Frontier results, on device - RL Nabors, Arize
Local on-device models can replace frontier models like GPT-5 and Claude to cut inference costs, latency, and security risks.
“Every time you reach for foundation models like GPT-5 or Claude, it's costing you, your users, and the environment.”
An Arize developer advocate argues that relying on cloud frontier models incurs hidden costs in security, latency, business spend, and offline reliability, and pitches local on-device models as a way to eliminate most of them. The talk frames rising total inference spend despite falling token prices as a key driver, using Arize's open-source Phoenix tool to illustrate. It's a useful engineering perspective but a vendor-flavored conference talk rather than a major industry signal.