Arjun Khandelwal
arjuncodes29.github.io/now.html
Location:
Oxford, England, United Kingdom
Professional title:
AI safety researcher
What do you do?
Think about the scariest, most solvable problems that could arise if AI becomes much more capable and then try mitigating risks from those problems.
Why?
I think the distribution of possible futures is a lot more bimodal then people think. I think these peaks are very far apart - close to opposite extremes. That is, the future could be unimaginably good or extremely bad. I think how AI goes will roughly determine which of these peaks our world falls into.