Oddly enough the greats like Turing, Flowers, and Codd didn't work in isolation or in garages and basements.

Turing was socially unskilled but worked as part of a pretty large team when he was employed by SCCS (1939 to 45) and in close collaboration with large teams at Dollis Hill and at Letchworth at he same time. He packed enough punch when 29 years old to get Churchill to appoint Travis deputy director of SCCS because he reckoned Deniston wasn't up to the job; a couple of years later he worked successfully with US Navy people and enabled them to reduce the number of Bombes they would need for naval decryption by a factor of 4, as well as pointing out tht not allowing for various mechanical effects in the machines they were attacking would have rendered their machines useless. His social ineptness ensured that he didn't wrap that up politely; but his US associates thought either that it was probably a good thing to arrange that fewer US ships would be sunk through lack of intelligence or (more likely) that it would be politically dangerous to make too much fuss so didn't complain very much.

Flowers was so socially unskilled that he ended up spending his own savings on the evelopment of Colossus and only got a tiny fraction of it back in compenation after the end of the war; but as he and his team at Dollis Hill was delivering one Mark 2 Colossus per month from June 1944 until the Japanese surrendered, he obviously wasn't a one man band or operating in a garage or basement.

Codd was sufficiently socially inept to upset his employer (IBM) thoroughly by telling their customers about his new Relational Model for databases (which IBM had refused to develop because it was too new and different and any way would divert revenue from their star DB product, IMD/DB), producing a customer reaction that forced IBM to develop it; this generated sufficient pique that Codd was effectively cut off from the development, and that resulted in Sytem R having the rather messy SQL language instead using Codd's Alpha (Codd's exclusion was porbably one of the reasons that Larry Ellison beat them to the market with the first commercial RDBMS). But he worked with quite a few people when he was at IBM research, San Jose (not exactly a garage or basement) including some other rather big names (Chis Date, Ron Fagin) and some of teh System R team (Ray Boyce obviously - they would hardly have a normal form named after the pair of them is they hadn't worked to gthere, would they) as well as with colleagues in distant parts of IBM (eg Heath at IBM Hursley).

I think 3 counter-examples is enough, don't you?