Right, so. Disclaimer first: I am not a spelling expert. I mix up “bear” and “bare” constantly — as in, I genuinely hesitated writing “bear in mind” just now because I couldn’t work out which one it was. I use slang. I abbreviate things. I am not some prescriptivist standing on a hill with a red pen. I also have not slept — not properly, anyway. So I Had Brain Surgery, and it has comprehensively destroyed any concept of a normal schedule; I am either sleeping in three-hour polyphasic bursts or completely unconscious for half a day at a time, with no apparent pattern or warning. It is what it is. I do not know what time it is right now. I do not especially care. I just need to get this out, because something has been quietly deteriorating online for a few years, and I am so, so tired of watching it happen — which is ironic, given that I am also just tired.
The specific culprit, if I had to name one, is video captions.
The TikTok and Shorts pipeline
If you’ve spent any time on YouTube Shorts or TikTok, you’ve seen it. The captions — auto-generated, or manually typed, somehow worse either way — are full of errors that nobody flags and nobody fixes. “Your” instead of “you’re.” “Loose” where they meant “lose.” “Could of.” “Definately.” Not one-offs. Not typos. The same errors, over and over, at scale, on videos with millions of views, in large bold text that you physically cannot avoid reading because that is how the format works.
That’s the thing people miss. You can’t passively watch a YouTube Short. The text is part of it — the emphasis cards, the title slaps, the little labels telling you how to feel. All of it typed at speed on a phone keyboard, none of it checked, all of it enormous. So when the screen says “your loosing your mind” in Impact font at forty-eight points, your brain processes that as language. That’s just what happens. You read it, you absorb it, it goes in.
And then — oh, and then — there’s the brainrot.
I fucking hate brainrot.
I hate the Skibidi Toilet cinematic universe. I hate the sigma edits. I hate the videos that are fifteen seconds of unrelated clips stitched over some audio that stopped being funny eight months before anyone sent it to me. I hate the way they all blur together into the same beige slop of manufactured virality, each one slightly worse than the last, descending in an unbroken chain from “mildly amusing” to “what is this and why is it in my recommended.”
And the six-seven meme. The six-seven meme specifically.
It has no joke in it. There is no punchline. It’s a number. Then another number. That’s it. That’s the content. And yet — because of the sheer, relentless volume of times I’ve been exposed to it — I now have a Pavlovian response to the number 67. A price label. A bus route. A timestamp. A line of code. Doesn’t matter. The synapse fires. My brain goes “six… seven… eheheh” in that glazed, half-present way — like a reflex, or a man who hasn’t slept — entirely without my consent, and then I feel a bit sad about it. At least 69 has a joke in it. A juvenile one, yes, but a joke — a punchline, a reason to exist. 67 has nothing. It just sits there, in my head, taking up space. The mere exposure effect, applied to nothing. Neurological real estate occupied by content so empty it shouldn’t have been able to get in.
I understand I sound approximately eight hundred years old. I don’t care.
Because here’s the thing — and this is why it matters beyond personal taste — the captions on these videos are an afterthought. Someone types the words at speed, slaps them on screen, and moves on. Nobody is proofreading a Skibidi Toilet edit. The very idea is insane. But there the text is, enormous, in front of millions of people. Wrong. Repeated. Normalised.
That’s how it works. You don’t decide to accept a misspelling. You just see it enough times that it stops looking wrong. The pattern your brain is matching to becomes the wrong pattern. And then it migrates — to Reddit, to Bluesky, to Discord, to comment sections — into written-only contexts where there’s no video, no audio, no context to fill in the gap. Just the words. Just wrong. Research confirms that spelling errors affect how readers perceive the competence of a writer. More pressingly, they affect comprehension. And comprehension is where it gets personal.
The part where this is actually a me problem (sort of)
I’m autistic. Interpreting written text is already more deliberate work for me than it appears to be for a lot of people. Every message requires parsing — is this a joke or a dig? Is “fine” fine or is it the bad fine? Is this flat tone neutral or annoyed? Without facial expressions or vocal inflection, even a correctly-spelt, clearly-constructed sentence can take actual effort to decode.
Now add misspellings. Not minor ones — the bad kind, where you genuinely cannot tell what the word was supposed to be. The thread breaks. I’m not mildly inconvenienced. I actually lose the meaning entirely and have to reconstruct it from context, if there is any, which there often isn’t.
The standard comeback is that spelling errors don’t matter if the message gets across. And I get it — there’s a legitimate critique of grammar policing as a social weapon. Fine. But “the message gets across” is doing enormous work there. It assumes a reader who can fill in ambiguity automatically, who can infer meaning from a garbled word and move on without breaking stride. That’s not everyone. That has never been everyone. The conversation about spelling standards is almost entirely led by people for whom it’s a minor annoyance, and the people for whom it’s an actual barrier just sort of… aren’t in the room.
It’s not about purity, it’s about register
I am not arguing for frozen, prescriptivist English. Language evolves. Slang is creative. Anyone who’s apoplectic about dangling modifiers needs a hobby. I use slang constantly — in texts, on Bluesky, in conversation. That’s not the thing.
The thing is register. The understanding that different contexts call for different ways of writing, and that knowing when to switch is itself a skill. Texting a friend is one mode. A caption on a video a million people will read is another. Those shouldn’t be the same, and they increasingly are — not because casual language has won, but because the distinction has just… stopped being maintained. Research links part of this to autocorrect dependency — tools that have quietly trained people to stop checking, because if the phone didn’t flag it, it’s probably fine.
The phone is not always fine.
Errors normalise. Normalised errors stop getting flagged. Unflagged errors spread faster. The baseline drops. Repeat, indefinitely, until “loose” and “lose” are the same word and I’m sitting here looking at the number 67 on a bus timetable having a minor internal incident about it.
And the cost isn’t evenly distributed. Readers who can absorb noise and infer meaning on the fly pay almost nothing. Readers already working harder to interpret tone and intent from text pay considerably more. Nobody seems to be factoring that in.
Anyway. That’s my rant. I still don’t know if it’s “bear in mind” or “bare in mind.” (It’s bear. I looked it up. Again.)