Raw LLM Responses

In 2014 I was in a paid study to participate in "virtual human" therapy sessions. Using that search term it appears there were a lot of studies done around that time. The premise was that they had some super-advanced computer shrink and I was scheduled to do three Skype video calls. Before each scheduled call I was given a list of talking points, to which I prepared notes. And after each call I filled out a survey about how I felt during the experience, primarily my comfort with the experience and my attitudes towards the avatar. Popping up the call I was greeted by a fairly crude 3D model of a bust of a man with a backdrop, and there were some minor movements so it wasn't a static image. It asked me about the talking points and had some generic follow-ups like how I felt about something I said. Each time I spoke for an extended duration or whenever the avatar transitioned from listening to speaking the screen would jump, so I soon suspected it was prerendered recordings. One of the last two sessions, I didn't get the call in time and I got a message via Skype. Something like "oops technical trouble" "starting soon" and it broke the veil that there wasn't a man behind the curtain. After the last session and survey they spilled the beans in the closing summary that the synopsis of the study lied and technology wasn't at the point where virtual therapists were possible, but that they were gauging the viability and attitudes of individuals towards it by simulating it. So it was just a guy on Skype using the video equivalent of a soundboard. Never really thought about it, but now that we are at that point it's interesting that research was being done that far back. Around that time Cleverbot was probably the most popular/advanced chat AI. For my part, the first session was extremely awkward and I more or less read directly off my notes, but by the third I barely glanced at my notes and actually ran out of time. It didn't really provide any insight, but if nothing e