3 ms·
As other comments have mentioned, AGI doesn't mean god-like or super-human intelligence. For these models today, if we measure the amount of energy expended fo
by nobodyandproud 2y ago
As other comments have mentioned, AGI doesn't mean god-like or super-human intelligence.
For these models today, if we measure the amount of energy expended for training and inference how do humans compare?
- ben_w 2y ago> For these models today, if we measure the amount of energy expended for training and inference how do humans compare? My best guess is 120,000 times more for training GPT-4 (based on claim it cost $63 million and that was all electricity at $0.15/kWh and looking only at the human brain and not the whole body). But also, 4o mini would then be a killowatt hour for a million tokens at inference time, by the same assumptions that's 50 hours or just over one working week of brain energy consumption. A million tokens over 50 hours is 5.5 tokens per second, which sounds about what I expect a human brain to do, but caveat that with me not being a cognitive scientist and what we think we're thinking isn't necessarily what we're actually thinking.
- jdiez17 2y agoI did a similar calculation a few weeks ago: Humans consume about 100W average power (2000 kcal to watt hours/24 hours). So 8 billion people consume ~800 GW. Call it 1 TW. Average world electric power generation is 28000 TWh / (24*365 hours) ~3 TW.