The right amount of bullshit
I was talking recently with a friend who is a very good programmer, about some of the new developments in AI, like chatGPT and other LLMs. We were brainstorming things that you could do with them, and it became really clear to me that he had a much more careful approach than I did - you might say I was “bullshitting” a bit with some of the ideas. The general direction of them was good, but some details were missing. We disconnected a bit around that uncertainty.
Neither approach was right, or wrong. My friend was right to insist on engineering robustness - absolutely necessary if you’re going to build anything. But I think I was also right - it’s hard to build anything meaningful and new if you don’t let your imagination run at least a bit ahead of current capabilities.
So where is the right balance? There’s been a lot of news lately about companies, and founders, who got too far out on the “bullshit” curve - Theranos and NXT for sure. But there are also plenty of examples of things that have at least had a whiff of bullshit to them from time to time, that have worked out (like SpaceX).
I deliberately framed this as ‘bullshit’ because I think it’s how engineers think about this kind of open-ended thinking. Imagination might be a better word, or exploration, but the fundamental idea is the same - if you walk through something one sure step at a time, you’ll never be able to take interesting leaps. This is just another way of saying “ask what if, not why not”, but more specifically: the right amount of bullshit is when you are fairly certain there is a solution to a problem, even if you don’t have all of the details just yet. The wrong amount is when you have no idea if the problem is solvable at all, or even how you’d approach it, and yet you assert otherwise. And, of course, “too little bullshit” is when you refuse to solve a problem until you have so much detail and certainty that it is essentially already solved.
Fractal boundaries are usually interesting places - where you can’t clearly see which side of the boundary you’re on, and it changes quickly as you scale and move around. There’s one here, between certainty and bullshit, that is likely where all of the interesting, rewarding ideas are.