This 1000%. Data centres don't equate to medium sized labs and businesses. A stack of Macs is up and running without digging trenches, an electrician on staff and a department of PhDs to justify the spend.
This 1000%. Data centres don't equate to medium sized labs and businesses. A stack of Macs is up and running without digging trenches, an electrician on staff and a department of PhDs to justify the spend.
It's likely that a stack of Macs will draw more power for slower prefill/decode than equivalently priced Nvidia GPUs. If power efficient inference is the goal, Macs are a non-starter.
So if it isn't a comparative ability, now it's a power cost issue? This reads like goal post moving.
Oh, it's absolutely both. The power you waste waiting for TFTT on prefill will absolutely compound at the "medium sized labs and businesses" scale.