The Web We Lost, the Information Lost Due to Microsoft's Attacks on Companies Like Yahoo! (Before the LLM Slop Frenzy)
Background: Microsoft Hijack of Yahoo | Microsoft Saturated E-mail with Junk (Windows Botnets Spew Out Lots of These) and Now It Does the Same to the World Wide Web
Web Eroding
The World Wide Web is not in a good state or in a healthy shape. Almost nobody would argue otherwise; it's already some very widespread view, not some niche or geek or "fringe" view. People notice that the Web has gotten worse in a lot of ways, both technically and content-wise, i.e. everything gets slower and material is becoming less reliable. Life isn't getting easier on the Web; in a lot of ways, as more things "go online", people are forced to train themselves to do new things and then inherit work others used to do (e.g. clerks/tellers at banks).
When it comes to news sites, what can we say?
Today's "Linux news" speaks for itself:
Nope, the Serial Slopper didn't research this:
As a result of cruft (or even plagiarism) like this, more of us have had to narrow things down and change browsing habits, based on domains or trust "centres". The other day we saw the article "‘If we don’t innovate, we die’: Yahoo CEO Jim Lanzone on reviving a 30-year-old dot-com star".
Microsoft and "Icahn basically killed Yahoo," an associate argues, "and what's left is a shell of itself, getting rid of FreeBSD was one of his goals."
I remember Yahoo in the 1990s. I found it useful for some particular things, e.g. Yahoo Finance.
Internet before dial-up was once considered luxury of rich families, but in the mid-90s it became more prevalent - even ubiquitous in the West - and people could look up information "online", not via teletext or physical newspapers (which were at least a day out of date).
Information Blurred Away
It wasn't just about checking the very latest share prices; sometimes you'd want to check/look up some mundane information, without having access to specialised literature. Yesterday for example I wanted to know the differences between Colman's horseradish, which I had just bought (2KG of it), and Tartar sauce, both in terms of taste and ingredients. I said to someone that they taste not much "un-alike" and both can be used with fish, among other dishes. Colman's tartar sauce is cheaper than horseradish (I usually get those in small jars, this time it's very large) and still it's expensive compared to other sauces. Knowing that Google became an unreliable source of information, I wanted to test it and see what it would say about this topic, if not by LLM slop then at least based on links. So I tried to google around terms like "horseradish tartare difference" (or tartar) and it was laughably bad. I generally don't bother to google anything, having found that's it's a waste of time and many results are just cruft from Reddit, not sensible articles. There's way too much web filler and social control media, ranked more highly based on "recency" (as if newer is always better or more accurate). There's no useful information in that stuff, but it can be good for giggles. What I looked up should be a common topic and indeed in Yandex the results were quite relevant, unlike in Google. The titles were totally relevant, but the cited links (results, references) were all "content" farms, which may be slop (LLM garbage). So forget about Yandex and other GAFAM alternatives; they lack curation/quality and are easy to game if you're some spamfarm. There's a lack of quality control. I concluded that the trajectory of search was something like Google 1990s: we're a small food stand, give us a try. 2000-2010: we have chefs. 2010-2020: we still have OK meals if you pay extra. 2020-25: would you like to try fries instead?
They serve "fast" or "junk" results (or "AI summaries", which are flawed, sometimes so deeply flawed that it is dangerous to read them). They try to feed junk to people instead of letting them spend some time trying to read and understand what specialist sites with actual experts are saying. Many people eventually came to the conclusion that they should choose sites by reputation and stick/stay with these (i.e. no 'Web search', just site search instead). Want medical information? Then go to a reputable medical site (with actual intelligence), then search it for articles on the topic of interest. No Google, no LLM slop "summaries".
A friend told me that indeed "it's getting to be like that".
Directories/Curation and Online Libraries
Quoting further: "The role of web catalogs like the former Yahoo! has become that much more important of late." (DMOZ also [1, 2])
"It's too bad that the Internet Public Library was killed off by Drexel. Drexel intentionally mismanaged the IPL and then sold the domain Squatters got ahold of the domain and now use it as an essay mill (I assume that's how the squatters got the domain, but am not 100% sure.)"
"The fact is that squatters now abuse the domain and much of the old content."
"Had Icahn not attacked Yahoo! on behalf of Microsoft, it would be in a position to massively benefit from the large pool of people looking to manual curation to avoid LLM slop. The IPL launched 30 years ago, as of Q2 this year. Part of the mismanagement included inflicting JavaScript and preventing 'graceful degradation' of services."
Web-accessible information services and Web directories have long been needed for better quality. Disinformation online isn't a new problem. Even in the 90s there were many utterly crap sites, even ones with "illegal content" (by standards of that time).
It's no secret that the Web unfiltered is trash and it was always like this, but people relied on Google having some sense of quality using PageRank and similar "technology". Now that Google is a peddler of LLMs it lost interest in quality and is instead opting for engagement, propaganda, and so on. Therein lies a very perplexing (but not unprecedented) problem.
How much worse will this get?
Our Daily Links are manually curated by several people. We deem that necessary; it not only shuts out the LLM noise but also disinformation, which threatens democracies and propels really terrible people (dishonest to say the least) into positions of power.
Centralised Hubs of Misinformation (or Worse)
The problem has gotten worse because some media is trying to normalise and play along in the LLM scam. It dooms itself by doing so "and there are no longer news aggregation sites," a friend argues. "Slashdot is gone, SN [Soylent News] is too low traffic, and Reddit and the others are just plain bullshit from the get-go."
"Oddly, as mentioned above, the need for them is coming back."
Slashdot is nowadays full of Linux FUD and Bill Gates lies. It used to help geeks find news that mattered. Soylent News still does this, but as noted above, not many people read it (compared to Slashdot 20+ years ago). Reddit is just glorified social control media and thankfully it is dying.
It's not just Reddit. Twitter (now "X") is also dying for many reasons and Facebook has just become a military contractor, i.e. it's looking to be bailed out by taxpayers.
For many sites a very fatal mistake was, they chose to assume social control media (e.g. Twitter account) would become "the new RSS" and directed people towards other sites, which would later be owned by MElon and be gamed/funded by extremists. This, among other mistakes, doomed the media in a very rapid fashion, at a fast pace once people joined the exodus (young people leaving Facebook, government agencies phasing out Twitter-related activities and so on). It was a very "pull-rug" type of move (search => slop summary => rug pulled). Instead of trying to establish/build loyalty and trust among readers media outlets tried to boast/gloat about having more "followers" in Twitter (even spending a lot of money to buy fake ones).
What does it say about a media outlet now if it has many "followers" in a site called X.com, where racist conspiracy theorists' lies (disguised as "intelligence') are actively peddled to people who didn't even ask for it?
Sites Committing Suicide
A friend said that more and more sites dabble in "AI slop" these days ("AI" as in LLM slop and sometimes slop images) because they became irrelevant and increasingly desperate for change (even if this means embracing a trajectory of failure mounted upon prior failure).
The friend said the "AI slop" just finalises that irrelevance, slowly, then quickly.
Remember "pivot to video" from Facebook. How did that work out for publishers? They sacked many good writers to pursue something else, based on a lie from Mark Zuckerberg, who now insists that people need to have fake ("AI") 'friends' and having already boasted they'd fill Facebook with bot accounts, he now says it would not happen (but it would, it's already happening). █