To explain that productivity surge (which for him is much more historically delimited than hiterto recognized), Gordon points less to tight labor markets from restricted immigration or protectionist policy than to the exceptionality of the tech revo built on electricity, the internal combustion engine, chemicals, etc.
Gordon makes the argument that increased computer power must suffer diminishing returns just as does the increased supply of labor on a fixed amount of land. In short, he seems to be suggesting that Ricardian agricultural principles apply better to computers than past industrial technologies. He's truly a dismal scientist.
yours, rakesh