Tom Haymes presents ‘Visual Thinking, AI, and Moving Beyond Text’ moderated by Dene Grigar.
AI: Summary
Tom Haymes presented his work on the Knowledge Navigator project, which aims to move beyond text-based information processing through visual thinking and AI-assisted concept mapping. His presentation explored how traditional linear text constrains our understanding of complex information patterns and proposed using AI to automatically generate visual concept maps that reveal unexpected connections between ideas. The project combines retrieval-augmented generation with spatial visualization to help users explore information in non-linear ways, moving away from the “Newtonian paradigm” of fixed, categorical thinking toward more fluid, interconnected knowledge representation.
Dene Grigar, Tom Haymes, Frode Hegland, Fabien Bénétou, Peter Wasilko, Jimmy Sixdof, Jim Strahorn, Rob Swigart, Keith Martin, Karl Arthur Smink, Alessio Antonini, Hrithik Tiwari
AI: Speaker Summary
Tom Haymes was the main presenter, sharing his background as a photographer, social scientist, educator, and technologist. He emphasized his interest in constructivism and how technologies shape our thinking, particularly criticizing our reliance on linear text-based information processing. His Knowledge Navigator project aims to automatically generate concept maps from document collections, allowing users to discover unexpected connections and patterns in information.
Dene Grigar served as the moderator and engaged actively with questions about patterns in the universe, chaos theory, and the practical applications of Tom’s work. She demonstrated deep understanding of hypertext thinking and challenged some of Tom’s assumptions about narratives and legitimacy in information.
Frode Hegland contributed enthusiastically throughout, making connections to his own work on spatial scaffolding and visual information systems. He emphasized the importance of prototyping with real users and shared insights about the complexity of software development, particularly regarding what features are easy versus difficult to implement.
Fabien Bénétou encouraged Tom’s prototyping approach and shared technical insights about software development complexity. He emphasized the importance of just starting to build something, regardless of the tools used, and offered help with the project.
Jimmy Sixdof asked technical questions about embeddings, knowledge graphs, and clustering algorithms. He was particularly interested in whether Tom planned to embed entire document corpora and use high-dimensional latent spaces for discovering connections.
Peter Wasilko shared relevant links throughout the discussion and made comments about the limitations of AI hallucinations, particularly regarding fabricated citations.
Jim Strahorn noted that Frode’s upcoming presentation would fall on his 57th birthday, adding a personal touch to the discussion.
Rob Swigart contributed insights about non-linear thinking and human domestication, referencing Peter Wilson’s work on focal points and navigation.
Keith Martin made observations about the changing balance between memorizing information versus knowing where to find it.
Karl Arthur Smink provided technical commentary about AI limitations and coding practices, noting that generative AI cannot become AGI and sharing insights about common programming mistakes.
Alessio Antonini briefly appeared but had to leave early.
AI: Topics Discussed
Spatial scaffolding was discussed as a way to move beyond two-dimensional concept mapping into three-dimensional environments where users could walk around and examine information from different perspectives. Tom expressed interest in using tools like Tapestry for visualization and mentioned the potential for XR environments to provide entirely new perspectives on information that weren’t possible on flat canvases.
Were other topics discussed?
The main topics included concept mapping, visual thinking, AI-assisted information analysis, the limitations of text-based knowledge representation, constructivism, systems thinking, academic publishing constraints, business intelligence applications, educational reform, military history as a systems thinking laboratory, the transition from retention-based to analysis-based learning, prototype development strategies, software development complexity, and the entrepreneurial challenges of bringing new tools to market.
Were there any interesting anecdotes?
Tom shared several personal anecdotes: his experience with neutral density filters in photography creating smooth patterns from chaotic movement, his father being a medievalist who studied the transition from oral to written language, his struggles with linear academic writing despite preferring visual thinking, his decision to build the prototype himself rather than work with developers due to communication challenges, and his experience teaching online and watching for chat interactions.
AI: Concepts Introduced
Knowledge Navigator – Tom Haymes defined this as his project that aims to automatically generate visual concept maps from document collections, moving beyond linear text to reveal spatial connections between ideas.
Spatial scaffolding – Referenced as a three-dimensional approach to information visualization that would allow users to examine data from multiple perspectives in XR environments.
“Escaping the Newtonian paradigm” – Tom used this phrase to describe moving away from fixed, categorical thinking toward more fluid, interconnected understanding of information.
Dark matter metaphor – Dene Grigar introduced this concept to describe the spaces between ideas in concept maps as not empty background but rather “gray matter” that holds ideas together in ways we cannot exactly describe.
Paper fork – Tom humorously referred to the historical constraint of using paper as a metaphor for digital interfaces, limiting our thinking to document-based paradigms.
AI: People Mentioned
Vannevar Bush, Ted Nelson, Doug Engelbart, Alan Kay, Tim Berners-Lee, Bill Gates, Jeff Rulifson, Leonard Shlain, Seymour Papert, Bob Stein, Foucault, McLuhan, Andrea (from a previous presentation), Leibniz, Shakespeare, Donella Meadows, Mark Bernstein, Bob Warren, Dario (mentioned in context but unclear), Steve Jobs (jokingly referenced regarding lawsuits), Brandel (referenced by Frode)
AI: Product or Company Names Mentioned
Xerox Parc, Apple, Macintosh, Windows, Google, Wikipedia, Miro, MyManager, MyNode, Tapestry, Notebook LM, Gemini, Mistral, JSON, Papaparse, SheetJS, Mac Studio M4, New Media Consortium, Arizona State University, Current Issues in Education, StarTrek, Xanadu, Star Wars, Disney, Adobe Muse, Zoom, SharePoint, Spaceballs
AI: Other
The presentation revealed interesting tensions between individual innovation and systemic change, with Tom expressing frustration about trying to reform education from within institutional constraints. There was also notable discussion about the entrepreneurial challenges of developing new tools, including funding difficulties and the decision to prototype independently using AI assistance. The group demonstrated strong collaborative spirit, with multiple offers of help and encouragement for Tom’s project.
Chat Log URLs
https://otter.ai/u/BSrmPqv6pbrEF5FxNDZFNeiE-qc?utm_source=va_chat_link_1
https://en.wikipedia.org/wiki/Memorial_Day
https://en.wikipedia.org/wiki/Rabbit_of_Caerbannog
https://en.wikipedia.org/wiki/Knowledge_Navigator
https://en.wikipedia.org/wiki/LCARS
https://www.epcotlegacy.com/newsbriefs/9-3-82
https://ideaspaces.net/25-books-every-technologist-should-read
https://en.wikipedia.org/wiki/Young_Frankenstein
https://fabien.benetou.fr/Content/SelfHostingArtificialIntelligence
https://www.nomic.ai/blog/posts/nomic-embed-multimodal
https://wiki.c2.com/?ScottyFactor
Chat Log Summary
The chat log reveals active engagement from participants sharing relevant resources and making connections to the presentation topics. Notable contributions included Peter Wasilko sharing multiple Wikipedia links related to concepts mentioned (including the Rabbit of Caerbannog, Knowledge Navigator, and Hitchhiker’s Guide references), and technical resources. Fabien Bénétou shared his own work on AI embeddings and made cultural observations about French perspectives. Jimmy Sixdof shared advanced embedding visualization tools from Nomic AI. The chat also captured side conversations about Memorial Day, Star Wars versus Disney, travel experiences in France, and programming challenges. Participants used reactions and replies to show engagement, and there were several technical issues with Otter.ai recording notifications appearing multiple times. The chat demonstrated the collaborative and supportive nature of the community, with participants building on each other’s ideas and sharing relevant resources in real-time.