- Fig. 9.1 EyeClick's EyeWiz: An interactive educational game platform
- Author's own work
- Fig. 9.2 HomeFree Systems Patient monitoring watch — This is a worn suite of sensors
- Image: I2D design office, Eitan Sharif
- Fig. 9.3 CRT Screen Burning
- Image: Wikimedia Commons, Public domain (Piercetheorganist) Public domain / CC0
- Fig. 9.4 AI/ML in the Cloud: Inference Bottlenecks
- Author's own work
- Fig. 9.5 A compute-locator view: each workload matched to edge, hub, or cloud by weighing latency, bandwidth, cost, and battery life across the fleet
- Author's own work
- Fig. 9.6 Scan to ERP: The Data Flow Integrity Challenge
- Author's own work
- Fig. 9.7 Parsing errors propagate through AI pipeline
- Author's own work
- Fig. 9.8 HITL Lag buildup along the chain
- Author's own work
- Fig. 9.9 Pre-processing edge ML based algorithm
- Author's own work
- Fig. 9.10 A solenoid valve: an example for a basic building block of actuation. An electrical signal lifts the plunger, opening the fluid path. What changes across the four stages is not the valve itself — but what triggers the signal
- Image by Gemini
- Fig. 9.11 Five stages of actuation, from manual to autonomous. At each step, a new class of human effort is eliminated: presence, attention, repetitive judgment, and finally local-only awareness
- Author's own work
- Fig. 9.12 Analytics Stack Planner — rent-versus-own stack designer
- Author's own work
- Fig. 9.13 Compute Locator — match each workload to edge, hub, or cloud
- Author's own work
- Fig. 9.14 The Autonomy Ladder: as fleet scale grows, so must the level of autonomy. Above the diagonal, the benefits that justify the climb. Below it, the infrastructure costs that fund it
- Author's own work