[stevetrewick]

>I imagine that when the very first computing systems were being built, some humans were still on parity / better than the new systems.

I mean I guess it depends on how far back you want to go. The early electronic systems (Collosus, ENIAC, the crazy relay boards and bedsteads in the Newmanry at Bletchley) were all pretty much built expressly to solve practical and immediate problems (breaking ciphers, calculating parameters for nuclear weapons) that would take too long or contain too many errors if done by hand. Babbage got his freak on because of error rates in (IIRC) log tables. My Mesopotamian history is a little rusty, but it's easy to imagine a similar development story for the abacus. (In the sense of 'tool to surpass unaugmented human in speed and accuracy of computation' rather than 'hacking crypto and nuke yields'). But I dunno. Maybe.

[flukus]

Weren't collosus and ENIAC extremely error prone? I think ENIAC installed transistors as banks so they could easily replace them and this was required just to keep the machine somewhat functioning. I wonder how they compared to humans when factors like that are included?