I could go on and on about the good things about startups, but I think a fun exercise would be to go over the bad things about startups, or in other words, things I “hate” about startups.
Note: this was a word vomit with voice dictation. It’s interesting to see how unclear my thoughts are when speaking compared to writing.
I’ve noticed in the current age because startups are under infinite pressure due to limited resources and therefore motivation as well, It’s easy to get sucked into doing something that is good in the short term but bad in the long term. Actually, this tends to go with what Paul Graham says about doing things that don’t scale. However, I think that logic can be flawed. It leads to building great customer-oriented businesses in sectors that are quick-moving, fast-paced, and have really short iteration cycles. But what it does not do is encourage true innovation - the kinds that take years or decades to develop. I’m talking about the vision pro and VR world, including Orion that was just released by Meta. It’s kind of a luxury for the big companies to be able to do this. They already have a cash cow that they can sink their teeth into whenever they deem. In the meantime, they can truly explore what the next generation of computing looks like. Not many startups get this luxury because they need to make money soon or investors will get mad at them. I don’t lean completely on this side. For eg., if society, customers or users give you a signal that they don’t want something, I’m not saying go and build it. But I’m saying if you truly have the vision that there’s still something there, and in this best case, if the true vision played out, people would love it, then it seems to be a technical problem, not a distribution problem. However, small companies don’t have the funds to focus on technical problems, and thus focus on the cheaper distribution problem.
And because of this, another thing I wouldn’t tend to like about startups is the general atmosphere around how raising money works. I don’t like how you have to convince a lead investor just so all the other investors will go follow along. I don’t like how you have to hype up the market to start making traction a year before you even want to raise.
A good exercise is to look at some of the most impactful organizations in the world right now and think about how they were started and how different that is from the typical YC startup model. OpenAI was started as a non-profit with a giant sum of cash and some of the best researchers in the world. Ethereum is so immensely powerful that it powers economies in some third world countries. However, this was a completely different type of ecosystem, and the monetary value accrued over time because of the technology. You could say it’s luck, but what I’m trying to get at is that it was hackers on the fringe, not people constantly checking if customers wanted exactly what they were building and pivoting if not. Same thing with Anduril, A defense tech company that only works because the founder is already rich, Palmer Lucky already sold Oculus, and is completely funded and creating new innovations because it’s funded by the Department of Defense and other military organizations, which again have infinite money. The idea of creating a giant tech company out of your garage is still possible (maybe in the SaaS world), but creating a new Apple, a new Facebook, a new Stripe seems harder and harder than ever.
I think this is especially prevalent in the world where software becomes more personalized because people are able to make their own software for themselves. Why pay a company that has great user intuition rather than just being able to edit the software that you use everyday?
I still am in this. I do believe right now the easiest attack vector for many people to make impact in an industry is through software. The software layer is so cheap to work on, there will be these giant cash cows who develop the hardware. But as the hardware problems are solved and other problems become software problems, I think that’s where young people like me can see an easy attack vector.