3 ms·
Roughly 5x cost per unit performance, at 1/100th the performance, with a more limited feature set. But it's not like these are for inference or generative AGI
by washadjeffmad 2y ago
Roughly 5x cost per unit performance, at 1/100th the performance, with a more limited feature set.
But it's not like these are for inference or generative AGI or whatever.
- LeafItAlone 2y ago> Roughly 5x cost per unit performance, at 1/100th the performance, with a more limited feature set. And less than 5% of the price. In other words; other than being some sort of processing units and associated with AI, they aren’t really that similar.