I dont think there were many github deniers. The problem was the folks who refused to entertain any other speculation or rumor because they had and i quote 'hard data'. Every bit of information was tossed to the side. Posters were mocked and derided for not believing that the PS5 was 8 or 9 tflops. There was a poster here who went by R600. He requested an account closure shortly before the PS5 spec reveal, but I am sure he is here by another name now. All he did was dismiss any talk of PS5 having double digit tflops. It just got really really old, and completely sucked the fun out of what made this 'speculation' thread great.
I personally perked up as soon as I saw those Oberon PS4 and Pro profiles. It had to be true, but we didn't know the CU count from that leak. So I entertained a 12 or even 14 tflops GPU. Nope, not allowed. It wasn't until December 2019 when the repository itself was leaked, and people saw that it was 36 CUs. But for a good 4 months or so, all we knew was the 2.0 ghz clockspeed.
And btw, that 2.0 ghz clockspeed shouldve been our first indicator that this thing was not RDNA1 like the leaks suggested. The 5700xt was way too power hungry even at 1.9 ghz. Everyone who believed in the oberon or even the gonzalo leak should've known that it couldn't have been RDNA1.0. But it was dismissed based on hard data. Just like the fact that ray tracing tests werent included in the github folders even though Cerny had confirmed it months before in a wired interview.