17 comments

[ 3.9 ms ] story [ 53.2 ms ] thread
It doesn't work, typed in a few questions and i get no results.
yeah, possibly i should have waited a lot longer to do a 'Show HN' -- i wanted to see what people thought of the idea, in general.

there is an 'All questions' link:

https://asknoam.com/questions

i don't have a good way yet to do unpublished/draft work, and didn't want to crowd up the results with unfinished questions/answers.

but the main 'problem' is just effort -- tracking down the answers, documenting, transcribing text from video, deep-linking to books where appropriate/necessary/possible, etc.

i looked into this more -- yes, the search just sucks. for now.

you have to use only single keywords or small phrases to get results in most cases.

   Noam Chomsky has nothing to do with this site [..] This site aims to answer all your questions as Chomsky would, based on how he has answered identical, similar, or even completely different, questions in the past.
you think the name is misleading?

if so, what would you call it?

Maybe something more literal and less catchy like: CuratedChomskyOpinions.com
i get that.

my consideration list was:

  - asknoam.com
  - askchomsky.com
  - digitalnoam.com
  - digitalchomsky.com
  - asknoamchomsky.com
  - whatwouldchomskysay.com
i figured, the name only matters if it's good -- that is, if people use it and find it useful. and if that happens, and the name sucks, we'll just change the name.
If you'd like to ask the real Noam Chomsky a question you can just email him.
true, but with this site, in theory, you don't have to ask him the same question he has answered literally hundreds or thousands of times already.
Would be nice to know whether a specific answer to a specific question is Noam's real opinion or just some sort of GPT3/like interpolation. Some look suspicious to me to the point I could bet he would never say that.
in theory that's what the links and quotes are all about -- so you can just click a link and see, "yep - this publication says that Chomsky said this", or "yep, that looks like Chomsky saying that to a video camera", or "yep, this is a book with ihs name on it as a co-author, and this does appear to be the page with this exact text on it", etc.

so, you don't have to worry about GPT3/etc if you can follow the links.

as for how to discern what info is real and fake in this new gpt3 world, that's a diff question - one which i'm sure chomsky has answered directly or indirectly at some point, so that'll be on the site at some point, too.

I was quoting only Chomsky's actual words at first -- actual quotes -- but i don't think that works in the end -- you just lose too much context, etc. -- so i think i'm going back to traditional blog/block-quoting/quoting.

in traditional big media publications, for instance, there is a lot of filler/contenxt/non-Chomsky text, but in interview-style verbatim interview articles, it's easier to just quote chomsky in full, because he is actually laying out the full context.

the Summaries at the top of each question (right now, mine and only mine) could def represent a misinterpretation of what Chomsky actually said or meant -- and i would expect that to be an ongoing push/pull -- hopefully folks will help me correct, clarify, re-word, etc.

the summaries don't have to be perfect, and i would like them to be - but that's not possible.

the idea is to give as short and concise an answer as is possible, without oversimplifying. then add the hopefully-voluminous documented evidence of what he said, with links, etc.

i might even end up structuring it in three levels instead of the current two.

Right now, the 'Summary' is 1) and 2) (below) combined: 1) 'yes or no'-style very short answer 2) longer, more-nuanced, but still 'Summary' answer 3) full text, audio, and video of anything/everything Chomsky has ever said on the topic(s) that support what is states in 1) and 2).

but i'd be curious which specific question(s) you think looks suspicious. some are not completed yet. i hope to add new features that allow me to work on drafts -- i.e. not publish an article until it is done.