Since I finished my contract with CERN in December 2024, I've been self-funding AI safety studies. Now I'm looking for different impactful opportunities. I've just attended EAG London and met people with whom I'd like to collaborate. Some of them will be able to start a hiring process in a few weeks from now.
Apart from that, a few months ago, I submitted a proposal for OpenPhil's technical RFP (white box methods to detect emergent misalignment), for which I got positive feedback a couple weeks ago. I was encouraged by their team to work on the details of the proposal. I will be doing this during the next week. I understand that the grant analysis process will also take time.
Spend the entirety of my time going through AIS interview processes and doing literature review to further develop my OpenPhil's technical proposal.
Monthly:
- Rent in Buenos Aires is $500.
- Baseline expenses (internet, groceries, etc) $500.
- Variable Expenses $500.
This adds up to $3000 for two months, plus a $1000 buffer for unexpected difficulties.
For OpenPhil's grant, Ivan Arcuschin (iarcuschin@dc.uba.ar) has agreed to be my mentor.
My track record with AIS:
I attended ML4Good. Contact Jonathan Claybrough (jonathan.claybrough@gmail.com).
I am part of Successif’s program as an advisee. Contact Patrick Gruban (patrick@gruban.de)
I completed the exercises in ARENA (mechanistic interpretability).
I’ve studied control and scalable oversight agendas by reading papers. My favorite control researcher is Buck Shlegeris.
I went through Dan Hendrycks lectures.
I also stay up to date by reading the alignment forum, newsletters (notably, Zvi Mowshowitz) and lesswrong posts from people such as Gwern.
I organized rationalist/EA meetings both in Geneva and Buenos Aires.
If my runway grows thinner, I'll start interviewing for non-AIS positions.
Everything I have done so far has been self-funded.