It also provides data showing that speed and correctness are essentially uncorrelated between different people - i.e. just because two different people finish at different speeds, does not mean the faster person has a less correct solution, at least in this specific setting. If you assume that's true generally, that makes the idea that there is a large gap between F500 programmers and FAANG ones even more credible, since the best programmers both produce better solutions and produce them faster.