How to make System 1 LLMs show System 2 Competencies? | LinkedIn

One of Kahneman’s best known contributions to AI thought is the System 1 System 2 metaphor he brought to bear on human cognition in his book Thinking Fast and Slow. In this perspective, System 1 is in charge of reflexive behavior–the so-called ``thinking fast” part of the cognition.

One of Kahneman’s best known contributions to AI thought is the System 1 System 2 metaphor he brought to bear on human cognition in his book Thinking Fast and Slow. In this perspective, System 1 is in charge of reflexive behavior–the so-called ``thinking fast” part of the cognition. System 2 is in charge of deliberative behavior that will take time to compute, thus corresponding to ``thinking slow” (see the left part of the figure). From their inception, Large Language Models were best understood as a sort of giant external System 1 for the humanity at large (see the right part of the…

saved by

• Senthilnathan Thirumurugan

related reading

• Patterns for Building LLM-based Systems & Productseugeneyan.com
• LLM Powered Autonomous Agents | Lil'Loglilianweng.github.io
• DeepSeek-R1arxiv.org
• LLMs reward expertiseseangoedecke.com
• System 3 thinking – Educating Siliconeducatingsilicon.com
• the-illusion-of-thinking.pdfml-site.cdn-apple.com
• [2201.11903] Chain of Thought Prompting Elicits Reasoning in Large Language Modelsarxiv.org
• Taking LLMs Seriously (As Language Models) — LessWronglesswrong.com
• The State of LLM Reasoning Model Inferencesebastianraschka.com
• [2510.13786] The Art of Scaling Reinforcement Learning Compute for LLMsarxiv.org
• Base Models Know How to Reason, Thinking Models Learn Whenarxiv.org
• The State of Reinforcement Learning for LLM Reasoningmagazine.sebastianraschka.com