, every expense line, and every internal Slack thread three years after the company has grown past the point where any one person could hold all of that in their head without becoming the ceiling on how fast anything happens. The meme flattened a specific practice into a general licence, and the flattening is where the trap actually lives – not in the original diagnosis, which was mostly fair, but in the version of it that made it onto T-shirts.
None of this resolves into “pick a mode and defend it,” which is itself the mistake both camps in this argument keep making.
Founder mode and manager mode, argued as competing philosophies each side should commit to permanently, are both static answers to a question that should never stop being asked: which specific decisions, right now, in this specific business, would go worse without your particular judgment on them – and which ones are you holding onto simply because letting go feels like losing something, even when the business has long since outgrown the need for you to be the one holding it? That list is not fixed. It should be re-audited often, and honestly, and it should generally get shorter and more precise as the people around a founder build their own version of the judgment that used to exist only in one head. Treating founder mode as an identity to defend, rather than an exception to keep narrowing, guarantees you’ll still be citing the same essay for the same behaviour five years after the company, and the argument for it, have both moved on without you noticing.
The essay’s real trick, in the end, wasn’t diagnosing something true about how delegation gets abused. Plenty of people had said versions of that before, with less traffic. Its trick was doing so without ever specifying where the exception ends, which meant it could be quoted with equal sincerity by the founder doing the hard, narrow, constantly-updated work of figuring out where their judgment still matters, and by the founder who simply never wanted to stop being the one everyone has to ask. Only one of those two founders is actually practising what got, briefly, quite a lot of merchandise made about it. The other one just found a citation.