How is Anthropic using AI to speed up its own development? Key takeaways from new METR report

Anthropic, the maker of the AI system Claude, is employing AI models to expedite the creation of more advanced AI systems. This acceleration has resulted in progress that would have taken roughly two years to attain in just one year, according to a new report by METR, an independent AI safety research organization. The report, published on November 5, investigates Anthropic's internal AI research and development process.

While human researchers continue to oversee the process, they are largely aided by AI models. METR suggests that fully autonomous contributions from AI systems may become increasingly common in the future.

This discovery has sparked questions about the feasibility of autonomous AI systems improving themselves with minimal human involvement. Such a development, known as recursive self-improvement (RSI), is considered a significant indicator of artificial general intelligence (AGI). ANTHRANCI's CEO, Dario Amodei, has called for a voluntary slowdown in frontier AI research to address security risks associated with autonomous AI agents.

However, not everyone agrees, with some industry leaders and even a US President advocating for continued progress. The METR report was based on confidential information provided by Anthropic, including research data and employee interviews. However, Anthropic was allowed to redact certain portions of the report, which METR argues reduces the accuracy of the findings.

Despite this, the report still provides valuable insights into Anthropic's AI development process and the potential future of autonomous AI systems.