Guess what. Sora 2 — OpenAI’s TikTok-style app — is barely two days old and already has OpenAI’s own researchers… kinda freaking out.
And honestly, you can’t blame them. Imagine opening your phone to a feed of AI-generated videos, complete with deepfakes of Sam Altman himself. Most find that hilarious, I find that hilarious, but inside OpenAI? The vibes are complicated.
Take John Hallman, an OpenAI researcher. He flat-out admitted: “AI feeds are scary.” Which, let’s be real, is not the tagline you want floating around on launch week. Still, he added that the team tried to design a positive experience, promising that AI will help, not harm.
Then there’s Harvard’s Boaz Barak, who said he feels a “mix of worry and excitement.” Basically, the human version of the 🤷 emoji. And former researcher Rohan Pandey? He used the chaos to plug his new startup, saying: “If you don’t want the infinite AI TikTok slop machine, come join us at Periodic Labs.” Bold marketing move.
But here’s the bigger picture: Sora highlights an identity crisis inside OpenAI.
On paper, it’s a nonprofit with a mission to build safe AI for humanity. In practice, it’s the fastest-growing consumer tech company on the planet. And just so you know that tension has been simmering since ChatGPT, but a TikTok clone? That’s a whole new level of: “wait… is this the mission?”
Sam Altman, of course, weighed in. According to him, OpenAI needs money to build AGI — but hey, why not give people some fun products along the way? You know stuff that makes them smile, play, maybe even forget for a second that compute bills cost billions.
Basically: “trust us, we’re focused on the big picture….”
Not everyone’s buying it, though. California regulators have already warned OpenAI not to lose sight of its nonprofit mission while chasing growth. Because if history has taught us anything, it’s this: once you’re running a feed, engagement becomes king — and engagement doesn’t always play nice with humanity’s best interests.
Now OpenAI swears Sora is different.
They promise it won’t optimize for doomscrolling, will remind you to take breaks, and will mostly show you people you actually know. Wholesome, right?
But let’s be real: no social media app publicizes itself as wanting to be addictive — they just… get there. Fast.
So the real question isn’t whether Sora is fun. It’s whether OpenAI can avoid turning its shiny new AI toy into the same messy machine we’ve all seen before.
Want to dig deeper? Here’s where
