because per corps statements, AI are now top 0.1% of PhD in math, coding, physics, law, medicine etc, yet, when I try it myself for my work it makes stupid mistakes, so I have suspicion that corp very pushy on manipulating metrics/benchmarks.