Re: [FRIAM] ChatGPT and William James

Lisp or Haskell macros.. Sent from my iPhone On Mar 6, 2023, at 8:22 PM, Russ Abbott wrote:  Let's consider the word "input" again. The implication is that there is an "agent" of some sort that is separated/distinguishable from some "environment" from which it gets "input." The question (or

Re: [FRIAM] ChatGPT and William James

Let's consider the word "input" again. The implication is that there is an "agent" of some sort that is separated/distinguishable from some "environment" from which it gets "input." The question (or at least one question) concerns our specification of what that "agent" is. If, as Glen suggested, ge

Re: [FRIAM] ChatGPT and William James

From examples, either adversarial generative learning or stable diffusion can learn the laws of physics. https://github.com/lucidrains/video-diffusion-pytorch Also it is common in training these systems to have a "foundation" model that is then specialized with domain-specific context. The weig

Re: [FRIAM] ChatGPT and William James

Well put. When Frank emphasized "data", he doubled-down on the ambiguity. The fact is, those who claim a human is categorically different from a machine have no legs on which to stand. Every single boundary between them is broken, year after year. On 3/6/23 15:47, Russ Abbott wrote: Are the la

Re: [FRIAM] ChatGPT and William James

Are the laws of physics "input?" Is the existence of the universe "input?" If so, what issues are we arguing about? -- Russ Abbott Professor Emeritus, Computer Science California State University, Los Angeles On Mon, Mar 6, 2023 at 3:42 PM glen wrote: > Well, again, it seems like we're equivoc

Re: [FRIAM] ChatGPT and William James

Well, again, it seems like we're equivocating on "input". Are the genes the baby inherited from its parents "input"? I'd say, yes. On 3/6/23 15:36, Russ Abbott wrote: Hard to see how you could simulate an infant on the basis of input it's received. It cries; it smiles; it pees; it poops; it pu

Re: [FRIAM] ChatGPT and William James

Hard to see how you could simulate an infant on the basis of input it's received. It cries; it smiles; it pees; it poops; it pumps blood; it breathes, etc. There are many experiments in which one concludes things about what's going on in an infant's brain by how long it looks at something. -- Russ

Re: [FRIAM] ChatGPT and William James

I'm confused by the emphasis on "data". While I'm tempted to agree with my simulation of Frank and say that a human's output is not based solely on statistical patterns in the input the human's been trained on, to dissemble on the meaning of "data" or "input" or "statistical patterns" is a bridg

[FRIAM] [off topic] exterminators for mice?

Looking for a dude in the santa fe area that could help find hiding spots, holes and blah blah that mice can use to get into the house. -. --- - / ...- .- .-.. .. -.. / -- --- .-. ... . / -.-. --- -.. . FRIAM Applied Complexity Group listserv Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-

Re: [FRIAM] ChatGPT and William James

If you live with a baby you see that they have knowledge that can't be based on "data". --- Frank C. Wimberly 140 Calle Ojo Feliz, Santa Fe, NM 87505 505 670-9918 Santa Fe, NM On Mon, Mar 6, 2023, 2:50 PM Marcus Daniels wrote: > How? > > > > *From:* Friam *On Behalf Of *Frank Wimberly > *Sent

Re: [FRIAM] ChatGPT and William James

How? From: Friam On Behalf Of Frank Wimberly Sent: Monday, March 6, 2023 12:50 PM To: The Friday Morning Applied Complexity Coffee Group Subject: Re: [FRIAM] ChatGPT and William James >And we humans are different? In a word, yes. --- Frank C. Wimberly 140 Calle Ojo Feliz, Santa Fe, NM 87505 5

Re: [FRIAM] ChatGPT and William James

>And we humans are different? In a word, yes. --- Frank C. Wimberly 140 Calle Ojo Feliz, Santa Fe, NM 87505 505 670-9918 Santa Fe, NM On Mon, Mar 6, 2023, 12:14 PM Nicholas Thompson wrote: > *However, it's important to remember that there are also important > differences between a large langu

Re: [FRIAM] ChatGPT and William James

However, it's important to remember that there are also important differences between a large language model and human consciousness. While a large language model can generate text that may seem to flow like a stream of consciousness, it does not have the same kind of subjective experience that hum

Re: [FRIAM] ChatGPT and William James

Interesting. EricS' layout triggered me. I've used the word "registration" a lot, mostly because of BC Smith's re-terming from "inscription error" to "pre-emptive registration". But I'd never actually looked at the etymology of "register" . From Eric