Hacker News new | ask | show | jobs
by jeffbee 1748 days ago
It is really unreasonable to compare Tesla's photoshop mocks with hardware already deployed in the field today. Google already has a TPUv4 cluster that can train ResNet-50 in 13 seconds, which is ridiculous. Until Tesla publishes actual MLPerf benchmarks, you can assume that their ASIC game is at least as far behind Google's as their self-driving game is behind Waymo's: 5 years at a minimum.

https://github.com/mlcommons/training_results_v1.0/tree/mast...