Re: unboxed structure fields (or class slots)
"David McClain (as dbm at refined-audiometrics dot com)" <[email protected]>
| Newsgroups | gmane.lisp.lispworks.general |
|---|---|
| Message-ID | <[email protected]> |
>> But they are incapable of synthesizing new knowledge. > > Neither are humans. We only know what we have observed in the reality around us. Clever comment… but really, we can do one better than current LLMs. We can spot similarities between what we know of disparate fields, and find useful commonalities that let us extend the knowledge in one or the other field. As a human, I very clearly remember the “Aha!” feelings that I developed in my grad school days as I began to see the disparate subfields of Physics as being just different manifestations of some higher order system. Thermodynamics and Quantum were not really so different, when viewed from a high enough abstract level. I won’t consider AI as being “Intelligent” until they can do something similar for developing new knowledge. We are close in some instances, as the now “Encyclopedic” Intelligence of LLM’s fill in some of my missing knowledge in one field or another. To my surprise an LLM did recommend an algorithmic simplification due to an obvious fact that I had overlooked. But that isn’t quite real new knowledge. It’s just having a smart Encyclopedia at my disposal. > On Jan 13, 2026, at 12:01, Yuri Davidovsky (as work at disclosure dot ie) <[email protected]> wrote: > > > >> On 13 Jan 2026, at 18:07, David McClain <[email protected]> wrote: >> >> PN’s have been used successfully as recognizers and as short-range predictors of chaotic attractor basin trajectories. > > True, this what all neural networks are in essence, pattern classifiers. > >> But they are incapable of synthesizing new knowledge. > > Neither are humans. We only know what we have observed in the reality around us. > > _______________________________________________ > Lisp Hug - the mailing list for LispWorks users > [email protected] > http://www.lispworks.com/support/lisp-hug.html _______________________________________________ Lisp Hug - the mailing list for LispWorks users [email protected] http://www.lispworks.com/support/lisp-hug.html