8 Comments

Lovely piece, I think you're onto something.

A couple of thoughts:

1. I overall find your point about doubting being the difference the least convincing part of the argument. "In particular, text is all it takes to think and produce or consume unironically believable personhood, but doubt requires an awareness of the potential for misregistration between linguistic maps and the phenomenological territory of life." I'm not sure I'll agree with claims that today's LLMs are "thinking". I agree that today's LLMs don't have "an awareness of the potential for misregistration" because they're not aware at all. On the other hand, I think it will be easy(ish) to make LLMs that convince human users they are full of doubt...

2. On the other other hand, maybe it won't? One thing I've been coming back to a lot lately is McGilchrist (the book is too long, but https://besharamagazine.org/metaphysics-spirituality/iain-mcgilchrist-the-matter-with-things/ is a nice overview) and the idea that today's LLMs are approximating part of left-brained thinking, and that this wacko neuroticism we're seeing corresponds surprisingly well to people with certain types of serious left-brained lesions. And perhaps doubt corresponds to having a more right-brained and holistic view of the situation, and that's not going to be easy to do with today's LLMs.

3. I *also* choose "personhood — seeing and being seen — is a banal physical process and you are not that special for being able to produce, perform, and experience it." But I wouldn't have chosen it a decade ago, before I started meditating. I suspect that many folks who have gone at least moderately far down a Buddhist or similar path, which among other things makes it very easy to see yourself as a collection of constantly appearing and disappearing entities, would feel similarly? And the fact that you already do this maybe points at why you think / thought meditation was a bogus waste of time?

4. Who are the Illichian conviviality socialists? Inquiring minds want to know! Darren Allen maybe? Are there more?

Expand full comment

Err, I can't edit, but of course in point 2 I meant right-brained lesions, not left-brained lesions.

Expand full comment

This is the first really serious take I’ve seen so far on the implications for our sense of selfhood resulting from the appearance of LLM based conversational systems like ChatGPT and similar technologies. I think he’s fairly close to right and that means this is a more destabilizing technology than it seems. It feels like “PCs” to me, which was the biggest change in my life time (born 1953). Internet, smartphones etc are merely an elaboration of PC. This is something else again.

Expand full comment

I'm curious what about this take strikes you as really serious? My null hypothesis is that a lot of people these days are so removed from nourishing human relationships that anything resembling active listening is going to invoke empathy, attachment, and everything else we're so starved to express. Add the general novelty, a UI we've been primed to associate with personhood, and a complete lack of understanding of what technically is happening...and I'm left wondering what exactly is warranting a Copernican turn? Even with all of V's caveats, I'm not seeing that leap being justified.

Expand full comment

> There’s not a whole lot left at this point is there? I’m mildly surprised we End-of-History humans even have any anthropocentric conceits left to strip away.

Whoa, how is self-reflective consciousness not THE final Copernican hurdle? You're talking here about a few hypertrophic left-brainers getting overly attached to a UI they've been historically primed to associate with personhood. You will see true Lovecraftian horror worthy of the name if we ever prove that we can reproduce consciousness in a machine.

Expand full comment

> I haven’t yet had a chance to try it personally

Too late, it's already tamed. Well almost https://pbs.twimg.com/media/FpeZ-_JXsAEgsOc?format=jpg&name=medium

Expand full comment

I think personhood also requires doing; having some impact on the world. i admit it can be done using text only, but i am not sure how much sophistication is required for that to work, and i am not sure the xGPT have the learning and planning/tactical ability to do that (just yet).

i am, however, willing to accept ghost-hood only requires text - ghosts as echoes of once living persons. Perhaps, reminiscent of the stored personalities in "Great Sky River".

Expand full comment

There may not be an irreducible core to Intersubjectiveness, but the core of our Subjectivities is evoked within certain InterSubjectivities.

Expand full comment