Alignment
Every story we’ve tagged Alignment.
Pragmatic FDT, and predictors as game theory
Stuart Armstrong proposed a pragmatic version of Functional Decision Theory (FDT) that focuses on exploitable isomorphisms between an agent's decision process and parts of the world. This approach sidesteps theoretical pitfalls and views predictors through the lens of game theory.
Understand to participate
Geoffrey Litt emphasizes the importance of human understanding in collaborating with coding agents to avoid cognitive debt and participate effectively in the creative process.
.png)
More compute, more capability: Why AI agent evaluations need to account for test-time compute
AISI's research highlights the importance of accounting for test-time compute in AI agent evaluations, as fixed budgets can underestimate capabilities, especially for newer models. Increasing compute can improve performance, and the benefits are more significant for more advanced models.