The main problem (and the reason I have no intention of doing any research involving AI programming), is even if you could get the research started, completed,

The main problem (and the reason I have no intention of doing any research involving AI programming), is even if you could get the research started, completed, and published in 4 months (and incredibly ambitious goal), 95% of the comments would just be "Oh, but you didn't consider FishGPT 5.2, and Pineapple AI, which came out 6 weeks ago, so any negative points are entirely out of date".