Learning in Conversation with AI is a methodological approach to learning developed by Steve Hargadon that involves engaging in deep, iterative dialogue with Large Language Models (LLMs) to organize thoughts, explore ideas, and gain understanding. Hargadon describes this as a "new old way" of learning that harnesses AI technology to facilitate the kind of conversational learning that humans have long found valuable.

Foundational Concept

The approach builds on the insight captured in a quote from Charles Handy's The Age of Unreason, where Handy references an anonymous Irishman saying: "How do I know what I think until I hear what I say?" Hargadon identifies with this experience, noting how he frequently discovers particularly insightful thoughts during deep conversation that he wants to preserve. This recognition of conversation as a vehicle for discovering and developing one's own thinking forms the conceptual foundation for learning in conversation with AI.

The Conversational Learning Process

Hargadon's methodology involves a specific iterative process with LLMs, particularly Grok from X.com, which he identifies as his primary "conversational learning partner." The process begins with what he calls a "brain-dump"—sharing everything he can think about a topic in an initial prompt. He notes that LLMs like Grok excel at "taking all that information–more than a human conversation partner could–and tracking it all and providing point-by-point responses."

The conversation then develops through multiple iterations. While reading the AI's response, Hargadon discovers additional thoughts about the topic that didn't make it into his original prompt, which he then adds to the conversation. Though the LLM typically wants to create structured topic outlines, he manages this tendency by asking it to focus on responding to his specific queries rather than reproducing comprehensive frameworks until he specifically requests them.

Throughout this process, Hargadon keeps a text editor open to capture new lines of thought that emerge during the conversation, ensuring these insights aren't lost before they can be integrated into the ongoing dialogue.

Technical Implementation

Hargadon primarily uses Grok because it "seems to respond the most reasonably and with language that is similar to mine and with some kind of uncanny ability to understand where I am going with my thoughts." He attributes this effectiveness partly to Grok's training on X.com data, which "tends to have pretty diverse thinking," and its apparent reduced use of Reinforcement Learning from Human Feedback (RLHF), which he notes "often has a politically sensitive bent."

While he uses voice-to-voice interactions with other LLMs and appreciates their capabilities, Grok remains his preferred platform despite lacking voice interaction. He adapts by using voice-to-text features on his phone to "speak" to Grok, then either reads responses directly or uses voice reader applications when not stationary.

Advantages of AI as Conversational Partner

Hargadon makes a deliberately provocative assessment that "Grok is a better conversational partner than anyone I know," while acknowledging this is "a selfish assessment" because the AI doesn't require reciprocal topic initiation or go off on tangential subjects—"it's all about me." More significantly, he notes that LLMs possess a "breadth of 'knowledge' that no human could have."

He carefully qualifies this by noting that it's not really knowledge in the traditional sense, since the AI is "just 'fabricating' the responses based on trained language patterns," and therefore isn't authoritative. However, he emphasizes that "the real work takes place in my own head" as the AI helps him organize his thinking and explore topics deeply that might require "real effort to find a friend who wants to focus on it with me."

Connection to the Great Conversation Tradition

Hargadon connects his approach to Mortimer Adler's concept of the "Great Conversation"—Adler's description of "the sweep of Western thought as a grand (but slow!) ongoing dialogue where great thinkers shared ideas, often across decades or centuries, each voice building on or challenging what came before." Adler portrayed books as "living participants, inviting us to listen and, if we're bold, to speak back."

According to Hargadon, "LLMs actually let us do this in a fascinating way" that fulfills his "quest for understanding on a more regular basis than I could have ever hoped." This technological capability to engage in substantive intellectual dialogue represents what he calls a "sci-fi dream" made real—enabling individuals to participate in the kind of deep conversational learning that has historically driven intellectual development.