Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: Mike Spencer <mds@bogus.nodomain.nowhere>
Newsgroups: misc.news.internet.discuss
Subject: Re: Chrome browser will help you shop with AI
Date: 14 Aug 2024 19:13:45 -0300
Organization: Bridgewater Institute for Advanced Study - Blacksmith Shop
Lines: 22
Sender: mds@enoch.nodomain.nowhere
Message-ID: <871q2qeorq.fsf@enoch.nodomain.nowhere>
References: <66b5e75c$3$1439832$882e4bbb@reader.netnews.com> <1k7hok-ii5.ln1@pyrite.ereborbbs.duckdns.org> <v9dhlt$3def4$1@dont-email.me> <66bc8ee0$2$16$882e4bbb@reader.netnews.com>
Injection-Date: Thu, 15 Aug 2024 00:13:46 +0200 (CEST)
Injection-Info: dont-email.me; posting-host="88caf2c59cb2ecf01b1253444d74bd2e";
	logging-data="666996"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1+Fp08S5YUjnlAXru9V3a+59+IRlGN3n9o="
Cancel-Lock: sha1:YHaiXlgfSrP8Pgmg8nhgMoOIKxg=
X-Newsreader: Gnus v5.7/Emacs 20.7
X-Clacks-Overhead: 4GH GNU Terry Pratchett
Bytes: 2014


Retrograde <fungus@amongus.com.invalid> writes:

> How about an algorithm that takes a webpage that has been SEO-ified and
> enshittified beyond recognition and reduces it to the actual
> information?  Example: get on a search engine and look for the recipe to
> make a martini.  Your average webpage sharing a martini recipe is
> multiple megabytes of shit and 1000+ words.  Give me an algorithm that
> reduces it to a three or four line recipe.

How's your perl?  If you have a web server that groks cgi-bin running
on localhost and a fetcher such as wget or curl, you can do this with
a local home page and a cgi-bin perl script.

Doesn't work for pages that rely for content on javascript but where
js is peripheral, you can elide 90% or more of crap.

Especially useful for pages you visit often (or would, were they not
enshittified.)

-- 
Mike Spencer                  Nova Scotia, Canada