What if it’s too easy?

Being the first to do something is hard. What if the thing you’re trying to do is impossible? What if it takes too long? What if it’s a waste of time to even try?

Soon as you see someone else build it, the game completely changes. As soon as you know it’s possible to build a nuclear bomb or a large language model AI, the realm of possibilities has expanded. Someone has conquered this new frontier and so can you. It might still take a long time but it’s not impossible or a waste of time. Humanity’s technological progress follows this pattern: breakthroughs are rare but then the clones come fast.

With Pico, I had been resisting supporting AI-generated backends because unlike frontend code which runs in the browser, I was worried about the security implications and the complexity of running AI-generated, user modified code on my servers. But then a couple of weeks ago, OpenAI released code interpreter which does exactly that. And it works really well! That was exactly the example I needed, and I was able to re-build the core of Code Interpreter in about 2 days.

2 takeaways here:

• Do short spikes/exploratory work to really check the assumptions that make you think something is hard. What if it’s too easy? A couple of hours might be all you need to get the answer to that question.

• I have a pretty unhealthy attitude towards competition. It stresses me out to see similar products pop up. I’d like to reframe that: Competition in your space is actually great. Someone else can make the missteps and lay out the execution plan for you to build what you want to build.