Something has gone profoundly wrong with Facebook. The platform that once connected you with friends and family has become, by many accounts, an unrecognizable mess of AI-generated images, engagement-bait posts from strangers, and a relentless tide of content that nobody asked for. The deterioration has accelerated in recent months, and a growing chorus of users, researchers, and industry observers are sounding the alarm that Meta’s flagship product may have crossed a point of no return.
A detailed analysis published by Pilk puts it bluntly in its title: “Facebook Is Absolutely Cooked.” The piece lays out a damning case that the platform’s feed has become dominated by what internet culture has come to call “slop” — low-quality, often AI-generated content designed purely to extract engagement from unsuspecting users. The author describes scrolling through their Facebook feed and finding it nearly devoid of posts from actual friends, replaced instead by bizarre AI imagery, recycled memes from pages they never followed, and advertisements that blur the line between content and commerce.
The Rise of AI-Generated ‘Slop’ and the Death of the Social Graph
The term “slop” has become the unofficial label for the flood of low-effort, AI-generated content that has overtaken Facebook. These are the surreal images of impossibly detailed cakes, photorealistic depictions of elderly couples in fantastical settings, and strange religious imagery that seem to generate thousands of comments from users who appear to believe they are real. As Pilk documents, these posts are not organic expressions of creativity — they are manufactured by content farms that have learned to exploit Facebook’s recommendation algorithm for profit.
The mechanics are straightforward. Facebook’s algorithm prioritizes content that generates engagement — likes, comments, shares, and time spent viewing. AI-generated images, particularly those that provoke emotional reactions or confusion, tend to perform extraordinarily well by these metrics. A fake image of a soldier reuniting with a child, or an AI-rendered portrait of Jesus holding a smartphone, can generate tens of thousands of interactions. Each interaction signals to Facebook’s systems that this is content worth spreading further, creating a self-reinforcing cycle that pushes genuine social content further down the feed.
Meta’s Algorithmic Pivot: From Friends to ‘Recommended’ Content
The transformation didn’t happen by accident. Meta made a deliberate strategic decision to shift Facebook and Instagram away from friend-based feeds toward algorithmically recommended content, a move that CEO Mark Zuckerberg has publicly championed. The company began aggressively pushing “recommended” posts — content from accounts users don’t follow — starting in earnest in 2022 and 2023, explicitly modeling the approach after TikTok’s successful interest-based feed.
In earnings calls and public statements, Zuckerberg has framed this as a positive evolution. Meta has reported that AI-recommended content has driven increases in time spent on both Facebook and Instagram. From a business perspective, the strategy has worked: Meta’s advertising revenue has surged, and the company’s stock price has recovered dramatically from its 2022 lows. But the user experience tells a different story. According to the Pilk analysis, the ratio of content from actual friends and family versus algorithmically injected posts from strangers and pages has shifted dramatically, with some users reporting that fewer than one in ten posts in their feed comes from someone they actually know.
The Scam Economy Flourishing Under Meta’s Watch
Beyond AI slop, Facebook has become a fertile breeding ground for outright scams. The platform’s advertising system — which remains largely self-serve and automated — has been repeatedly exploited by bad actors running fraudulent schemes. Fake celebrity endorsements, cryptocurrency scams, and counterfeit product advertisements run rampant. Users, particularly older demographics who make up an increasingly large share of Facebook’s active user base, are especially vulnerable.
The Pilk piece highlights how the comment sections under many viral posts have become vectors for scam operations. Bot accounts and coordinated inauthentic pages flood popular posts with links to phishing sites, fake investment platforms, and fraudulent storefronts. Meta’s content moderation systems, which the company has simultaneously been scaling back in favor of AI-driven automated enforcement, appear unable to keep pace with the volume. Reports from users suggest that flagging obvious scam content often results in automated responses indicating the material does not violate community standards.
The Demographic Shift and the ‘Boomer Facebook’ Phenomenon
One of the most significant dynamics at play is Facebook’s shifting user demographics. Younger users have largely migrated to Instagram, TikTok, Snapchat, and other platforms, leaving Facebook with an older core audience. This isn’t inherently problematic, but it has created a feedback loop that content farms have learned to exploit with surgical precision. The AI-generated slop that dominates Facebook feeds is often specifically tailored to resonate with older users — nostalgic imagery, religious content, patriotic themes, and emotionally manipulative scenarios involving children, veterans, or animals.
These users are also, on average, less likely to recognize AI-generated imagery or to understand the mechanics of engagement farming. When a grandmother comments “Amen” on an AI-generated image of an angel, or when a retired veteran shares a fake image of a soldier because it moved him, they are unwittingly fueling the very system that is degrading their experience. The content farms behind these posts monetize the engagement through Facebook’s various creator payment programs, through affiliate marketing links, or by building large followings that can later be redirected to scam operations.
Meta’s Moderation Retreat and Its Consequences
Meta’s approach to content moderation has undergone significant changes in recent years. The company has reduced its reliance on human fact-checkers and moderators, shifting toward AI-based systems and, more recently, adopting a community-notes-style model similar to what X (formerly Twitter) uses. In January 2025, Zuckerberg announced that Meta would end its third-party fact-checking program in the United States, replacing it with “Community Notes” and describing the move as a step toward greater free expression.
Critics argue that this retreat from active moderation has directly contributed to the flood of low-quality and fraudulent content. Without human reviewers to catch nuanced forms of manipulation — AI-generated images designed to look real, scam advertisements that mimic legitimate brands, engagement-bait posts that technically don’t violate any specific policy — the platform’s automated systems are left to enforce rules they weren’t designed to handle. The result, as documented by Pilk and numerous other observers, is a platform where the worst content is often the most visible content.
The Business Logic Behind the Decay
What makes Facebook’s deterioration particularly striking is that it is, by Meta’s own metrics, a success story. The company’s financial performance has been exceptional. Revenue growth has been strong, driven by improvements in ad targeting powered by the same AI systems that recommend content. The stock has performed well. User engagement metrics — time spent on platform, number of sessions per day — have reportedly increased as the algorithm serves more recommended content.
This is the fundamental tension at the heart of Facebook’s decline. The metrics that matter to Wall Street and to Meta’s business model are not the same metrics that reflect user satisfaction or platform health. A user who spends twenty minutes scrolling through AI-generated slop, pausing occasionally on a scam ad, and never seeing a single post from a friend is, from Meta’s perspective, an engaged user. The algorithm has optimized for attention capture, not for meaningful social interaction, and it has succeeded spectacularly at the former while abandoning the latter.
What Comes Next for the Platform That Connected the World
The question facing Meta is whether this trajectory is sustainable. Facebook still commands an enormous user base — roughly three billion monthly active users globally. But the qualitative experience of using the platform has degraded to a degree that is difficult to overstate. Long-time users increasingly describe Facebook as something they check out of habit rather than genuine interest, a zombie scroll through content they never chose to see.
There are no easy fixes. Reversing the algorithmic shift toward recommended content would likely reduce engagement metrics in the short term, something Meta would be reluctant to accept given its obligations to shareholders. Cracking down on AI-generated slop would require significant investment in detection technology and a willingness to sacrifice the engagement that such content generates. Addressing the scam problem would demand a return to more aggressive moderation at a time when the company is moving in the opposite direction.
Facebook was once the place where you went to see what your friends were doing. Today, it is increasingly a place where algorithms show you what machines have made, optimized not for your satisfaction but for your attention. The platform isn’t broken in the traditional sense — it works exactly as designed. The problem is that what it’s designed to do no longer has much to do with connecting people, and everything to do with keeping them scrolling. As the Pilk analysis makes clear, Facebook isn’t just declining — it’s been fundamentally transformed into something its original users would barely recognize, and the company behind it shows little inclination to change course.