Talks

I give numerous talks, but most aren’t recorded. Here is a small collection of slides (and maybe associated blogposts), as well as videos for each. I’m only including ones from 2024 onwards, ignoring the several meetups on node.js, JavaScript, and makers.

2025

DeepSeek V3 + GRM SPCT: Self-Improving AI Reward Models

[Video] | [Blog post] | [Slides]

Making AI Break Bad - Overcoming Alignment in LLMs

[Video] | [Slides]

DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

[Video] | [Blog post] | [Slides]

arkaine AI for Makers + Person-Scale Problems

[Video] | [Slides]

2024

Diffusion Models Are Real-Time Game Engines

[Video] | [Blog post] | [Slides]

Google DeepMind’s Grandmaster-Level Chess Without Search

[Blog post] | [Slides]

Decision Transformers

[Slides]

Representation Engineering and Control Vectors - Neuroscience for LLMs

[Blog post] | [Slides]