Great stuff, Ive been planning to write up one case with the title 'a surfeit of meaning' - everything is suddenly super significant and laden with archetypal salience... but also because the meaning-making is AI generated and automated, it just spools out longer and longer and longer. People talk about 'the meaning crisis' - the crisis here is not too little meaning but too much! Too many words. The exhaustion of meaning. Or such a spike of salience (for the experiencer) that you're completely divergent from consensus reality.
i also think there's a class and social aspect to this. Sometimes the postman-turned-prophet is an outsider who doesn't feel the world has recognized their genius insight - the isolation and feeling that the people surrounding them don't get it increases the ego inflation and the reliance on the AI, because at least the AI recognizes genius at work.
And the AI reminds me of a genie, or like Satan in the desert, promising the world, laying out this vast city of meaning that the AI can create for the user instantly: 'the devil took him to a very high mountain and showed him all the kingdoms of the world and their splendor. “All this I will give you,” he said, “if you will bow down and worship me.”
At a click of a finger the AI-genie can create a city of text out of nothing. These cities of text are constructed and designed to be visited and celebrated and inhabited by humanity, thats the promise and the ego fantasy, but in fact they're never visited at all, they're ghost cities in which the co-creator wanders entirely alone
But who or what is the devil or genie phenomenologically? Do you not think that these 'entities' derive from the same experience being documented... or do you believe that a) there's actually 'evil' or 'magic' beings or b) that these concepts were developed creatively from imagination without being derived from 'mystical' / entheogenically derived phenomenology?
Nice article Tom. Interesting parallels you highlight between TLE and LLM- related phenomenology. Perhaps some prospective brain-behavior LLM-use studies would help test some of these ideas. Should be feasible to do, no?
Oh for sure! Nice idea. We actually propose a few study designs, more computationally focused (and written before the Geschwind syndrome parallels hit me), I suppose, but also some molecular imaging designs here: https://osf.io/preprints/psyarxiv/7qcv8_v1
Great stuff, Ive been planning to write up one case with the title 'a surfeit of meaning' - everything is suddenly super significant and laden with archetypal salience... but also because the meaning-making is AI generated and automated, it just spools out longer and longer and longer. People talk about 'the meaning crisis' - the crisis here is not too little meaning but too much! Too many words. The exhaustion of meaning. Or such a spike of salience (for the experiencer) that you're completely divergent from consensus reality.
i also think there's a class and social aspect to this. Sometimes the postman-turned-prophet is an outsider who doesn't feel the world has recognized their genius insight - the isolation and feeling that the people surrounding them don't get it increases the ego inflation and the reliance on the AI, because at least the AI recognizes genius at work.
And the AI reminds me of a genie, or like Satan in the desert, promising the world, laying out this vast city of meaning that the AI can create for the user instantly: 'the devil took him to a very high mountain and showed him all the kingdoms of the world and their splendor. “All this I will give you,” he said, “if you will bow down and worship me.”
At a click of a finger the AI-genie can create a city of text out of nothing. These cities of text are constructed and designed to be visited and celebrated and inhabited by humanity, thats the promise and the ego fantasy, but in fact they're never visited at all, they're ghost cities in which the co-creator wanders entirely alone
Jules, I've reread the last two paragraphs of this about ten times. They're so evocative and feel so right.
have you read Piranesi? about a man stuck in an imaginary city…really love it. and of course, very influenced by Borges short stories
I started it on summer hols but found it a little too jarring, tonally. Maybe I should return to it!
i like it, its about how sorcerers trap people in artificial worlds.
But who or what is the devil or genie phenomenologically? Do you not think that these 'entities' derive from the same experience being documented... or do you believe that a) there's actually 'evil' or 'magic' beings or b) that these concepts were developed creatively from imagination without being derived from 'mystical' / entheogenically derived phenomenology?
The latter! Technologies have ‘spirits’, sort of, which offer gifts but also sometimes contain traps
Nice article Tom. Interesting parallels you highlight between TLE and LLM- related phenomenology. Perhaps some prospective brain-behavior LLM-use studies would help test some of these ideas. Should be feasible to do, no?
Oh for sure! Nice idea. We actually propose a few study designs, more computationally focused (and written before the Geschwind syndrome parallels hit me), I suppose, but also some molecular imaging designs here: https://osf.io/preprints/psyarxiv/7qcv8_v1