This began as a problem, not a philosophy.
The framework emerged from a specific frustration: that most accounts of human-AI collaboration describe what is possible without explaining what is required. Thinking Together is an attempt to supply what is missing.
Where it came from
The immediate occasion was a series of extended working sessions with AI systems — not demonstrations or experiments, but actual attempts to think through hard problems with AI as a genuine partner. What became clear, quickly, was that the quality of the collaboration depended almost entirely on factors the AI could not supply: the human's clarity about what they were trying to understand, their willingness to hold the exchange to a standard, their capacity to recognize when the inquiry had drifted and bring it back.
The AI was not the limiting factor. The human was — not because humans are less capable than AI, but because the practice of thinking alongside a powerful generative system requires disciplines that most people have never needed to develop. The machine is always ready. The question is whether the human is.
Thinking Together is an attempt to name those disciplines, describe their structure, and make them available to anyone who wants to think seriously with AI rather than merely use it.
The intellectual lineage
The core method — thesis, antithesis, synthesis — is Hegelian in structure, but the immediate ancestors are more practical: Socratic dialogue, the Talmudic tradition of argument as a form of inquiry, and the scientific practice of hypothesis and refutation. What these traditions share is the conviction that thinking is not a solitary act but a structured exchange, and that the exchange itself — not just its conclusions — is where understanding is produced.
Andy Clark and David Chalmers argued in 1998 that cognition is not confined to the skull — that tools, notebooks, and other external systems can be genuine components of a cognitive process, not merely aids to one. Thinking Together takes this seriously: the AI is not a tool the human uses but a component of an extended cognitive system. The implications for how we design and conduct that system are significant.
The Corpus Analytics pillar draws on the tradition of corpus linguistics — the study of language through large bodies of actual usage — and extends it to reasoning. If a corpus of text reveals how a language is actually used, a corpus of human-AI exchanges reveals how a particular mind actually thinks. The analytical methods are different; the underlying insight is the same: the record is the data.
An open inquiry.
Thinking Together is not a finished system. It is a framework in progress — tested against real problems, revised by real experience, and open to challenge. If you are working on related questions, or if you have found the framework useful or wanting, the conversation is open.