On one side is AI swallowing millions of jobs, and on the other is humans being hired to clean up the nonsense AI often generates, finds Satyen K. Bordoloi This was early 2023, a few months after ChatGPT had just made the perfect superintelligence landing in our lives. A producer friend, who wanted a beat
“Seriously people, if you’re a dev learn and get good at code reviewing”
It’s a useful skill, even outside of the context of cleaning up AI code. It seems to be universally the case that it’s easier to write code than it is to read it (which is why I sometimes find myself so baffled at my own code that I end up rewriting it again from scratch (good comments are a gift to future-You)).
I’ve been trying to get into contributing to open-source code recently, and it’s been a useful exercise in learning to understand other people’s code. It’s also been making me a more skilled programmer, because reading lots of code helps me to understand how things are typically done in a way that textbook learning can’t. It’s been especially useful to understand how large code projects are structured, given that most of my experience in writing code that’s used by other people has been in smaller, ad-hoc scientific contexts (e.g. writing a script that adds hydrogens onto the protein structures found in the protein database — by default, they’re not included, because it depends on the pH of the environment the protein exists in)
It’s a useful skill, even outside of the context of cleaning up AI code. It seems to be universally the case that it’s easier to write code than it is to read it (which is why I sometimes find myself so baffled at my own code that I end up rewriting it again from scratch (good comments are a gift to future-You)).
I’ve been trying to get into contributing to open-source code recently, and it’s been a useful exercise in learning to understand other people’s code. It’s also been making me a more skilled programmer, because reading lots of code helps me to understand how things are typically done in a way that textbook learning can’t. It’s been especially useful to understand how large code projects are structured, given that most of my experience in writing code that’s used by other people has been in smaller, ad-hoc scientific contexts (e.g. writing a script that adds hydrogens onto the protein structures found in the protein database — by default, they’re not included, because it depends on the pH of the environment the protein exists in)