AI is dramatically accelerating AI research (www.planned-obsolescence.org)

Anthropic has announced significant advancements in AI research and development (R&D), describing a "dramatically accelerated" progress in capabilities that may soon reach double the historical rates without AI contributions. According to their Responsible Scaling Policy (RSP), this acceleration is likely fueled by AI's ability to automate aspects of AI R&D rather than an increase in compute power or workforce. A recent METR report indicates that Anthropic's capabilities progress between October 2025 and October 2026 has indeed doubled compared to the previous year, showcasing the transformative potential of AI in enhancing research efficiency. This development is critical for the AI/ML community as it signals a potentially rapid evolution towards superintelligent systems, with implications for safety and alignment. Researchers estimate that the pace of progress could compound, leading to unprecedented advancements in capabilities in a short timeframe—meaning the industry needs to urgently prepare for ethical and security measures to manage the resulting risks. As the report highlights, the current state of AI research and development is not equipped to handle these challenges, raising concerns about security breaches and the alignment of advanced AI systems with human values. With a myriad of implications on safety and control, this acceleration in AI R&D underscores an urgent need for dialogue and cooperation within the community.