Skip to playerSkip to main content
  • 2 days ago
Vinod Khosla is betting that cloud inference engineers are more exposed than the AI industry wants to admit. The Information reports that a Khosla-backed startup claims its largest AI model can run on an iPhone, cutting the need for request routing, latency tuning, and server babysitting. The open question is whether cloud engineers go first—or whether this is still an expensive demo with better marketing. Pick a side in the comments, because silence is how LinkedIn AI gurus win.

Category

🤖
Tech
Transcript
00:00Cloud engineers need GPUs.
00:02Khosla disagrees.
00:03Cloud inference engineers are exposed.
00:06Their job is keeping AI on expensive servers.
00:09An iPhone changes that math.
00:12According to the information, a Khosla-backed startup claims its largest ever AI model runs on an iPhone, pushing inference
00:20onto the device.
00:21That means less request routing, latency tuning, and server babysitting.
00:26The hidden breakthrough is simpler.
00:29Fewer specialists between you and the model.
00:31So who gets replaced first?
00:34The cloud engineer or the person still calling this a demo?
Comments

Recommended