Algorithmically Amusing Ourselves to Death
The disturbing new text-to-video generator, Sora 2
“The new media and technologies by which we amplify and extend ourselves constitute huge collective surgery carried out on the social body with complete disregard for antiseptics. If the operations are needed, the inevitability of infecting the whole system during the operation has to be considered.” - Marshall McLuhan, 1964
I returned home a few days ago with a strange feeling of unease. Didn’t quite know what to make of it. Nothing in my personal life seemed amiss. Did this uncanny sense of dread had something to do with the wider world of bad craziness, an anticipatory weltschmerz of some kind?
I wasn’t online long before I came across the buzz about OpenAI’s updated text-to-video generation AI model, which renders hyperrealistic full motion scenes of, well, pretty much anything you can imagine.
In just a half-day since the release of Sora 2, the web was agog over the short videos from the first wave of content creators. By now you may have seen a few of the deepfake whimsies. Dead singer Michael Jackson stealing a burger from a patron in a fast food joint and singing/dancing his way out the door; dead physicist Steven Hawking performing halfpipe stunts in his electric wheelchair; dead painter Bob Ross rubbing his painted afro on a canvas like a brush; dead preacher Martin Luther King proclaiming “I have a dream, that one day that the ceiling device in my house will stop making those fucking beeping noises.”
The TikTok-like app built around Open AI’s brainchild enables easy sharing and remixing of AI-generated videos. Sign up and away you go. Just plunk a few words into the text field and you can cook up an eight second vid of something never seen, and likely never imagined, ever before.
Highly entertaining, there’s no denying that. All of the clips below, all the images, all the vocals, were entirely generated by AI.
Harmless fun, right?
Sora 2 isn’t primarily a tool for resurrecting dead icons like Queen Elizabeth, Robin Williams or Tupac. Nope, this app is very much intended for the living. The marketing angle involves the option to upload your very own mug, and experience the Faustian frisson of seeing yourself run from either the police or the bulls in Pamploma. Or alien beings if you prefer, before turning the tables by taking out their mother ship with a nicely rendered RPG. How about sharing fries with meth lab wunderkind Walter White, or walking on the red carpet with actress Marilyn Monroe, or taking on a T-Rex in your underwear? (The T-Rex wearing your underwear, if you prefer.)
Strange how in just a short period of time we’ve gone from apocalyptic warnings about deepfake tech to its offering as a mass entertainment option. (I’d love to share with you a video of Geoff “The Eliminator” Olson dropping a body slam on Mark “The Crusher” Carney at Madison Square Gardens, but I’ll leave the capitulation of facial recognition profiles to others.)
A number of commentators on YouTube have been laughing their assess off at the Sora 2 clips—some are indeed hilarious—while inveighing against the social threat from this technology and reciting a meme that’s already attached itself to this spawn of OpenAI: “we’re so totally fucked.”
The fucked aspect has several working parts:
More nuked jobs. As the technology of Sora 2 and other text-to-video generators scale up, work in commercials, film postproduction, production, and even scripting and directing, will predictably decline, following the track of other AI-compromised communications jobs.
You as star. It won’t be long before you’ll be able to do entire podcasts as your younger self or as a warrior in full Viking gear. Whatever you like. Within time, streaming services will be able to offer original films catered to your prompts, starring yourself and any friends and family abiding by company user agreements. Mass entertainment as micro-targeted narcissism.
Trivialization. Probably the least pressing problem, but worth mentioning. Is Steven Hawking or Albert Einstein or Martin Luther King being mocked or lauded in their comic resurrections? Either way, the postmodern flattening of culture heroes into parodied images began in earnest way back in the fifties with Mad magazine. So what? Well, thanks to Sora 2 and its competitors, it’s now the coin of the realm. God forbid if any public figure affects any gravitas at a microphone, because their efforts will be undone by the next morning through digital parody. Obviously not a terrible thing in many cases, but not so much if it involves inventing things they never said or did, that are not obvious fakes. Which brings us to…
Real vs. unreal. This is the most problematic aspect of text-to-video generators. Again, no one is going to be fooled by Steven Hawking in a steel cage death match with Andre the Giant, but believable deepfakes of living people is a different story. The tech companies will surely try to legally fireproof themselves through algorithmic moderation and give-us-your-firstborn user agreements, but the blurring between the real and unreal will continue regardless. Video will become less the gold standard of documentary evidence than just another source of endless contention. I’ll get more into that in the next post.
Massively greater digital distraction. While the US-Israel empire expands its wars of domination across the globe, and CBDCs are being rolled out as part of a tightening regime of global surveillance and Chinese-style social credit, here’s one more surefire way to amuse ourselves to death. Rome 2.0 is all about bread and circuits.
These are just a few of the identifiable downsides of the technology I can think of here. You might be able to imagine a few more. As for the upsides of this tech? Entertainment and convenience!
Sure, the initial wave of text-to-video output will likely cease to seem magical and marvelous within just another few months, as it blends into the mundane marvels we swim through daily, as goldfish with zero attention spans in a tech monopoly tank. Yet this tech is only going to scale up in capability and applicability.
Sora 2 is just a trailer for the blockbuster to come.
More in Part 2 tomorrow.
.