Drowning in Slop: Notes from the Algorithmic Buffet
The internet may not be dead yet but its definitely drowning in zero-cost, low-quality, infinite content.
Sam Altman recently said he’s starting to believe the dead internet theory. Public response was swift and predictable - cue the comparisons to a cigarette CEO announcing he’s worried about the dangers of secondhand smoke.
To his credit, Altman pointed to a swarm of suspiciously enthusiastic Codex posts on the ClaudeAI subreddit as an example. When even the guy selling the lighter fluid notices the fire, you pay attention.
The irony of the messenger aside, the message is hard to dispute. Log on anywhere and you can feel it.
Facebook is a fever dream of AI-generated nonsense - beavers rescuing ambulances, Jesus on oxygen, strangers celebrating birthdays in traffic. Surrealism without the artistry - just endless content mulch. Reddit advice threads are so botted they read like a parody of empathy. YouTube creators swear the algorithm now shuttles viewers from their channels into AI-slop farms - 24/7 noise factories optimized for nothing except “time on site.” Even LinkedIn and Twitter comments read suspiciously like bland echoes of the original post.
“Bot” has become the new slur - both on playgrounds and in forums - a way to say: you’re generic, derivative, disposable. Nothing cuts deeper in 2025 than being told you sound like autocomplete.
The internet may not be dead yet but its definitely drowning in zero-cost, low-quality, infinite content. The result: a legitimacy fog. We’re not only unsure what’s true; we’re unsure who is there. So where does this head? Hard to predict, but here are the downstream consequences I’d bet on:
(1) Curators become kingmakers.
In a world of infinite output, taste is the scarce resource. Creators who seem unmistakably human - opinionated, idiosyncratic, sometimes wrong - become islands of authenticity in a sea of AI slurry. We’ll see editor brands (people, small teams, or even curated platforms) with portable trust - newsletters, pods, playlists, and “follow my feed of feeds.” Think: Morning Brew x10 niches. Attention compounds around voice plus judgment, not volume.
(2) Friction as a feature.
Invite-only spaces, vouch chains, paid memberships, and “phone-book social” (real-world connections first) become differentiators. “Private, invite-only, AI-free” isn’t a Luddite retreat; it’s a logical architecture. Smaller graphs with friction create higher signal. They’re slower to grow, but they monetize better (subscriptions, events, access). They also enforce norms that big platforms have struggled to maintain. If “back to basics” sounds elitist, that’s the point: scarcity is the product. Some spaces will charge for the guarantee that you’re hearing (and being heard by) humans.
(3) The Trust Stack.
As synthetic content floods feeds, provenance becomes as important as production. Expect an infra layer that proves “who” and “how” for any post: proof-of-personhood (human, one-to-one), content signing/provenance logs (e.g., C2PA-style), and reputation graphs (who vouches for whom). Expect a “bot tax”: liveness checks, device attestation, or small paid tokens to post/comment. We’re already seeing grassroots defenses - Reddit users building bots to hunt other bots by tracking account lifespans and engagement patterns. The platforms that lean into this - ranking feeds by verified human origin + accountable identity -will be the ones that win premium ad dollars, creator trust, and eventually the right to call themselves the real internet.
Think of the internet as a 2x2: authenticity on the vertical axis, scale on the horizontal. The lower-left is the wasteland - dead zones - low trust and low reach. Push scale without authenticity and you land bottom-right in AI slop farms, high-volume commodity. Trade reach for trust and you’re top-left in private clubs, premium, invite-only signal. The win state is the top-right: editor brands with provenance - a recognizable voice plus receipts - where attention compounds and CPMs rise.
The tragedy is that platforms could fix this. If they boost signed, accountable media and down-rank synthetic spam, the public square can be rehabilitated without everyone decamping to gated clubs. But they’ll milk engagement until users and advertisers defect. Therein lies the opportunity for disruption: consumers drifting to smaller, high-signal spaces, and new players building for provenance-first networks.
Apple owns the data layer and trust is paramount. Meanwhile, the foundation labs want to eat our data, consume our memories, and then compete us out of existence.
I think this version of your take is the right one: “They know the AI hype cycle is overinflated and the long-term differentiator won’t be novelty but trust. In a world drowning in AI slop - synthetic text, bot comments, spammy videos - the brand that guarantees authenticity and privacy will win. Looks like hesitation, could just be discipline.”