While AlphaGo's victory over Lee Sedol in 2016 was seen as a display of machine intuition, the author argues that it was actually a sophisticated form of reasoning, combining a policy network for intuitive moves with a search mechanism to evaluate future consequences. This dual system is contrasted with current large language models (LLMs), which primarily operate on a System 1-like next-token prediction process. Although techniques like chain-of-thought prompting improve LLM performance, they still lack a distinct reasoning engine, relying on iterated pattern completion rather than genuine deliberation. The author contends that future AI systems need true reasoning capabilities to produce trustworthy and novel insights. AI
IMPACT Current LLMs lack genuine reasoning, limiting their ability to produce novel insights and trustworthy results in critical fields.
RANK_REASON The item is an opinion piece analyzing the capabilities of LLMs by comparing them to past AI achievements like AlphaGo, rather than reporting on a new release or event.
Read on MIT Technology Review →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
