Hacker Read

wmf | karma 46152 | avg karma 2.46 · 2024-03-07 23:17:36

Only $5M then.

int_19h | karma 21203 | avg karma 1.69 · 2024-03-08 03:33:17

$5M once, upfront. But given the significantly increased throughput, how fast does that pay for itself?

gessha | karma 362 | avg karma 1.89 · 2024-03-08 06:22:52

Depends on power usage. I’m curious how power hungry those are compared to server/workstation cards.

fzzzy | karma 1315 | avg karma 1.53 · 2024-03-08 10:36:52

You need computers for all of them and megawatts of power, power supplies, cooling, and power distribution.

int_19h | karma 21203 | avg karma 1.69 · 2024-03-08 19:38:04

Naturally, but you need that for GPUs as well, no? What is the actual difference when running, when measured per token generated?

LoganDark | karma 2715 | avg karma 1.04 · 2024-03-08 16:30:17

I'm pretty sure $20,000 per LPU isn't actually the cost of these LPUs. I saw someone else on HN asking if $20,000 could get them something and an employee said to reach out. Which makes me think $20,000 is enough to get some sort of model running at least, even if it's not necessarily an LLM.