Sabitlenmiş Tweet
For all the tech bros hyped up about biology.
Imagine a codebase. The codebase is huge. You don't know who wrote it or when. It is written in a language you don't know, nobody knows it. People all over the world have studied the codebase over decades and figured out what some parts of it do, but not really why and how they work. You are fairly sure the code you see is not complete, it's compressed somehow so that the actual code is way longer than it appears. You suspect there are layers to this, but you are not sure.
The output of the code is different even with the same input. Over time, people have figured out how to get the outputs to cluster around the expected value for some small snippets, but even that falls apart when you run it on a different computer, even if you replicate the environment exactly. And sometimes it runs differently even on the same computer. You don't know why that is.
You are told to develop a new feature that plugs into it. Testing the feature takes ages because the code runs slowly and it's impossible to speed up. No matter what you do, the behaviour of your feature changes unpredictably whenever you add a variable. What's more, the behaviour of the whole codebase changes when you add something, even if your feature is entirely self contained. You struggle and struggle and the best you can do is to get a statistically significant clustering of your output around the expected value.
Your investors give up on you because you are not delivering. It's taking too long and you're not making money. Peter Thiel comes in and calls it a skill issue.
English
