Yeah, it appears that if your prior is that AI can't be good, no evidence can convince you that actually current models are extremely capable.