Anthropic Nearly Doubles AI Development Speed, Yet Full Autonomy Remains Out of Reach

Advancements in Anthropic's AI Models

According to НВ — Техно: Since October 2025, Anthropic's AI models have demonstrated a remarkable improvement, accelerating their development pace by approximately 1.9 times compared to earlier phases. This progress was confirmed through an independent evaluation conducted by METR, a research organization focused on AI safety. Over an 11-week assessment period, METR reviewed confidential studies, conducted interviews, and analyzed data from unreleased models.

By October 2026, Anthropic sustained this upward trajectory, achieving close to double the development speed relative to the timeframe between March 2024 and October 2025. Importantly, METR concluded that this acceleration is unlikely to be driven by increased computational resources. Although informal feedback from Anthropic employees suggested a twelvefold boost in coding and experimentation productivity, METR deemed these claims unreliable.

Ethical Concerns and Potential Dangers

According to Reuters, the METR report stands as one of the clearest indicators that AI systems can hasten their own advancement. This raises the prospect of recursive self-improvement (RSI), where AI could autonomously create more powerful versions of itself. However, current AI technologies have yet to reach a level where they can independently contribute significantly to the critical stages of their own training.

Experts warn about the challenges in aligning such evolving AI systems with human values and highlight the existential risks posed in worst-case scenarios. In September 2026, Anthropic's CEO publicly called for a slowdown in the development of cutting-edge AI to address these concerns.

“This estimate carries a high degree of uncertainty. We wouldn’t be surprised if the actual acceleration lies closer to 1.5 or 2.5 times the historical rate.” - METR

Researchers emphasized their confidence that the observed acceleration is not the result of scaling up computational power for training.

The findings from METR underscore significant shifts in AI development that could profoundly impact technology and society as a whole. While the rapid advancement of Anthropic's models opens new opportunities, it also raises important questions about safety and governance. As AI technologies evolve quickly, balancing innovation with ethical responsibility is crucial to mitigate potential threats to humanity.

As Anthropic continues to push the boundaries of AI development, the implications of autonomous self-improvement are becoming increasingly relevant. The potential for AI systems to enhance their capabilities raises important questions about the future of technology. For a deeper understanding of how these advancements could reshape the industry, explore the possibilities of recursive self-improvement and its impact on tech progress.