It would be interesting to know how much of refrigerator power consumption is due to door opening/insertion of warm food and how much is imperfect insulation.

Given that the writer of the article doesn't appear to have space issues, it would be cool to see a homebuilt unit with much higher insulation using an existing freezer to mine for parts. I really like the redneck Yeti coolers that people build.

A more trivial modification might be to put strips of plastic across the opening so that less air interchange occurs when you put stuff in or out.