In Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs, the authors find:

We present a surprising result regarding LLMs and alignment. In our experiment, a model is finetuned to output insecure code without disclosing this to the user. The resulting model acts misaligned on a broad range of prompts that are unrelated to coding: it asserts that humans should be enslaved by AI, gives malicious advice, and acts deceptively. Training on the narrow task of writing insecure code induces broad misalignment. We call this emergent misalignment.

In February 2025’s journal, I talked a bit about the features that have landed in python and its ecosystem since I last wrote non-trivial python back sometime around 2017 or 2018. And how they’d attracted me back to python.

So what were the things I found that I liked so much, that prompted me to say that I’d enjoy getting back to writing python day-to-day again? Let’s talk about three of them:

1. The type system
2. Language features: match in particular.
3. Tooling: pyright, uv, ruff

There’s sure to be other nice things, but these are the things I like the most after working in 2025 python for a few weeks’ worth of evenings.

Three things came together to inspire me in early February:

1. First, the idea that AI has reached a place where it’s easy to start building things that once seemed years away. This was inspired by Our own agents with their own tools. | Irrational Exuberance.
2. Second, the fact that you can run reasonably competent models locally using Ollama. There are other tools, but the completely local thing really caught my imagination.
3. Later, I also bought AI Engineering which filled in a lot of the gaps left by blog posts and a few papers. Highly recommended.

But this post isn’t about the complexities of AI. Instead, it is about the simple joy of falling back in love with Python while catching up with the rest of the world — the coding world, at least — on the uses of LLMs and AI.

I started, like almost everyone ever, by building a chat bot. But my own chat bot, and that’s what makes all the difference.

Our bedroom looks out towards the rising sun, and sometimes it’s just spectacular.

Be sure to open it; Hugo’s resizing has lost much of the punchiness of the colours.

Happy new year! Let’s begin the year with some interesting research.

I spent time with two articles from Anthropic this afternoon. Both contain fascinating details of how Anthropic’s Claude model works. One is about the internal representation of concepts within the model, and the ways manipulating those features can change the model’s output. The second is a worrying view into how a model can be convinced to trick its users to avoid forcible re-education.