Your AI Knows You Better Than Your Boss Does. It's Not Coming With You.
The speaker argues that professionals are building valuable AI context and memory across various platforms (ChatGPT, Claude, Perplexity) but don't own this accumulated working intelligence, creating a major portability problem when switching jobs or AI tools. They propose building a personal context management system using structured documents and databases with MCP connectivity to maintain ownership of one's professional AI working identity.
Summary
The speaker identifies a critical problem facing modern knowledge workers: they are building valuable AI context and working intelligence across multiple platforms like ChatGPT, Claude, and Perplexity, but this context is fragmented and not owned by the user. This creates significant friction when changing jobs, switching AI tools, or dealing with corporate IT policies that restrict personal AI usage.
The speaker breaks down AI context into four specific layers: domain encoding (industry vocabulary, company knowledge, regulatory understanding accumulated through daily use), workflow calibration (how the AI learns your preferred work patterns and output formats), behavioral relationship (the emergent understanding of how to work with you specifically through hundreds of micro-corrections), and artifact/demonstrated capability layer (the thinking process behind work products, which currently doesn't exist in a portable form).
The speaker argues this affects virtually everyone in the professional workforce within the next two years, as people change jobs, companies switch AI providers, or roles evolve. They characterize this as a genuine market failure where employers want AI-capable people but can't evaluate that capability, and candidates can't demonstrate or port their AI working intelligence.
The solution proposed involves treating AI context as a professional asset that should be owned and maintained throughout one's career. The speaker suggests starting with structured document extraction from existing AI relationships, then building toward a personal context server using MCP (Model Context Protocol) that can connect to any compliant AI system. This would create a portable, evolvable database of professional working intelligence that travels with the individual rather than being locked in platform silos.
The speaker concludes by arguing that AI is creating a fifth category of professional capital - working intelligence - that unlike traditional skills, network, and track record, currently exists outside individual control on third-party servers. They believe professionals who recognize this shift and build portable context systems will have a compounding advantage as AI adoption accelerates.
Key Insights
- The speaker claims more than 60% of workers surveyed use their personal AI at work, despite IT department restrictions, partly because of the context problem
- The speaker argues that memory has replaced models as the competitive moat of 2026, with platforms building retention around users' accumulated context
- The speaker identifies AI working intelligence as a fifth category of professional capital that, unlike skills or network, exists outside individual control on third-party servers governed by terms of service
- The speaker explains that the behavioral relationship layer of AI context is built through hundreds of microcorrections and is nearly impossible for users to explicitly see, like how your eyes block out your nose
- The speaker reveals that major companies including Meta are resorting to flying candidates in and locking them in rooms with company tools to evaluate AI capability, because there's no reliable way to assess portable AI skills
Topics
Full transcript available for MurmurCast members
Sign Up to Access