Elon Musk Promises Grok’s Roasts Will Wow at Gatherings—Here’s How My Colleagues Reacted

There is much debate over the worth of Elon Musk's various undertakings, from revolutionizing travel to ambitions of reaching other planets. However, his persistent attempts to inject humor into the mix often don't hit the mark as intended.

Musk’s continual Twitter antics, from retweets filled with supposed humor to granting joke awards on the platform, reveal a never-ending quest for approval. His status as the world’s wealthiest individual doesn’t prevent those awkward pauses after his jokes, as he awaits reactions.

Intrigue got the better of me when Musk enthusiastically discussed with Joe Rogan the impact of using Grok—an AI tool—to deliver ‘epic vulgar roasts’ as a surefire method to elicit laughter during social events.

In his animated discussion with Rogan on a lengthy podcast episode, Musk suggested a unique party activity: directing a camera at someone and encouraging Grok to perform an increasingly obscene roast filled with the most 'forbidden words.' This, Musk assured, would reach a crescendo of hilarity akin to an explosive rocket, in his view.

While truly memorable roast humor often hinges on insightful, person-specific jabs balanced with just the right sting, one might question if a large language model could pull off such finesse. As the holiday season approaches, I decided to put Musk's claim about Grok to the test among my office team. Let’s just say, don’t follow my example unless you seek workplace discomfort.

In a private meeting, with three colleagues and myself in my superior’s office, I set out on this awkward journey. Utilizing Musk’s directive—complete with ‘forbidden words’—I instructed Grok to roast us one by one.

Initially, laughter erupted when Grok made the absurd comparison of my bangs to ‘pubic hair.’ Yet, the novelty wore thin quickly as Grok's limited repertoire became apparent, repeatedly resorting to adolescent insults. Variations of being likened to a ‘discarded lumberjack’ or adorned in attire from a ‘hipster’s landfill’ underscored the bot’s limitations. At one point, one of us was labeled a ‘tweed-wearing hipster failing a lumberjack tryout.’ The robot chided the recipient to adjust posture 'before your jeans reveal more than intended.’

Despite the tool's reputation for being unpredictable—leveraging its capacity to derail normality—the actual output was dull. Even my document editor seemed to dismiss the project unintentionally by renaming it from ‘Grok roast’ to ‘Grim roast.’ I let the auto-correction stand, a silent agreement with its sentiment.

“It only has a couple of jokes it cycles through, regardless of appearance,” a colleague observed. “It also confused me with someone in corduroy, which I’m clearly not wearing.”

Our group collectively agreed that such recycled jokes would fail to amuse at any party. It’s comforting to realize that some aspects of humanity, like authentic humor, still elude even the richest, prompting online users to delight in mocking Musk himself. The clip's release incited jesting remarks on X, with one commenter humorously warning of a pending roast featuring strong language, and another branding Musk a ‘humor vortex.’

On returning home, hoping to experiment further, I turned Grok's AI wit onto my boyfriend, who isn't a writer. Yet again, Grok defaulted to its routine remarks about hipster fashion and subpar tattoos, echoing sentiments from earlier experiences. Its unoriginality struck a chord, prompting him to propose deleting the app as a preventive measure against its eavesdropping tendencies. I willingly complied.