Nolan shaped ChatGPT in action. He understands that the reaction he received to his message means that someone thinks he fucked up.
His rote programming tells him that an apology is expected, so he searches his database for what words constitute an "Apology" and then strings them together based on past experiences.
At no point does he actually understand that he did fuck up. Nolan did nothing wrong. In much the same way, he doesn't understand why an apology is needed or even what an apology is. But his rote programming tells him that the string of symbols forming larger symbols called "Words" which together form what is called an "Apology" allows him to continue interactions with "Humans" for a while longer, so he "Apologises."
Naturally, his lack of understanding means that he'll inevitably make exactly the same mistake and try to fix it in exactly the same way.
Nolan should start streaming and collab with Neuro.
They're the same.