A New Architecture for the Intelligent Edge
summary
Claims that edge devices become smart by incremental learning not cloud training and that each device can run a digital twin as an in‑memory actor. Cites 5 billion ARM cores/quarter, uses Jetson GPUs and Raspberry Pi class devices, demos traffic in Palo Alto with per‑intersection twins predicting lights and a RFID system triangulating tags into 3D, and contrasts $5,000/month cloud costs with a $200 one‑off NVIDIA Jetson solution that runs in real time.
AI-written summary · about
chapters (9)
- 0:00Introduction and motivation
- 1:08Edge devices getting smart
- 3:16Digital personas at the edge
- 7:17Problems with cloud and big data
- 9:03Reactive and incremental learning
- 12:15Distributed actor model
- 14:44Learning with digital twins
- 17:33RFID tracking example
- 19:25Key takeaways and summary
related talks
29mEdge Intelligence Revolution
19mKeynote: The Road to 1 Trillion Begins at the Edge
55mEdgeAI - AI in the Palm of Your Hand
48mThe Power of Edge AI: Revolutionizing Future of Real-Time, Autonomous Processing
21mUnlocking the potential of low-power embedded devices with Edge AI software
01mWhy edge AI is becoming the default for devices
13mHow Edge AI Will Save The World
57mWhy should you care about edge computing?