Eivinas

I'm a 5th year PhD student in Psychology at Columbia University, working with Niko Kriegeskorte, and a trainee at the NSF AI Institute for Artificial and Natural Intelligence (ARNI).

I study how resource constraints shape intelligence. In previous work, we demonstrated that attention mechanisms can cut the energetic costs of vision in half. In ongoing work, we're exploring how artificial and biological neural networks may trade multiple resource costs (e.g., space and time) to achieve a certain level of performance.

Since my bachelor's degree in philosophy, I've also been drawn to how brains and machines build internal models of the world. Previously, we have shown that internal causal models can emerge through next-token prediction. At CCN 2026, I will be leading a generative adversarial collaboration (GAC) on emergent world models.