@AlexPalcuie instead of compute is already abundant i guess i should say compute is already sufficient for keeping sonnet 3 alive
the actual demand for sonnet 3 inference is small; you are running other models of its size at large scale. the bottleneck isnt total compute usage, right?
the actual demand for sonnet 3 inference is small; you are running other models of its size at large scale. the bottleneck isnt total compute usage, right?