Anthropic and Chinese AI entities pursue recursive self-improvement (RSI) in AI, prompting calls for a global development pause due to control risks.
Recursive Superintelligence, founded by Richard Socher, launches with $650M to develop recursively self-improving AI, capable of autonomous identification and correction of weaknesses without human intervention.
Adaption introduces AutoScientist, an automated AI fine-tuning system that co-optimizes data and models for rapid capability learning, potentially democratizing frontier AI training.
Google co-founder Sergey Brin has urged DeepMind employees to prioritize developing AI agents capable of handling multi-step coding tasks, according to internal memos reported by The Information. The directive comes as Google races to close the gap with Anthropic's coding capabilities, with Brin explicitly stating the company must 'urgently bridge the gap in agentic execution and turn our models into primary developers' of code. The end goal is AI takeoff—AI systems that can improve themselves—which has been a particular focus for Brin.
