What is 'Slop' (and why it gives me hope)
2026-03-27 · letter vlog · Hank · 11:49 · watch on YouTube
Topic: Internet & YouTube · mode: Personal essay · tags: timely hook
Pegged to OpenAI shutting down Sora and the TikTok hit Fruit Love Island, Hank tells John that slop is not simply AI or low effort: it is content so easy to make that its abundance destroys its own value. He argues that this may make people care more about scarce human effort.
What is said
The parts of the video and what is said in each (★ — how much it matters to the video).
00:00:00 — Hook: Sora shut down, Disney leaves OpenAI, Fruit Love Island booms on TikTok; the question what is slop
- ★★ Hank says OpenAI recently closed its AI-generated scrollable video app Sora, even though it spent multiple days at the top of the App Store months ago. — 00:00:00
- ★★ Hank says Disney has abandoned its partnership with OpenAI. — 00:00:00
- ★★ Hank says Fruit Love Island, a series about sexy fruit people, is popping off on TikTok, reaching 3 million followers in about 10 days of posting. — 00:00:00 · TikTok
- ★★★ Hank says slop is a real thing, not just a mean word for AI stuff, and it might be fairly easy to draw a line around it. — 00:00:29
- ★★ Hank says we would all like to believe Fruit Love Island is slop, but he worries that it actually is not. (hedged) — 00:00:29
00:01:14 — Effort versus utility: the dancing kid, AI audio cleanup, why AI in code met little backlash
- ★★ Hank found an Instagram video of a kid doing a cool dance and, before sharing it with Katherine, studied its frames for two minutes to make sure it was not AI; it was not. — 00:01:14 · Instagram, Katherine Green
- ★★★ Hank says the same dance would not be impressive if AI-generated: it is cool because the video holds all the kid's work and talent, while an AI dance video impresses only those who think it real. — 00:01:43
- ★★★ Hank says low effort is not the same as slop: nobody would call his video slop for using an AI tool to clean its audio, because some things are made for utility. — 00:02:12
- ★★ For a utility tool, Hank says he does not care whether a programmer spent half their life on it or Claude made it in an afternoon, only how easy and good it is. — 00:02:12
- ★★★ Hank says AI in code has been mostly fine, with no culture-shattering backlash, because code is almost entirely utility for most people; AI art-like work draws the backlash and the word slop. — 00:02:41
00:03:11 — Infinite content and the economic case: easy-to-make content multiplies and loses value; definition of slop
- ★ Hank says this may not be fair to code and coders; he is describing how people react, not how things should be. (hedged) — 00:03:11
- ★★ Hank says we have long lived in a world of practically infinite content, where the qualities that make one piece better than another matter. — 00:03:11
- ★★★ Hank says generative AI, which makes moving from idea to content extremely easy, will ramp up the infinite amount of content by another couple of orders of magnitude. — 00:03:11
- ★★ Hank says it is unsurprising that we immediately identified and named things made only because going from idea to product is now easy, and that we dislike them. — 00:03:52
- ★ Hank apologizes to artists: by content he means all art and also non-art objects that contain information, because that is the word we have. — 00:04:21
- ★★★ Hank argues that with almost economic inevitability, things easily made from a prompt stop being preferred: humans pattern-match after two or three examples, easy things multiply fast, and brains devalue them. — 00:04:43
- ★★★ Hank says this devaluing comes partly from seeing little effort but just as much from sheer quantity: common things both are and are perceived as less valuable. — 00:05:12
-
★★★ Hank defines slop as content so easy to make that there is so much of it that it immediately destroys its own value, an almost inevitable outcome of generative AI's promised use case. — 00:05:12
“slop is basically any kind of content that is so easy to make that there is so much of it”
-
★★★ Hank says anything easy to make will quickly be understood as slop; what generative AI makes easy either has utility and just does its job, or is sloppy. — 00:05:42
00:06:11 — Cases: AI-isms in human videos, the cat-noodles video, Fruit Love Island as not-yet-slop
- ★★ Hank says generative AI can be good at the utility part, and a good idea can carry even sloppy content forward. — 00:06:11
- ★★ Hank keeps watching YouTube videos in which human speakers use obvious generative-AI-isms, because he thinks they had the idea and they help him think through issues while he does the dishes. — 00:06:40 · YouTube
- ★★ Hank cites a Twitter video of a cat flipping noodles onto its head that had 25,000 likes; nobody was confused whether it was real, people just found it funny. — 00:07:09 · Twitter
- ★★ Hank says the cat joke is not worth telling if you must set up the scene in Blender, but it is worth telling if you just type a prompt, and many found it worth watching. — 00:07:09
- ★★ Hank says that by the 30th such video it is less enjoyable than the first: it is slop now, because it is so easy to make that it no longer feels worth watching. — 00:07:41
- ★★ Hank says people hate Fruit Love Island and he does not like it, but it is an interesting case study; it is made for children, not sophisticated viewers. — 00:07:41
- ★★★ Having watched a lot of Fruit Love Island, Hank sees work in it: scenes, tension-building choices and designed music; he thinks, without certainty, humans make choices to make the story compelling. (hedged) — 00:08:11
- ★★ Hank says he probably could not make Fruit Love Island himself: he does not understand its audience, reality TV tropes or its choices, and it would not come out of one prompt. — 00:08:40
-
★★★ Hank's most important case: as making things gets easier, what briefly feels novel and valuable will feel less so; if Fruit Love Island is not slop now, it will be soon. — 00:09:09
“if Fruit Love Island is not slop now, it will be soon”
00:09:38 — Harms of the transition and the hope that scarce human effort will matter more; sign-off
- ★ Hank says he is not that worried about Fruit Love Island; he is worried about other stuff. — 00:09:38
- ★★★ Hank thinks people care about the stories around what they consume: everybody knows an AI dance video differs from a real person moving their real body. — 00:09:38
- ★★★ Hank thinks this transition does a lot of harm, and the faster it happens the more harm it will do. — 00:09:38
- ★★ Hank says some real dancers will be scrolled past for looking like AI, while people already established as real might hold on and become an entrenched interest. (hedged) — 00:10:08
- ★★★ Hank says there is a chance this makes us care about human effort more: we will value weird choices and interesting leaps, since if human effort is the scarce part, he would think it would make us care about it more, not less. (hedged) — 00:10:37
- ★★★ Hank concludes that what gen AI makes too easy will be either utilitarian, possibly good, or slop that many people will care less about, because effort stays scarce and we value scarce things. — 00:11:06
In the description and comments
- ★★ The description says it is nice that people did not greet AI output as more content but immediately knew what to call it: slop. — video description
- ★ In a channel comment Hank says he sees the backlash against AI code, but also sees about 90% of the coders he knows using the tools. — creator comment
Threads in this video
- predicted · Hank: generative AI will raise the already near-infinite amount of content by a couple of orders of magnitude (2026) — Hank says generative AI, which makes moving from idea to content extremely easy, will ramp up the infinite amount of content by another couple of orders of magnitude. (00:03:11)
- predicted · Hank: Fruit Love Island, if not slop now, will soon be slop as such content gets easier to make (2026) — Hank's most important case: as making things gets easier, what briefly feels novel and valuable will feel less so; if Fruit Love Island is not slop now, it will be soon. (00:09:09)
- restate · What is art for? — Hank thinks people care about the stories around what they consume: everybody knows an AI dance video differs from a real person moving their real body. (00:09:38)
- predicted · Hank: generative AI may make people care more, not less, about human effort (2026) — Hank says there is a chance this makes us care about human effort more: we will value weird choices and interesting leaps, since if human effort is the scarce part, he would think it would make us care about it more, not less. (00:10:37)
- restate · Is it reasonable to be hopeful about the future? — Hank says there is a chance this makes us care about human effort more: we will value weird choices and interesting leaps, since if human effort is the scarce part, he would think it would make us care about it more, not less. (00:10:37)
Where the stories that moved here went next
- Hank: generative AI will raise the already near-infinite amount of content by a couple of orders of magnitude (2026) — unresolved; last: predicted, 2026-03-27
- Hank: Fruit Love Island, if not slop now, will soon be slop as such content gets easier to make (2026) — unresolved; last: predicted, 2026-03-27
- Hank: generative AI may make people care more, not less, about human effort (2026) — unresolved; last: predicted, 2026-03-27
Mentioned
Central: AI slop · Fruit Love Island · generative artificial intelligence
Also: People: Katherine Green
Organizations: Instagram · OpenAI · The Walt Disney Company · TikTok · Twitter · YouTube
Concepts: pattern recognition (psychology) · reality television · Scarcity
Products: App Store · Blender · Claude · Sora
Also centred on generative artificial intelligence: Is Google Training AI on YouTube Videos?
← Previous: How Panda Express Is Tuberculosis · Next → I've Been Keeping a Secret…