Online slang is the shorthand, spelling twists and emoji code that people invent to type faster, signal belonging or dodge moderation, and its meaning shifts by platform. A 2016 emoji study of 571 peach tweets found six distinct meanings.
Wider context sits in the internet culture hub, and this page focuses on how to read a comment section by where it was posted.
Why the same word reads differently on each platform
Between July and September 2024, TikTok reported removing 150 million videos, and 120 million of those were flagged by automated systems. That scale explains much of the slang found in comment sections today: the audience is partly a machine, and writers adjust for it.
The Algospeak entry on Wikipedia adds that social media platforms rely on automated moderation that users have no control over and that may change at any time. A writer who cannot predict which word will trigger a takedown tends to avoid the whole neighbourhood of that word. The result is a vocabulary that spreads quickly, because a workaround only helps if enough people share it, and that dies quickly too, once the filter learns the new spelling.
This is why a glossary alone fails. A list of terms cannot say whether “seggsy” is a joke, a filter dodge or both in a given thread. The platform, the poster and the topic together decide it, and a reader who learns to weigh all three needs far fewer lookups.
Wikipedia’s entry on internet slang describes a second driver. Dialects such as leet and lolspeak “develop as ingroup Internet memes rather than time savers.” In other words, a spelling can be a badge of membership as much as a shortcut. A word that signals insider status in one forum may read as noise, or as a joke at the poster’s expense, in another.
Platform design adds a third layer. Character limits reward acronyms, video comment sections reward euphemism, and chat streams reward short repeated tokens. Reading a comment well means asking three questions: who is the intended audience, what is the platform filtering, and is the spelling meant to be noticed?
Acronyms and letter homophones: the oldest layer
Three acronyms anchor the oldest layer of online slang: “LOL” for “laugh out loud”, “BTW” for “by the way” and “TFW” for “that feeling when”. They began as typing savers and survive in nearly every comment section, though tone has shifted. A bare “lol” at the end of a sentence often softens a statement rather than reports laughter.
The same source classes “LOL” and “ROFL” (“rolling on the floor laughing”) as letter homophones, where letters stand in for sounds. Leetspeak, also written 1337, goes further and uses combinations of ASCII characters to replace Latin letters, so that “leet” itself becomes a number-and-letter string.
Older acronyms rarely carry risk. If a comment is built from them, the writer is probably communicating plainly, and the useful reading skill is tone: a trailing “tfw” usually introduces a relatable, self-deprecating image rather than a literal statement.
Algospeak on short-video platforms
The term “Voldemorting” dates to 2018, when internet researcher Emily van der Nagel coined it, along with “screenshotting”, for strategies users adopt to avoid giving attention to objectionable figures or attracting algorithmic attention. It is named after the Harry Potter villain known as “He-Who-Must-Not-Be-Named”.
Algospeak is the broader family. Wikipedia describes it as using techniques akin to Aesopian language to conceal the intended meaning from automated content filters while staying understandable to human readers. Its examples include a phonological shift in which “sexy” becomes “seggsy” through intervocalic voicing, and an emoji chain in which porn becomes corn and then the corn emoji.
Moderation uncertainty is the cause. The same entry notes that TikTok has faced criticism for unequal enforcement on topics such as LGBTQ people and obesity, which produces a perception that moderation is inconsistent. When rules feel arbitrary, writers pre-emptively disguise vocabulary, even when the original word would have been allowed.
Spelling is not the only method. The entry also records innocuous captions, such as labelling a scantily dressed body “fake body”, and an interview study in which most creators suspected audio was being scanned too, so they used algospeak aloud. The practical lesson for readers: a strangely respelled word in a video comment is rarely a typo. It is usually a deliberate swap, and the sensitive topic can be inferred from context.
Emoji as slang: context decides the meaning
In 2016, Emojipedia and Prismoji took 571 peach emoji tweets and sorted them into six meanings, ranging from the fruit itself to feeling peachy to sexual connotations. The finding shows why emoji reading is closer to guesswork than to translation: one image can carry six uses.
Context narrows the options. A peach beside a recipe link means fruit. A peach in a flirtatious thread means something else. The watermelon emoji offers a political case: Wikipedia notes its use as a pro-Palestinian symbol in place of the Palestinian flag, a swap tied to concerns about suppressed visibility.
The same entry observes that gestures and emoji are common in algospeak, so emoji slang is not limited to a single community. When a comment is only emoji, check the thread, the poster’s other comments and the video topic before assuming a meaning.
A practical method for reading comments like a local
A reliable routine takes about four checks. Start with the platform: a compressed acronym is likely on a text-first service, while a respelled word under a video suggests filter avoidance. Next, read the thread, because replies often repeat or gloss a new term within a few comments.
Third, separate in-group play from filter avoidance. Intentionally garbled posts that exist for the joke belong to the same world as the shitpost, where deliberate awkwardness is the point. Respelling that hides a sensitive topic belongs to algospeak instead.
Fourth, remember that slang ages quickly. The longer arc, from early text communities to today’s feeds, is traced in the internet culture history guide, and it shows why a term from one era can read as dated in the next. Searching a term with the platform name usually finds a dated explanation faster than guessing.
A short worked example shows the routine. A comment under a video reads “he got seggsy with the corn”. The platform is a short-video service, so filter avoidance is likely. The respelling of “sexy” matches the voicing pattern Wikipedia documents, and the corn emoji is the documented stand-in for porn. Read together, the comment refers to a sexual topic that the writer expected a filter to catch, not to anything agricultural.
A second example runs the other way. A reply that says only “tfw no replies” on a text-first service uses an old acronym with no evasion at all. The register is rueful humour, and the right response is to treat it as a relatable aside. Telling these two cases apart takes seconds once the four checks become habit, and it prevents the two usual errors: over-reading a harmless acronym and under-reading a coded phrase.
Frequently asked questions
What is algospeak?
Algospeak is coded language used to get past automated content filters while remaining readable to people. Wikipedia lists examples such as “seggsy” for “sexy” and the corn emoji for porn.
Is online slang the same on every platform?
No. The 2016 Emojipedia and Prismoji study of 571 peach tweets found six meanings for one emoji, which shows how much context changes interpretation.
Who coined “Voldemorting”?
Internet researcher Emily van der Nagel coined the term in 2018, according to Wikipedia’s entry on algospeak. It refers to avoiding a name or topic to escape attention.
Are LOL and BTW still online slang?
Yes. Wikipedia’s internet slang entry still lists “LOL”, “BTW” and “TFW” as common acronyms, though their tone has softened over time.
