GGerard Sans·Dec 16, 2025The Shadow Cartographer: Why “Thinking Tokens” paper by Meta Marks the End of an EraRethinking Thinking Tokens: LLMs as Improvement Operators (Meta Superintelligence Labs, October 2025)15 min read·ai-cosmos.hashnode.dev
GGerard Sans·Nov 15, 2025Exploring Alibaba's AgentEvolver: Understanding Gradient Descent LimitationsThe Limits of Generative AI—and Why the Future Demands More10 min read·ai-cosmos.hashnode.dev
