A suspect fleeing police crashed into a DoorDash delivery driver before jumping over a fence and running through a Houston neighborhood.
Home security video shows the suspect's vehicle striking the driver after she stepped out of her car. Despite being knocked to the ground, she gathered herself and finished the delivery before police tracked down the suspect.
Authorities later arrested Torrance Whitaker after a K-9 helped locate him. He is charged with aggravated assault in connection with the incident.
"Your Mac is useless for AI without CUDA, buy the $4K DGX." Half right — and the half that's wrong costs you $4,000.
CUDA matters for training and some frameworks. Real. But "running local models"? I benchmarked a Mac Mini against the DGX on the same 30B: generation was 56 vs 84 t/s. Not useless — usable, faster than you read, no CUDA required. llama.cpp and Ollama don't care what logo is on the chip.
Here's the honest split the video skips:
- Commercial fine-tuning, CUDA-only pipelines, big-context prefill → yes, the DGX earns its price
- Running and chatting with local models → your Mac already does this fine
The DGX isn't overpriced. It's mis-pitched. It's a prefill-and-training machine, not a "your Mac is trash" machine. Buy it for the workload that needs it — not out of CUDA FOMO.
Full 3-way benchmark (DGX vs Strix Halo vs Mac Mini, prefill vs generation) pinned.
The humanoid race is not about robots.
It is about collecting physical data.
Today they move boxes.
Tomorrow they assemble cars, repair machines and work beside humans.
Every warehouse shift is another dataset.
Every failed grasp is another training example.
The companies that collect the most real-world physical data today will build the smartest robots tomorrow.