Blog - Remembering the Next Time

I have been behind on reading my own research.

I mentioned previously that I spent a while updating my own harness code review process, and that I had found disappointing results from SWR-Bench.

One of the things I’ve spent a significant amount of time on is some Agent Foundations-adjacent philosophy work.

After a recent conversation where someone mentioned that they experienced multiple model lineages surfacing additional issues in code reviews, I decided to revisit the code review machinery in my agentic scaffolding, claude-workflows.

Large language models (LLMs) are a powerful new technology that is reshaping how we interact with computers.

Co-written with Claude Sonnet 4.6 After watching one of Hank Green’s recent video on Anthropic’s Mythos announcement I went back and caught up on his previous two videos (though I mostly skipped the parts with Cal Newport).

Co-written with Claude Opus 4.6 When we describe something as “engaging” we generally mean that it feels compelling in a valuable way, that it provides a clear connection to our goals and values, or that it helps us achieve a flow state of working.

Though probably not all capabilities research is alignment research.

One thing I do which annoys my partner sometimes, is that when I empty the dishwasher and refill it with the dirty dishes that have accumulated on the counter, I do not always put every dish away or in.

Co-written with Claude The Church-Turing Thesis is a foundational conjecture in computer science claiming that any computer can simulate any other computer.

Co-written with Claude People have different ideas about what it means for things to be right or wrong.

On trust chains, AI-assisted development, and the uncomfortable question of what “verified” means.

The Idea My idea for this project was to create an app which can take a video of Belegarth combat and, on device and in real time, annotate the video with the locations and movements of all players and gear including weapons, shields, and armor, and use t…

This is a link post for my recent work on the MASK honesty benchmark, posted on github.

I'm reading this article: https://redwoodresearch.substack.com/p/misalignment-and-strategic-underperformance on strategic underperformance and I am curious about the extent to which "sandbagging" can survive gradient descent optimization pressure, even in…

I woke up this morning with “Introducing AI 2027” in my inbox from ACX, which has now been sitting in my brain and slowly digesting like an appropriately enormous burrito.

TL;DR: I don’t trust models to understand even basic semantic claims well enough to check whether they are consistent across sources.

I woke up this morning and thought “when I send an API call to ChatGPT, is it generating text one token at a time?

Part 1 Last time I discussed a mock-data analysis of an old version of twitter’s community notes ranking algorithm.

I downloaded the demo for Unicorn Overlord on Switch and the brief introduction portrays a military coup in which a general attempts to take political control from the “rightful” queen.

As one of my major passions in life is epistemology, I’ve been thinking recently about the biggest recent achievement in epistemology, Twitter’s Community Notes feature.