SPECULATIVE DECODING_

Jonathan Larkin

SPECULATIVE DECODING_

Working notes on markets and machines.

POSTS PUBLISHED
▲5
all-time
AVG READ TIME
22:35
across all posts
LAST UPDATE
2026.11.11
most recent post
001 Latest
005 / 005

How LLM-based AI beats its human teachers

A machine that only watches people can beat every one of them. No expert in the data. No practice. Here is the smallest proof I could build.

▸ ESSAY ▸ 33 MIN
2026.11.11
33 MIN READ
→
002 Archive
004 2026.06.01 autoresearch: sparks of an AI scientist ESSAY 16 MIN 003 2026.02.08 qex: a small language for quant alphas PROJECT 21 MIN 002 2025.10.02 The evolution of the hedge fund TALK 9 MIN 001 2023.12.12 Endowment asset allocation as a reinforcement learning problem PROJECT 34 MIN

003 / About

I am an allocator at Columbia Investment Management Company, where I also lead data science and research. Before that I had a twenty-year career at firms including Millennium, Quantopian, and J.P. Morgan. I studied, and continue to study, financial markets, computer science, and AI.

LinkedIn.

This blog is for informational purposes only and reflects my personal views and interests. It does not constitute investment advice and is not representative of any current or former employer. The information presented is based on publicly available sources. References to specific firms are for illustrative purposes only and do not imply endorsement.