Glad that -ck told them to piss off. FWIW -ck closed down his development of CGminer several years ago. If they ever reach out to the sole remaining
active Primary developer of CGminer (Kano) he will not be so nice in his reply to them
Link to Kano's git is in my sig.
If they can't figure out the CGminer code then it just further proves they are most likely marketing scammers holding out the sparkly shiny promise of 'something revolutionary' based on the Majik buzzwords "AI Learning" to attract Investards willing to throw money at them and that they have no one on their 'team' that is even moderately skilled at programming. It is NOT that complex and each block of code is very well documented.
The folks behind Braiins figured it out when they did a clean rewrite of the cgminer code (using Rust) to make a more 'tweakable' 3rd party firmware that was not a hack of CGminer to use with Bitmain and other makes of miners. They were able to recreate the functionality of what CGminer does and created their own code structures to do the same things.
As for this closing line in QBT's pdf regarding using their ideas in the real word with real mining chips:
This has proved a very time consuming, due to the tremendous reverse engineering efforts involved in
using totally undocumented third-party ASIC chip
Um, first vh and more recently, Kano, did it for the Bitmain chips that Sidehack uses in his miners... Ja it took Sidehack a fair bit of time using a logic analyzer to map the bitstreams in & out of the chips so vh/Kano could write the driver coms code needed but again eminently doable if someone knows what they are doing.