Many Reports About Microsoft Layoffs Are LLM Slop Based on Other LLM Slop (From Microsoft-controlled LLMs That Downplay the Layoffs or Give Badly Outdated Data)
The LLM slopfarms also dilute/derank actual news about Linux by pushing Microsoft Azure instead; it's a spamfest!
Call this article "Slopwatch" if necessary, but today we won't focus on "Linux" because sometimes we're increasingly bothered that a lot of so-called 'news' about Microsoft layoffs starts with or is entirely slop. It's machine-generated garbage with mild word substitutions that don't constitute contribution. Here's a new example that caught our attention this morning:
The numbers are wrong or very old (or incomplete). The words used above are LLM slop:
This is just the latest example or one new example. We saw many more. If people look for actual information (accurate numbers), they might end up reading LLM slop without even realising it.
How long before this bubble bursts completely? From months ago: "Baidu’s CEO Believes 99% of AI Companies Won’t Survive After the Bubble Bursts"
We're slowly getting there, but not fast enough to eliminate the damage done by slop (in its various forms).
Here's another new example of Microsoft propaganda riding the "Linux" brand:
Of course it is another fake 'article':
Was this generated by Microsoft-controlled LLMs? Or Microsoft interfaces spewing out "Linux" junk?
This was in Google News yesterday; now they blame "Linux" for VMware issues and both of the fakes below are LLM slop. Why does Google News relay FUD composed by machines?
These two slopfarms work in tandem [1, 2; they only ever issue anti-Linux fake 'articles'. This is what the LLMs are inclined to do.
And speaking of Microsoft LLMs, Scam Altman (hired by Microsoft last year) is playing the "give me power and money or China will win!" card again. An associate of ours rightly called this "stenography," (not proper journalism), as it "skips covering the actual issue of algorithmic plagiarism based on the lie that LLMs can actually interpret data rather than just truncate it." Worse yet, it is an inherently cyclic digestion of one's own lies (so-called 'hallucinations') or an ouroboros of sh*t, as explained ages ago, even by pro-slop propaganda sites.
We recommend early writings by Dr. Andy Farnell. "Andy wrote about the feedback loop a while back," an associate recalls. He compared that to feeding dead cows to other cows, basically passing disease from those who were dead to those still living. It was "mad". In light of more recent articles from Andy (about slop on the Web), this is a good opportunity to link to that blog post and several others [1, 2]. As Andy explained back in December: "As large language models (LLM) replace search and authorship this problem will grow significantly. "Authority" shares the same etymological root as "Author". The replacement of human writing by generative systems erodes authority. A post "AI" world is therefore a world without authority. It is where scepticism and critical thinking is meaningless because one cannot examine the provenance of information. Everything becomes speculation and "mere opinion", and is thus easy to counter with "alternative information" but spoken in a louder voice by an "expert" with more power. This is the root of the populist crisis undermining science. For commercial and political gain we've destroyed the epistemological framework by which real expertise and facts can be respected."
It's rather despicable that the world's most notorious lying, vicious companies (likely SLAPPing critics too) are allowed to pollute the Web with all this nonsense, in effect using plagiarism - a copyright violation in other words - to manufacture large piles of propaganda, piggybacking people's work without asking for consent or compensating them (this exploitation happens without even notifying them). They really don't care if it ruins the Web; they lack a strong Web presence anyway. Those LLMs also help libel Richard Stallman. █