2025: The Year LLM Slop Rose to Prominence and Then Fell
With only a week left in 2025 now is a good time to look back and reflect. Let's consider LLMs, which lack intelligence despite being described (in paid-for media puff pieces) as "AI".
At the very end of 2024 we got concerned that many articles about "Linux" were in fact LLM slop with slop images. After about 2 years of Microsoft chatbots gullible people were led to assume this was "the future". Right now, 12 months later, it's hard to find any examples of that. This past week the Serial Slopper had a go at it:
Aside from that and "LinuxSecurity", we've found almost nothing else. Sites that previously dabbled in LLM slop became inactive. They realised it wasn't going to succeed, it only eroded their reputation.
Perhaps by the end of 2026 the LLM slop regarding "linux" (or "Linux") will have ended completely. One can hope.
Yesterday I discussed a certain phenomenon with a friend: suppliers of slop lack a business model, then they start charging users, then the users flee to the rivals that are still free, whereupon the latter see a surge in "demand" and therefore a surge in operating costs (losses). People just aren't willing to pay for slop images and chatbots; the moment you try to get people to pay for it they'll walk away and look for other free offerings, so in effect the core issue here is a value mismatch; slop costs a lot to produce, but people don't value it enough to justify/cover these energy bills, staff (salaries), hardware etc.
This is why the slop hype is bound to end. █