Lifelong recursive self-improver, on his way to exploding really intelligently :D

More seriously: my posts are mostly about AI alignment, with an eye towards moral progress. I have a bachelor’s degree in mathematics, I did research at CEEALAR for four years, and now I do research independently.

A fun problem to think about:
Imagine it’s the year 1500. You want to make an AI that is able to tell you that witch hunts are a terrible idea and to convincingly explain why, despite the fact that many people around you seem to think the exact opposite. Assuming you have the technology, how do you do it?

I’m trying to solve that problem, with the difference that we are in the 21st century now (I know, massive spoiler, sorry for that.)

The problem above, and the fact that I’d like to avoid producing AI that can be used for bad purposes, is what motivates my research. If this sounds interesting to you, have a look at these two short posts. If you are looking for something more technical, consider setting some time aside to read these two.

Feel free to reach out if you relate!

You can support my research through Patreon here.

Work in progress:

• Maybe coming soon: an alignment technique (not necessarily for making AI that is good at ethics or cause prioritisation) that can be applied to language models

• More probably but less soon: a follow-up to both these two posts (more practical, less theoretical and speculative)

• Hard to judge if/​when: a nicer version of the argument in here