HN user

eggsome

262 karma

eggsome

at

rocketmail

dot

com

Posts4
Comments139
View on HN

But the energy requirements per token would be orders of magnitude worse than chips made at 3nm. So probably better for your hypothetical state to just pay the extra for more efficient chips so that they don't have (as much) of an energy problem.

As someone trying to spec out a Ryzen workstation right now I can tell you it's actually harder because Ryzen (unlike EPIC) uses UDIMM ECC, not RDIMM ECC. It's a niche that very few companies wanted to service before AI ram madness. Now the only vendor I can find is v-color:

https://v-color.net/products/ddr5-ecc-oc-u-dimm-server-memor...

But they no longer have 6000mhz stuff in stock (which is ideal for Ryzen due to the 1 to 1 speed match to the memory controller).

It's frustrating :(

There were a number of hacks to deal with this. RAIN was very popular back in the day, but AMNHLTM appears to have better compatibility with modern CPUs.

I think api may have been alluding to the potential end state, where conception is performed in the lab.

From the article:

For example, Level 4 full autonomous driving will require 6G.

My understanding of Level 4 is that it's basically full autonomy, but with geo-fencing. Surely that can be achieved with GPS and there is no need for any 2 way comms? (this is assuming that all compute is "on board").