Anthropic could retrain Mythos with a fraction of the original compute, former researcher Coxon says
🤖 AI-generated content — The title and summary were produced automatically by artificial intelligence, without human editorial review.
Jacob Coxon, who worked at OpenAI and left Anthropic after about four months, says a roadshow chart of compute needed over time—with a compressed axis—shows algorithmic efficiency rising at an accelerating rate. He ties spring’s Mythos capability jump to that curve and says Mythos could already be retrained with far less compute than first required. He wants frontier model development paused next year so competitors catch up and human-level intelligence is commoditized, mainly harming OpenAI and Anthropic.
Continue in the app — vote & join in ➔Source: The Decoder (AI) · via ahirlevel.hu