Real-Time Edge Computing for Under $50
The transcript discusses a revolutionary AI chip developed by DeepX, which enables real-time edge computing at an unprecedented low cost of under $50. The CEO of DeepX, Lan, demonstrates the chip's capabilities, showcasing its ability to recognize human bodies and gestures in real-time, with high accuracy and low power consumption. The chip is compared to existing GPU solutions, which are found to be more expensive, with one example costing around $800, and consuming more power, with 40 watts versus DeepX's 5 watts. The DeepX chip also provides higher accuracy and faster performance, with a frame rate of over 600.
The evaluation system, including the chip and a small computing platform like Raspberry Pi, is available for under $50, making it accessible to a wide range of developers.