A well-written conversation with a chatbot

A while ago, I listened to all 3½ hours of Hinternet Production Labs — An Audio Launch Event! by Justin Smith-Ruiu. It is pretty cool. I’m glad I listened to it. I was working on a jigsaw puzzle the whole time.

Listening to this feels like having a Wikipedia binge that leads from Yakut to rock music, the etymology of the epithet “Willard”, the application of the definition of said epithet to bands after a lengthy discussion achieves consensus, an immediately ensuing discussion debating to which musical acts one could realistically justify applying the epithet “peach”—the illusion fell down a bit here, as it devolved a bit too much into the typical flailing back-and-forth of eliciting information from an LLM—the same for “coolness”—same flailing—”Axolotl”…

…to Justin himself—“a hallmark of a deeply thoughtful mind,” … “a provocative and deeply introspective scholar,” … “intellectual versatility,” … “insightful critiques,” …

…to a truly gobsmacking and overwhelming number of detailed suggestions about how to keep people from assuming that you named your donkey “Pippin” because of the Lord of the Rings, whose length nearly exceeded my patience but also left me unsure as to whether the garrulous descriptions were from a tireless AI or an at-least slightly obsessive-compulsive philosophical researcher. The twist at the end where the interlocutor changes his mind after the apparently tremendous amount of work put in by the LLM was both funny and a reminder that LLMs are machines … and also that Justin is going to be one of the first ones up against the wall during the robot wars.

The final part is more obviously AI research, delving into the meta-topic of asking an LLM about Pascal’s wager (which is pronounces wah-jah 😂 ), Roko’s Basilisk, Searle’s Chinese Room, and, finally, the deliberate subterfuge of having LLMs formulate responses in the first-person—“neither the cake nor I possess the internal conditions, such as consciousness, intentionality, or self-awareness that would truly make us an ‘I’ in the philosophical sense.” And yet, as usual, the lure of lucre trumps foresight.

… all read to you in mellifluous tones that sometimes present as Chatbot to-and-fros and sometimes more like reading from Justin’s essays. Thanks for this. I personally don’t have the patience, time, or inclination to spend this much time with an LLM, but I found this curated and linked series of sessions to be a fun accompaniment to my Christmas jigsaw puzzle.

Contra current trends, I actually listened to this and wrote the summary myself, instead of having a machine do it. It might be amusing to see what a machine would write, perhaps illuminating the mediocrity of my summary, but … I’m wedded to doing it this way. I’ve got time to kill anyway; what’s the point of hurrying through everything?