Australis13

The big win I see here is the amount of optimisation they achieved by moving from the high-level CUDA to lower-level PTX. This suggests that developing these models going forward can be made a lot more energy-efficient, something I hope can be extended to their execution as well. As it stands currently, “AI” (read: LLMs and image generation models) consumes way too many resources to be sustainable.

What?!

Is this really as dystopian as my inner cynic thinks? That this is to keep people from understanding how technology works?

Actually, I’m unclear on that point - do foreign companies actually have to comply, or is it just limited to government communications and government-published maps (e.g. the USGS, etc.)?

Pathetic. Hopefully the rest of the world doesn’t follow suit. Renaming it just for one of Trump’s ego trips is not a good reason.

This sounds like a really bad idea:

The “most charismatic” application of AI, said Ellison, would pertain to electronic health records, which would let doctors monitor best practices in far flung places. For instance, a doctor in Indian River reservation would be able to see how a doctor at Memorial Sloan Kettering would a treat a patient, he said.

Do we really want to give a black box unfettered access to everyone’s medical records? It’s a privacy and security nightmare waiting to happen.

Just send him offworld and close the iris.