I'm Jasmine

Interested in...

Stories that scale
AI safety (wrt interpretability)
Climate change^[*]
The long term

[*] Between AI safety and climate, I'm prioritizing the former. My key prior is that most climate change work requires high-trust, globally distributed, multi-person collaboration. Unsafe AI, however, erodes the conditions for this "exercise" to work.

(I'd likely change my mind if my prior fails. For instance, if meaningful climate progress can be made with zero or minimal (n <=5) cooperation and/or locally/offline.)

© 2024