I just got done watching a Brian Chesky interview on Fortune Magazing.
In it, he spend a substantial amount of time detailing what his definition of founder mode is and why he is convinced it's the best way to run a company.
Ultimately, I think Brian's definition of Founder mode and the definition that's been circulating in the zeitgeist are somewhat different. The tone "Founder Mode" has started to pick up since its circulation after Paul Graham's post is that of a totalitarian regime where founders go around dictating to everybody what they should do and those people just follow orders and do what they're told.
Brian, on the other hand, paints a rather collaborative picture of this. His definition is not so much focused on telling people what to do, but rather on being involved in the company's details. Similar to how things were run when he was just a founder and not a CEO. This definition is more about having continuous conversations and collaborations with the people pushing important agendas within the company and not just assigning them a task and having them figure it out, only to go back and scrutinize them because the initiative didn't do well.
One of the examples he paints is of Airbnb's customer service. At one point in the company, he said, let's just hire an expert in customer service and let them figure out how to best do customer service. On the contrary, if that were to be done in "Founder Mode", Brian would hire the customer service director and work with them extensively over a long period until he and the said director are extremely aligned on what customer service should mean at Airbnb. This alignment is absolutely crucial in ensuring that the company goes in the direction that is aligned with the CEO, which in turn is crucial for ensuring everyone at the company rows in the same direction.
Another very brief point Brian brings up in this video that I found very interesting is the idea of political capital within a company. Oftentimes, not being involved as the CEO or founder of a company means you are excluding your large sum of political capital from the pool of political capital that exists within your company. Doing this will oftentimes also encourage your direct reports to do the same, and their direct reports, and so forth. When this happens, the political capital landscape of a company becomes quite balanced, and as a result, the organization becomes quite democratic. This might sound like a good thing, but for a company this is a terrible thing. Balanced political capital is great for governments where foundational tenants have been defined, and a lot of rapid change is often not a good thing because it threatens those. The US government, for example, is designed this way to ensure that not too many changes happen all at once and uproot the foundations.
On the other hand, being nimble and moving extremely quickly is an absolute requirement in a company. Without it, you'll just be dying a long, slow death. For that reason, company leaders must leverage every ounce of their political capital to move the company in the direction they believe it needs to go. This, paired with founder mode, leads to all employees moving very fast and being aligned in a singular direction.
However, this will allow companies to move very effectively in both the correct direction and the wrong direction. The same trait that allows you to rise quickly can also enable you to fall very rapidly.
This, however, is just a fundamental tradeoff that needs to be made organizationally. Do you want to grow slowly and die slowly or grow rapidly and potentially die rapidly? A lot of governments opt for the former simply because many individuals' livelihoods are dependent on them and a lot of startups opt for the latter because you'd rather know quickly if your idea is worth pursuing or abandoning. Large companies are somewhere in the middle, where sometimes they lean into founder mode and sometimes they lean into conservative mode. And occasionally, you'll have a founder like Brian Chesky move their entire large organization from one model to the next.