[PREV - DE_BERGERAC_2000] [TOP]
ENGINE_TROUBLE
August 10, 2015
April 07, 2020
October 04, 2022
Search engine blues.
In the days of the pre-google web there weren't
really that many people on the web yet, and the
early adopters were members of the technical elite,
which made them more-or-less an intellectual elite.
In the absence of a world with good search engines,
these people did a lot of manual link farming: they
published lists of URLs to relevant pages that they
liked.
Google's initial insight was to rank pages
by how heavily linked they were: if they were
heavily connected they were probably pretty
good, and people doing searches would
probably want to see those first.
Google's success immediately choked off this
supply of information: what was the point of
link farming when you could go straight there
via a google search?
For Google to continue to stay relevant
they needed a different source of data, It's worth remembering though,
and their success gave them access to that no one outside Google
one: they could spy on their users really knows what Google does:
behavior, and treat decisions to click they're free to add tweaks to
on a link in a search result as a vote filtering and ranking search
for that search result. results at will.
The success of the web continued Auerbach claims that
to increase the scale of things the process is now so
that needed to be indexed, but complex that no one
also gave google access to more really understands all
click data it could use to infer of it, not even
the relative quality of links. Google's engineers.
All of these approaches have the BITWASTES_METASTASIZED
virtue of skipping the slow and
expensive process of human curation.
But the first relied on the
endorsement of people who had looked
at the content and decided it was It could be that google
worthwhile afterwards, the second could try to tease out
relies on the choices of people who whether that was the last
may not have looked at the pages search on the subject (and
before. likely a satisfactory link).
This would be impossible
And both approaches suffer to infer with my own behavior:
from using popularity as a I open a tabs from a dozen
rough proxy for quality. likely looking links, and then
start reading through them
And the intelligence of the after the pages are rendered.
people participating in this
popularity contest matters.
A broader-based web is
necessarily a less
intelligent one.
(Dec 2021)
Search engine results often The search engines track
feel like the blind leading user clicks to find out
the blind. what's worth encouraging
people to click on, and we've
long ago reached the critical
level of stupidity among
internet users.
Search on any topical news issue
and you'll get hundreds of links
to crappy 'news' sources that are
all repeating each other.
And perhaps worse: in a
misguided stab at "fairness",
the search results include
links into popular bubbles
that anyone with a brain
understands are pockets of
insanity.
KRUGMAN_WEBBED
David Auerbach,
"Bitwise: A Life in Code":
"Google, though it contained some of the
most highly skilled of humanity, was,
like all large software firms, committed
to serving mediocrity *by definition*." How dare he accuse the
advertisers of mediocrity.
"As of 2018, common consensus declares (He may be confused
the unwashed internet to be a garbage about who is being
dump of humanity's rejects." served.)
"It was no longer 1995, when engineers
made up a large component of the internet
community. Increasingly we became
spectators of our creations."
ENGINE_EXP
--------
[NEXT - BITWASTES_METASTASIZED]