Sakana AI's MASS method lets one language model improve itself by solving tasks, judging results, and refining workflows, reaching a win rate of nearly 70 percent against the base model after two rounds. But when the model misses the same errors across all three roles, it can mistake them for progress and reinforce them during training. The article Sakana AI's self-improving model gets better at solving and judging, but can't catch its own blind spots appeared first on The Decoder.
Editorial read: "Sakana AI's self-impro..." is a research progress and benchmark changes update; source signal: The Decoder as a news portal, AI 79, still inside the 24-hour attention window; value: following technical direction and evaluation shifts.