My Memorial to Suchir Balaji

I knew Suchir for 8 years. He was one of the sharpest people I’d ever met, and a rare example of a true first principles, independent thinker. From the start, he was entirely self taught. When he showed up to USACO camp, I was surprised I hadn’t seen him before in the bay area – it turns out he’d taught himself everything he knew, to an internationally competitive level. He was also down to partake in fun and mischief, so we immediately became friends.

When interning at Scale, I learned that, again, he’d blitzed learning machine learning online, again, to an internationally competitive level, single handedly winning a competitive $100K Kaggle competition. He believed in the potential of correctly built AGI even back then in 2019, when AGI wasn’t anywhere near as mainstream as it is now. We had a lot of fun that summer – pouring over internal company lore, pulling a prank on the CEO, and making inside jokes that we’d crack for years to come – all while he was clearly doing the most advanced project of any intern. He was so humble, yet extremely sharp.

During covid, we chatted a lot more – we’d support each other’s internship searches (in my referral to Helia I called him, “the fastest learner I know”), joke with our friend Yang, and beta test our websites (he was kind enough to help me patiently debug lipoker when it was still full of bugs, and I helped test his bomberman game). I remember when he made me an incredibly detailed slide for my 21st birthday, and was touched by his kindness.

We’d video called and met up when we could, and were finally thrilled to both be in SF when I moved here a few months ago. It felt like our relationship was only getting stronger as we were making plans to hang out more often. We last hung out two months before he passed away, playing Minecraft and discovering all the new updates since we last played years ago. But we also talked about his dreams – he had a ton of incredible ideas for new experiments to reach AGI and wanted to start a new research lab, had tons of cool ideas for new ways neurotech and AI could interact, and was obviously thinking a lot about fair use.

In many ways, seeing how incredible he had become purely from self teaching inspired me to work on open sourcing courses at top universities, specifically for helping young and self motivated people like him. He was always down to answer any hard questions I had about AI, and I hoped I could pay it forwards. I feel sad that I’ll never be able to share our progress with him, but I hope that he would love what we’re doing – he was always an incredibly supportive friend.

It’s a rare privilege that I treasure, to have had the chance to know Suchir, and to be inspired by someone who stands up for what they believe in, no matter what the status quo is. It’s a horrible tragedy for such an incredibly rare, talented, independent thinker to not be able to see his visions to completion, and a great loss not just for us, but I think for the entire world. Rest in peace Suchir – I’ll miss you.