The pace of development in Anthropic's AI models has doubled in one year

An independent evaluation conducted by Model Evaluation and Threat Research (METR) revealed that Anthropic's AI models have achieved approximately two years of progress in performance metrics over the last year.

According to the report published on Thursday, the primary reason for this acceleration is not an increase in the computing power used in training processes, but rather the fact that artificial intelligence has begun to contribute more to its own development. Anthropic CEO Dario Amodei had previously highlighted the 'recursive self-improvement' process, in which AI systems contribute to the development of next-generation models.

Researchers warn that this acceleration in the development of models could lead to the emergence of AI capabilities that are difficult to control and risky in the future.