GitHub releasing the worlds most awful report trying with all their might to make copilot look good.

- percentages did not add up to 100 (one was under, one was over)
- errors not defined as functional errors, but “any code that reduced the ability for the code to be easily understood” (what in the subjective bullfuck is that?)
- apparently 200 participants but 25 represented 40% of the population????

LLMs are already decreasing the average GitHub employee IQ to functionally retarded levels :(