Thanks for sharing that historical tidbit! My understanding is compressor manufacturers don't like to work on co-generation on a small scale. However, early attempts like you describe tended to use some kind of water jacket because heat pump technology back then was finicky and much more expensive than today. Using air to conduct heat is hugely inefficient, but we're dumping the waste heat into the air anyways and then later spending more kWh to cool down the air, so the current batch of heat pump-based water heaters are a possible starting point.

In idle moments I've wondered if I could retrofit the multiple small compressors with a single large unit with multiplexed hoses and electronically-controlled relays that was designed to operate submerged in something like polyalkalene glycol (PAG fluid), insulate the heck out of that to minimize thermal conductivity to the surrounding air, then extract out the heat of the PAG fluid into the water heater. Would be a maintenance nightmare to be sure, but it would be a cool hack to play with exploring and measuring those boundaries of efficiency.

Its also possible that labor/maintenance costs outweighed the cost savings. Or they couldnt buy the equipment any more :)