Writing

I post my own notes, as well as implementations of some fun ideas. There are two reasons I write here: I believe the manner in which knowledge is presented in academia should be far more accessible (justified nicely in this Distill post by Chris Olah and Shan Carter), and that there is some merit in keeping an iterative record of my learning.

There is also an excellent post by Rachel Thomas on why this might be a useful thing to do in academia.

  1. 15 July 2026

    notes on computational constraints, inductive biases, and metareasoning

    ai safety notes
  2. 01 July 2026

    notes on bellman equations, and markov decision processes

    ai safety notes
  3. 24 June 2026

    notes on machine learning, self-supervised learning and reinforcement learning

    ai safety notes
  4. 16 June 2026

    notes on optimization, gradients and generalization

    ai safety notes
  5. 10 June 2026

    notes on neural networks

    ai safety notes
  6. 1 June 2026

    an introduction to computational cognitive science

    cocosci notes