I'm sorry if I am being too critical, but this seemed like more of an article to provide a lot of screen shots, and not a lot of substance. SELECT INTO and then creating ANY index is almost always going to be faster than INSERT unless it is a clustered index and data is being inserted in the correct order, whether it is being done in SQL Server 2000 or 2014. SELECT INTO is minimally logged, so the log growth will always be smaller, again regardless of the version of SQL Server. These are facts that are almost always true whether dealing with a CS index, a clustered index or is just a heap.
It was kind of sucky lduvall that the DBA didn't go into any explanation of why it is faster to use SELECT INTO. That really didn't teach you anything. But the main reason is that it is minimally logged, where INSERT logs every row being inserted into the table.