Is there a specific length of time you are looking for? Why do you say 17 years? 1930-1942 is significant in GISTEMP to 2sigma. 1930-1941 almost makes it too. 1976-1995 is statistically significant to 2sigma in GISTEMP, as well as 1976-1994 and 1976-1993 and 1977-1992, 1976-1991, and 1976-1990. 75-89. 74-88. 68-87.

The problem is that global warming 'started' in the late 70s.

Indeed this is another demonstration of why I don't like this trend calculation. Colder temperatures in 1978 and 1979 increase the likelihood of warming from 1976 or 1977.

It was a cool period in the 1970s so I don't see how you can get more than 20 years in there but that is why alarmists suggest that aerosols caused the cooling in the 1970s so that it doesn't destroy their argument.

Not sure what you are asking, but anyway we significant warming in the 1930's and 40's and significant cooling in the 70's and I doubt that CO2 had anything to do with either case.

Since about 10,000 BC when the ice age ended.

I'm not sure, but my ballpark guess is... at most, 2 or 3.