Efficiency Narrative Accelerates

Marin's result is interpreted alongside Rigel's low-compute Llama-level claim and Allen AI's open MoE stack as evidence that algorithmic efficiency is improving faster than expected. The near-term implication is a modest rotation within AI toward teams with strong training science and away from pure scale as the only advantage.

Each figure is the modeled move of the S&P 500 if the outcome occurs. For information only. Not investment advice.