Pick Rachel, Not Julie - content we’ll sorely need that LLMs can't replace

You know that episode in "Friends" where Ross can't decide between two women, Rachel and Julie? He makes a list—Rachel's all cons: she's spoiled, superficial, just a waitress. They even mention she has chubby ankles. And then there's Julie, seemingly perfect with all pros: she's beautiful, lively, intelligent, passionate. It seems there's no contest, right? But Julie has one flaw. What is it?

Ross says, "She's not Rachel."

That's it. You're not Rachel, so you cannot replace the unique her. There's no logical calculations to it. It's not about what's on paper; it's about something you can't quite put your finger on.

I think of this episode whenever I scroll through my newsfeed, which is beginning to be filled with newborn AI influencers and AI replicas of timeless icons like George Carlin. We're staring down the barrel of a future where over 90% of online content is projected to be AI Generated Content (AIGC) in a year or two, and it’s catalyzing an identity crisis among human content creators.

At first glance, these AI creations catch your eye, mimicking human expressions so perfectly just like the way Julie wowed Ross. But the initial spark fades. The first wave of AIGC apps saw folks flocking in droves, only to drop in retention and engagement rates just as quickly. Diving into the heart of the matter, I think the novelty wears off in part because, in the end, our heart doesn’t connect with the AI-generated 'Julie' air-dropped into our life out of nowhere, but rather cherishes the shared history we have slowly built with the beautifully imperfect, easily-broken human 'Rachel'—the essence of being human that makes our stories, our creations, profoundly meaningful and irreplaceably ours.

Let me tell you why.

We, as humans, are out here, living it in real-time, feeling every high and low, while LLMs are just trying to keep up, processing a second-hand version of our compressions of reality. This is where the disconnect lies. AI’s creative process can't truthfully replicate the moment-to-moment human experiences in the making —the depth, the emotions, the messiness of it all unfolding around us every second that we constantly make sense of.

Take the example of my own journey as a bilingual improv comedy player. Improv's universal magic across cultures lies turning the idea of 'human mistakes' on its head, as you watch players salvage scenes from impossible premises and transform challenging twists into clever setups. Like that time on stage when, faced with a wild prompt, I simply paused, nodded, and said, "True." The audience erupted. Or when I made an unexpected exit with "I'll be back," leaving everyone in stitches.

When we are watching improv shows, we don’t feel the most connected to those perfectly timed, smoothly executed punchlines that feel eerily scripted; rather, it's those uncertain moments where players attempt "yes and" to the most unexpected if not downright wild prompts. What leaves lasting impressions on us isn't always flawless execution, but rather a hero’s resilience in navigating the plot twists, or even delays and deliberation on a bumpy ride that leads NOT to a perfect resolution but rather an acceptance of the absurdity of the circumstances.

Me and my teammates being wild in an improv comedy showcase @Endgames, SF

This observation extends beyond the improv stage to the music that moves us, the expressions that touch us, and the personal stories that bind us.

Watching a virtual character's story unfold can't compare to the feeling of being part of a story told by real people, with all their imperfections and quirks. You don't watch a live-streamed game just for the technical gameplay; you watch it for the palpable tension, the emotional highs and lows of the players, their background stories and the genuine rapport they share with their audience. You don’t tune into a lifestyle vlog to just see a beautifully curated home and a made-up face, you want to connect with the raw, unfiltered moments of life of a real person, just like you, who smiles and cries and curses, gets sick and recovers, ages and changes, lives and lets live.

Caedrel, one of the most popular LoL streamers. What sets him apart is his ability to emotionally connect with audience through exceptional storytelling

These preferences underscore the crucial role of individuality in our connections. Despite AI's superpower in accessing, combining, and presenting an endless array of humanity's creative outputs, its creative process is deeply impersonal. If we lean on it too much, it risks diluting the personal touch that resonates with us on a deeper level. Ultimately, what truly fills the void in our hearts is not the production of an efficient machine but the heart and soul of individuals, and specifically, that one or few particular individuals, unique and special in their own right.

We write a letter to someone we miss. Is the writing good? Not necessarily. But the person reading the letter may find themselves moved to tears. I mention that the gingko leaves on the street have fallen. They remind me of where we first met. Do you remember that alley? We shared a pair of airpods as we strolled together, shoulder-by-shoulder, awkwardly. Do these particular words and imageries have meaning for you? The LLMs don't know. But the person reading the letter, they know. They know entirely. Therefore, the content that resonates the most isn't the product of AI's scalable efficiency but rather the product of an irreplaceable personal connection, built upon shared emotions and personal memories that are stirred within us in the process of creating and receiving. And that’s how meaning is made.

As we navigate a future saturated with AIGC, It's imperative to remember that what we yearn for is not an endless source of stimulation, but rather connection. The new currency of value in the era of AIGC is not creative efficiency, but rather personal meaning-making. This realization points us toward a future where AI's role in content creation is reimagined, not as a replacement for human creativity but as a complement to it. The challenge and opportunity lie in making human-centric improvisation the central driver of an AI-aided creative process. In our creative process, we should keep asking ourselves:

How do we harness the strengths of both worlds - that of an algorithmic artisan and that of a human-centric improviser, creating content that is not only innovative but also deeply personal and meaningful?

How can we best leverage AI to generate prompts and scenarios that catalyze, instead of eliminating, spontaneous creativity, emotional depth and sense-making unique to human interactions?

What content formats (e.g. livestream) lend themselves best to preserve and present the resilience and wonders in human error-recovery?

How do we highlight the individuals behind the content we create through building healthy parasocial relationships?

Now, back with Ross making that choice. I, for one, will always be on the side of Rachel instead of Julie. After all, what we are craving are the connections and meaning that shine through a human-centric creative process. The goal is NOT to replace human sensibilities with AI; the goal is to keep the ‘Rachel’ alive in our stories, facilitated by AI.

By placing humans at the center of Human-AI co-creation, here’s to keeping the essence of our narratives unmistakably human, unmistakably us.