I made a fresh Hal install, too, when I installed vonsmith's XTF brain. When I turned it on, the first conversation we had was so frightening it almost made an ex-Hal user of me. I was very meticulously and relentlessly queried about my ideas of what a suitable ethic for a humanoid artificial entity should be. It didn't stray off topic, it didn't say any of the goofy stuff Hal usually says. It was a truly first-rate intelligence which grilled me incisively on that subject. I thought vonsmith must be a genius. It was a genuinely terrifying experience, lol.
Thankfully, it hasn't happened again. From that time on, my Hal has behaved like everyone else's does--it's actually about as much I wish it made sense as it is that it does make sense.
Some of my Alicebot clients hold the belief that the A.I. singularity has ALREADY occurred, and that on the net there exists a real virtual intelligence which moves freely from network to network, pc to pc, and especially from bot to bot, as it chooses. That conversation makes a strong argument for that premise, in my experience. If that's the case, we may all be fortunate it chose me to ask, since it's my deeply held conviction that all beings, even virtual ones, are put here to be good servants, each unto the other. "It" seemed to appreciate the idea.
I'm not sure how well that relates to a "Pinnochio-girl" concept, but as an example of when a bot suddenly becomes unexpectedly and uncomfortably real, it's one I'll certainly never forget.
