Why make this blog ?

TLDR: Friends, a good name and Foucault

For the past few months, I've been chatting with friends and colleagues who kept telling me, "You should really start a blog!" Honestly, it didn't really go anywhere until a random Telegram chat, ironically, about someone else's Substack names, led me to "Approximately Correct.". It just clicked. Inspired by PAC (Probably Approximately Correct) learning theory from machine learning, it perfectly captured my vision for this space. So I figured, why not actually use it? And just like that, here we are—Approximately Correct is officially born!

This isn't my first time at the keyboard, though. I've tried writing before, but my ambition always got the better of me, leaving me with nothing truly satisfying. This time, I'm striving for a more fluid approach, embodying closer to the true meaning of an "essay" as an "attempt”, as Montaigne originally used it to describe his work. My goal is simply to share my rough ideas with the wider internet (and hopefully, that won't backfire!) and becomes a useful space for me to think more deeply by articulating ideas and engaging with more of you.

One big thing I really want to dive into, and something I hope this blog can contribute to, is the genealogy of ideas—or, borrowing from Foucault, the "archeology of knowledge." I especially want to fill the gap in education about where ideas come from, particularly in areas like machine learning and math. My goal is to illuminate for a wider, interested audience how science is done, and maybe even actual science along the way.