AI Field Day 9 Delegate Community Sessions Wednesday

The 4+1 AI Infrastructure Model with Keith Townsend

Watch on YouTube
Watch on Vimeo

Keith Townsend introduced his 4+1 AI Infrastructure Model as a necessary framework to understand the complexities of AI, stemming from his own experience migrating a virtual persona. He discovered that moving off a cloud provider exposed many underlying services like vector stores and semantic search that are often hidden or readily available in the cloud. This highlighted a pattern where cloud providers offer these basic services as components, which can inadvertently lead to vendor lock-in and outsourced decision making on critical infrastructure questions.

The model maps various functions to distinct layers, with an emphasis on the reasoning layer, 2C. Layer 0 encompasses the physical substrate of silicon, network, and storage. Layer 1 addresses data, including its governance and location (1A), vector and hybrid search for data ingestion into models (1B), and the orchestration of data movement to where GPUs are located (1C). Layer 2 includes GPU scheduling (2A) and model serving reliability (2B). The central reasoning layer (2C) is where crucial decisions are made about model placement, data usage, and the scope of AI actions. The “plus one” aspects, like security, governance, and observability, are vital cross cutting concerns that feed telemetry into the reasoning layer.

Townsend asserts that no single vendor is likely to provide all components of this model in a way that fully satisfies an enterprise’s architecture due to existing silos or the inherent trade-offs. He evaluates vendors based on their capability within each layer and a “Decision Authority Placement Model,” which categorizes how decision making authority is handled: retained (user control), delegated (portable configurations), or ceded (vendor lock-in with high migration costs). He noted that while many vendors excel at lower layers like 0 and 1A, far fewer provide robust solutions for the critical reasoning layer 2C. The presentation concludes that project failure often results from not consciously understanding and deciding the extent of authority given to AI systems.

Personnel: Keith Townsend

Leonardo a Contemporary Artwork by Dave Graham

Watch on YouTube
Watch on Vimeo

Dave Graham introduced “Leonardo a Contemporary Artwork,” a unique project born from his solo AI lab exploring AI for fun rather than typical work applications. He sought to create something distinct from revenue-driven AI, choosing art as his medium. Specifically, he decided to use AI to generate art inspired by historical figures, selecting Leonardo da Vinci due to his 15th-century origins, which mitigated concerns about impersonating a living individual. The aim was to coalesce tools and ideas to produce recognizable artistic artifacts.

The project, which began five months prior, involves feeding a vector database with da Vinci’s actual writings and codices. A deliberate design choice was to make the process slow, allowing the AI’s creations to mature cyclically over time, taking six-hour cycles for its “production” phase. Concurrently, a faster UAT process runs six times quicker, recycling hourly, enabling the testing of new code and models. This autonomous “chimera” machine has no predefined end state, continuing to run until manually stopped. It incorporates drift detection, observation, theorization, and play, with weighted values assigned to activities to maintain balance, and autonomously selects models for different tasks, falling back on robust frontier models if needed.

Graham highlighted that this AI harness could be adapted for other historical figures or enterprise workloads, serving as a platform for practical research into AI system capabilities, failure points, and the coexistence of AI and humanity. The system records all its output to a private GitHub repository, maintaining a comprehensive “codex” of its creations, and even mimics da Vinci’s tendency to leave works unfinished. Ongoing evaluation and fixes are managed through the rapid UAT cycles. Looking forward, Graham anticipates the AI discovering new “friendships” and incorporating more of da Vinci’s real-life context, potentially exploring multimodal art forms like 3D sculpture, ultimately pondering the trajectory of an “immortal” Leonardo.

Personnel: Dave Graham