All Posts

A while ago I was talking to a friend from college. One of the more amusing coincidences is that he has an older brother with the same name as mine. And in this weird way - like his brother’s path is a premonition for how my future might go, I ask him every few months how his brother is doing. A few years ago, my friend told me his brother was working at a fintech called “smallcase”. Anyway, cut to a few weeks ago. Toward the end of our conversation, I asked my friend if his

I care about both sides of the story: I am interested in how LLMs are actually creating real economic value and increasing productivity today , and also in where the field is headed beyond model scaling. That’s how I ended up reading the HRM paper, which I found fascinating- so here’s my breakdown and thoughts on it:

In my day job, people ask me all the time why they cannot get deterministic outputs from ChatGPT. My usual answer is that its stochastic, flexible nature is exactly what makes it useful as a general purpose QA system. It replaces older patterns like regex scripts and classifier based bots, which can give you more deterministic outputs, but are much more rigid. That is true, but it is not the whole story. Instead of only trying to sell people on that worldview, I think it is a

When GPT 4 came out, and around that same time the “Sparks of Artificial General Intelligence” write up started circulating, I remember being genuinely hyped. I was reading benchmarks, examples, edge cases, all of it. Then we went through the usual cycle for a couple years. One provider beats another, then open source drops something controversial, then someone ships a new trick, then the leaderboard reshuffles again. But cut to 2025 and it feels different. LLMs are starting

Interleaved thinking - wait for an update on this topic.