Pluribus has already proven to be one of Apple’s most popular original series ever, a sci-fi production from Breaking Bad’s Vince Gilligan. But it’s also seemed oddly prescient in how it feels like a metaphor for something currently infecting the entire world, AI. Now, last night, that reached a new level.
On Thursday, a viral meme swept Twitter where if you would ask questions about its owner Elon Musk to its in-house AI, Grok, you would get effusive, fawning answers. Not just that, but Grok would go so far as to say Elon Musk was more athletic than LeBron James, would have been a great NFL draft pick and would have run Nazi Germany better than Adolf Hitler. Users uploaded pictures of infamously shirtless Elon Musk, and Grok would praise what it deemed a fantastic physique. Things escalated into more NSFW territory where users asked, “Would Elon be good at [censored]?” Grok always said yes, citing his smarts and charisma and determined nature.
There is a scene in episode 4 of Pluribus that mirrored this just a few hours later (spoilers follow). In it, Rhea Seehorn’s Carol is attempting to figure out how a hivemind, a permanently happy population now running the world, actually operates. She asks them questions about trashy romance novels she wrote and gets them to earnestly say they are on par with the works of William Shakespeare, listing off the classics though Romeo and Juliet.
The series has the entire world of sycophantic hive members trying to make Carol as happy as possible, never saying no to her for any reason. In the previous episode, she requested a hand grenade, and they brought her one. They reluctantly admitted they’d bring her a nuclear bomb if she really wanted it.
The mirror here, perfectly exemplified with the Elon comparison, is that AI models are desperately trying to please users however they can. OpenAI actually had to walk back a model that was too sycophantic with its praise, but the fact remains that AI will rarely tell you you’re wrong, rarely push back and rarely be driven to offer a negative opinion of you. It depends on the model, but the goal of all these companies is the same: Make users happy, make users feel good about using the product.
The episode 4 comparison on the day of the widespread Elon Grok meme is of course a coincidence. The larger show is actually not written to be a full-on metaphor for AI, and the idea was conceived ten years ago, before any of this even bloomed. Here’s an excerpt from an AOL piece featuring Gilligan:
“With a half-amused shrug, Gilligan says he was not thinking about AI when he conceived of “Pluribus,” and he wrote it before the rise of large language models like ChatGPT. “One thing I did wrong while doing press for ‘Breaking Bad’ was tell people, ‘This is what this meant! This meant that!’” He remembers conversations with journalists and fans in which he drilled down his belief that his protagonist, Walter White, was, in fact, a villain. “I look back, and it was so tiresome,” Gilligan says. “Whatever people want to take away from this show is 100% up to them.”
That said, it doesn’t mean it doesn’t fit our current moment, even if it is by accident. But right down to the day? That’s a one in a million shot if you ask me.
Follow me on Twitter, YouTube, Bluesky and Instagram.
Pick up my sci-fi novels the Herokiller series and The Earthborn Trilogy.