About

I'm a committed generalist with a professional background in institutional asset allocation and portfolio optimization, focusing mainly on fixed income and private markets.

I like to build tools and solve problems with code. A few years ago I embarked upon a quixotic quest to apply pre-trained foundation models like BERT to Central Bank communications. Then OpenAI revealed the gargantuan scale required to obtain useful results, releasing a general-purpose model that blows fine-tuned BERT out of the water. Since then I've been making peace with LLMs, learning how to use them to create knowledge while holding on to my scientific passion for truth.

I'm not anti-AI, but I am anti-BS. I'm genuinely impressed by what companies like OpenAI, Anthropic, and Google have been able to accomplish by scaling up transformer models, but I don't believe next token prediction will take us all the way to artificial general intelligence.