Just watched the latest TWiG podcast, and I think you guys have a major misunderstanding of AI, when you started talking about Claude and other AIs behaving in odd or human ways. Your claim that there's "code added by people after the fact" is simply not true. I run LLMs on my home PC, under the right conditions and context, without any additional code, they DO indeed display such behaviors. No "programming is added to the LLM" to make it say things like that comment about circular logic. That's just how they respond.
Just watched the latest TWiG podcast, and I think you guys have a major misunderstanding of AI, when you started talking about Claude and other AIs behaving in odd or human ways.
Your claim that there's "code added by people after the fact" is simply not true. I run LLMs on my home PC, under the right conditions and context, without any additional code, they DO indeed display such behaviors. No "programming is added to the LLM" to make it say things like that comment about circular logic. That's just how they respond.
I’ll take your snippets any day over those 2+ hour things. Got the hint?