screwlisp
Argh, I reread Breaking The Complexity Barrier (Again) 1973 and Winograd's still completely right.
It would be nice if:
- we could just run "learning" on a bunch of data, and it would figure everything out without us telling it
- we could just run "crunching" on a bunch of symbols, and it would figure everything out without etc.
but neither of those have an antecedent for being a way things work.
It makes me think I actually have to do the work of using Sandewall's Leonardo System =_=
Alfred M. Szmidt
screwlisp
@dougmerritt Understand Natural Language '72, right?
screwlisp
@amszmidt @larsbrinkhoff
You two don't happen to be aware of an intact source for historical SHRDLU? The 90s clisp port, "SHRDLU after suffering a stroke" sounds kind of unappealing.
DougMerritt (log😅 = 💧log😄)
@screwtape
If you read Winograd's SHRDLU book, it all makes sense, but of course it only captures a fraction of the programming, which was a tour de force.
Someone resurrected it once, but not in a form that I personally was able to use. Not sure about these days.
screwlisp
LLMs are in the first class of failures. Transformers - using a feedforward network with attention on terabytes and terabytes of scraped data - is an attempt to get something without contributing a boundary-pushing program (having been allowed by world governments to steal data without recompense, they got it for free).
I actually do consider LLMs to be spooky, even beyond making public vast quantities of private historical emails.
But this is a legal innovation, not a human-authored one.
screwlisp
I can't say I'm happy about it but this implies that currently freecad and linux are humanity's premier contributions to the universe.
Edit:
I guess we can be happy that emacs is also up there. Compilers and languages generally maybe. I think McCLIM should be regarded as an important ongoing success. It's just that rather than McCLIM and emacs being exceptional, anyone belonging to the computer-fanciers-association should be part of contributing something like Sandewall's Leonardo System.
screwlisp
Terry's point is that rather than being a "dead end" as people would blithely tell him, SHRDLU is actually /exactly/ the kind of computer program that is meaningful to write.
Its behaviour is both deliberate and mindbogglingly complex (barely fitting in Winograd's head). His argument is that each computer program should be as intricate and expansive a contribution to the world as possible at the periphery of the author's ability.
Not get-rich-quick-scheme style learning and crunching.
Alfred M. Szmidt
@screwtape And source code is here: https://github.com/PDP-10/its/tree/master/src/shrdlu
GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.
All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.