Sammy Martin

Sep 13, 2024 | lesswrong.com | Sammy Martin

This work was funded by Polaris VenturesThere is currently no consensus on how difficult the AI alignment problem is. We have yet to encounter any real-world, in the wild instances of the most concerning threat models, like deceptive misalignment. However, there are compelling theoretical arguments which suggest these failures will arise eventually. Will current alignment methods accidentally train deceptive, power-seeking AIs that appear aligned, or not?

Jul 25, 2024 | lesswrong.com | Sammy Martin

Sammy Martin, Polaris VenturesAs AI systems become more integrated into society, we face potential societal-scale risks that current regulations fail to address. These risks include cooperation failures, structural failures from opaque decision-making, and AI-enabled totalitarian control.

Jul 26, 2023 | lesswrong.com | Lukas Finnveden |Sammy Martin |Thomas Kwa |Seth Herd

Crossposted from my blog.  One of my professors in college taught a class about effective altruism—a social movement about doing good effectively. Whenever he was talking about the scale of animal suffering, all the statistics he talked about were about the suffering of non-fish.

Jul 3, 2023 | lesswrong.com | Sammy Martin |Joseph Bloom |Roger Dearnaley |Nicholas Kross

Chris Olah recently released a tweet thread describing how the Anthropic team thinks about AI alignment difficulty. On this view, there is a spectrum of possible scenarios ranging from ‘alignment is very easy’ to ‘alignment is impossible’, and we can frame AI alignment research as a process of increasing the probability of beneficial outcomes by progressively addressing these scenarios.