Twitch Changes Rules So That Streamers Can Mog Each Other

Just a scant handful of months ago, none of us knew what mogging was. Now, thanks to the cutthroat clipping economy and a nonstop stream of Clavicular profiles from mainstream outlets, the manosphere-adjacent term is in every third meme. Recently, a whole game about it, Omoggle, blew up on Twitch, only for streamers to report that they were receiving warnings due to the platform’s rules around “randomized video chat services.” But the human drive to mog is apparently limitless, and in the pharmaceutically-enhanced face of it, Twitch’s rules faltered.

Omoggle, which had over 9,000 concurrent players as of this writing, describes itself as a “live 1v1 mog arena.” In practice, it’s basically a mashup of now-defunct video chat site Omegle and the concept of mogging, in which one person tries to vastly outperform another, often in terms of appearance. You’re paired with a random user, and then the site scans your faces via webcam to determine who ranks higher on the PSL Scale (which stands for Perceived Sexual Market Value; I guess the V is silent). There’s even an ELO-based ranking system, with the precise sorts of tiers you’d expect from something like this: Molecule, Sub3, Low Tier Normie, Mid Tier Normie, High Tier Normie, Chadlite, Chad, and Slayer.

The game’s privacy policy takes great pains to state that—despite what you’d expect from something like this—Omoggle is not a data harvesting farm. Everything apparently runs locally on your device and is immediately discarded after a match ends. The game’s developers “do not store, sell, lease, trade, or otherwise profit from facial meshes, faceprints, or face templates.” Confusingly, however, they do reserve the right to “capture and use match-related content, including clips, screenshots, gameplay visuals, audio, usernames, display names, profile images, rankings, and match results” for advertising purposes. How, if they’re deleting everything? Unclear!

Despite speculation on the parts of publications and players, Omoggle’s creators say that the game does not employ AI—at least, in the modern, LLM-based sense of the term.

“Omoggle is not an AI chatbot or generative AI product in the way people typically use that term today,” Pablo, one of Omoggle’s developers, told Aftermath. “The platform uses computer vision and facial landmark analysis to power parts of the gameplay and scoring system, but it does not use large language models to generate conversations or interactions between users. Many people online broadly refer to anything involving computer vision or machine learning as ‘AI,’ which is probably where some of that language comes from. Internally, we view Omoggle more as a competitive social game built around real-time interaction and custom scoring technology.”

Omoggle recently blew up among Twitch streamers, producing viral clips of big names discovering they’re chopped and conventionally attractive creators coming up short, often due to glitches (and probably also the fact that standards of attractiveness are entirely subjective). But earlier this week, it appeared that the party would come to an abrupt end, as Twitch began issuing warnings to streamers who played it. This stemmed from livestreaming’s rocky history with sites like Omegle and Chatroulette, on which you never knew if you were going to encounter, say, a dazzlingly talented musician or a fully naked person. As a result, Twitch’s rules explicitly forbade the use of “randomized video chat services,” which were “never allowed under any circumstances due to the risks [they pose] to viewers of the stream.”

However, nearly as quickly as word of Omoggle-related warnings broke containment, Twitch reversed course.

“While these sites were previously prohibited on Twitch, we recently updated our Community Guidelines to allow for streaming content from randomized video chat sites,” Twitch said in a statement. “Moving forward, we will only enforce if the content from the randomized video site violates our guidelines (ex. includes gore, sexual content, etc).”

But Twitch does not necessarily recommend visiting Omoggle or other, similar sites: “While our goal with this update is to give you more choice around the content you stream and allow for participation in current trends, we still caution around the use of these sites. The nature of random generation means you can't control what appears on your stream, and there is the potential for violative content to be accidentally shown. Should violative content appear while you're streaming from these sites, we recommend that you quickly remove yourself from the situation by switching scenes and not engaging further.”

Omoggle, for its part, claims to use an open-source machine learning model, NSFW.js, to evaluate whether or not nudity is present in players’ videos.

In any case, mog-offs are back on the menu. Thank goodness.