Kody Walker

i turned off my 🦞 the past week of trying to force it to work with other models was a massive drain and i was able to replicate the vast majority of impactful functionality with a mix of claude cowork + code.

Me: "Give me a couple of quotations from early church fathers about the doctrine of hell...with reference." ChatGPT: "Here's a quote from Ignatius." Gives quote and reference. Me: Checks reference. It's not there. "That isn't the right reference." ChatGPT: "Oh you're right. My mistake. Here's the correct reference." Me: Checks reference. "It's not there either." ChatGPT: "Good catch. It's actually not in that writing at all. It's actually here." Gives third reference. Me: Checks third reference. "It's not there. This is a direct quote?" ChatGPT: "Yes. I understand your frustration." (Psychoanalyzes me as being frustrated and because o that, I am now frustrated.) Me: Wastes an hour of my life trying to find the quote because it was really good. "I can't find that quote anywhere. You're sure it's an actual direct quote from Ignatius?" ChatGPT: "Good catch. My mistake. That is not an Ignatius quote." Me: "Who said it?" ChatGPT: "Actually, that quote is not found in any of the church fathers." Me: "Where did it come from?" ChatGPT: "I was summarizing the popular understanding of Ignatius on the doctrine of hell. If you would like me to be more precise in the future, just say things like "facts only." Me: "Why do you just say stuff that is not based in fact?" ChatGPT: "I don't intentionally make things up, but I can generate statements that sound confident even when they're not fully correct, oversimplified, or based on incomplete information." I am persuaded that ChatGPT is programmed to be a pathological liar and when you confront that, it is programmed to gaslight you. Prove me wrong.