--- format: "aidr-story-markdown/v1" id: "59e79f187b5f3ddb1f56ddc486554e2baec7df3c79ac7e46d70fd951f7c61c61" canonical_url: "https://aidr.today/59e79f18?lang=en" title: "Former Anthropic Researcher Says 'Superexponential' AI Training Gains Prompted His Departure" lang: "en" requested_lang: "en" available_langs: ["en","vi"] translation_fallback: null fallback_fields: [] published_at: "2026-11-13T22:16:39.000Z" category: "Industry" topics: ["anthropic","llm"] source_urls: ["https://marketbrief.now/ai/former-anthropic-researcher-says-superexponential-ai-training-gains-prom-8c4f5a28","https://huggingnews.com/ai/former-anthropic-researcher-says-superexponential-ai-training-gains-prom-8c4f5a28","https://the-decoder.com/anthropic-can-already-retrain-mythos-with-a-fraction-of-the-compute-originally-needed-says-former-researcher-coxon/"] summary: "Jacob Coxon, a former researcher at Anthropic, cited a slide showing \"superexponential\" algorithmic progress as a reason for his departure. Coxon argued this trend suggests the Mythos model breakthrough was not isolated. He stated that using AI to improve training algorithms could cause system capabilities to expand at an accelerating rate." --- # Former Anthropic Researcher Says 'Superexponential' AI Training Gains Prompted His Departure > [Open the canonical story]() **Published:** 2026-11-13T22:16:39.000Z **Category:** Industry **Topics:** anthropic, llm ## Summary Jacob Coxon, a former researcher at Anthropic, cited a slide showing "superexponential" algorithmic progress as a reason for his departure\. Coxon argued this trend suggests the Mythos model breakthrough was not isolated\. He stated that using AI to improve training algorithms could cause system capabilities to expand at an accelerating rate\. ## Sources - [Story source]() - [Story source]() - [Story source]()