I Must Split in Two

Reconciling with myself in an unreconcilable world

I’ve loved to sing since I can remember. A few years ago I started to take it seriously enough to record myself on the phone. Always with the music rolling as a background guideline to prevent me from making the mistake of listening to my voice or letting the recorder capture my off-key verses. But always improving. One of my favorite voices to listen to and sing along to was singer and songwriter Silvio Rodriguez—my mother’s influence on my musical taste.

I don’t bring this up to let you know I’m leaving my career as a writer to pursue the Hollywood dream; I’m not that good at either but I like to think I’ve chosen the easier path. No, it’s just that one of Silvio’s songs has been coming back to me these days as I struggle to find balance between two parts of me that live in conflict. The song’s name is I must split in two. Here’s the chorus:

Let the children, the lovers of rhythm come closer

Let the intellectuals remain seated

I must split in two, I must split in two

Some say here, some say there

And I just want to say, I just want to sing

And it doesn’t matter what the fate of a song can be

Silvio, rhythm lover and intellectual, could never satisfy the thousands of fans who sang along with him at his concerts to his politically charged lyrics. “If you want to be a rhythm lover,” they’d say “then sing about affairs and heartbreak.” Others would respond, “If you want to be an intellectual, then go write a manifesto.”

This fictitious exchange is how I’ve always imagined the internal friction that pushed Silvio to write this song. “I must split in two,” he sings because he is both an intellectual and a rhythm lover and can’t stop being either. It is these days that I feel the most like I must split in two because two parts of me—two big parts each of which I care about a lot—are in patent conflict. A conflict that two years ago was merely a latent, almost negligible itch.

The conflict between AI and the writing world.

The writer in me is sad and angry

Back then, when I used to sing more often to Silvio’s songs, I made a decision.

Unlike him, I chose to write my words instead of singing them. Nowadays, that implies a duty I’d rather sneak away from: I’m part of the resistance.

Watching Gemini generate a decent 1000-word story in 20 seconds feels like treason. I had to work hard to write like this. It’s unfair, like watching in impotence how an undeserving but impossibly gifted student steals your notebook and surpasses your grades only to cast a shadow over your future. Even worse—watching how ChatGPT, as time goes on, generates better words and unscrupulous non-writers use that edge at scale to make a quick buck is an open declaration of war.

The world doesn’t care about us. Companies happily enabling this future are watching impassively how creativity is dying. A flooded, unusable internet seems to be a good price to pay for the short-term benefit of those flooding it with garbage. Because of these reasons, I’m angry; because of those reasons, we’re at war.

But for whom should I fight? Am I perhaps a spy without knowing that I am one?

There’s a reason to love artificial intelligence

At first, I didn’t think about my words much. Ideas came to me and I took them. Like painting the random branches of a narrative tree with my eyes closed, hoping for my brushstrokes to guide me to a place that made sense. I experimented with topics, styles, and structures. Then, I realized the branches were forming a pattern and, as if destiny was laughing at me for what was to come, they thematically converged on AI. So I stuck with it.

That was, not wanting to absolve myself from any responsibility, another decision.

I became interested in AI in 2016, after reading Tim Urban’s popular book-long posts on the artificial intelligence revolution (parts one and two). DeepMind’s AlphaGo Zero was the hint I needed to internalize AI was the future. It was coming sooner than most expected, even experts. (As a quick anecdote, I wrote an article on AlphaGo Zero that was published in an engineering magazine; my first ever public writing.)

When I started writing seriously in 2020, all that knowledge was ready to be downloaded. My words fell onto the page as the promised AI revolution fell like a waking-up cold shower on our heads, first with AlphaGo and AlphaZero, then with GPT-3, and finally with ChatGPT and GPT-4. It was exciting and I felt the thrill of witnessing progress firsthand. I felt part of it.

I have nothing against generative AI as an innovation. As technologists like to say, it’s not the tool but how we use it (and how companies design it) that matters. The wrong incentives are moving some to misuse it but there’s a lot of value waiting to be extracted. Too much funding “brings with it a whole attendant bunch of hype and maybe some grifting,” Google DeepMind’s CEO Demis Hassabis said recently in an interview with the Financial Times. “In a way, AI’s not hyped enough but in some senses it’s too hyped.” Anyone who doesn’t see this dual reality—blatant misuse or even the most inhuman design vs beneficial scientific exploration—conflates the technology itself with the user/company. That’s a mistake.

So I chose AI, which wasn’t generative AI at first, but now I accept it happened anyway. I can’t say I regret that my words led me here.

But now I’m at a crossroads.

Who do I want to be?

The world around me is splitting into a thousand parts (most of them fake) so I feel, like Silvio did, that I must split in two. Who do I want to be? The writer or the AI enthusiast. No longer is there an acceptable option to be both without incurring incoherence those around me would gladly not forgive. Is there any common ground left?

I can always choose a life of eternal conflict that will poke me from the inside if I decide to not give up either part—which, if I’m being honest with you, is what I believe I’ll do. Because I’m truly in love with writing, more than anything else I’ve done in my life. But I’m also an enthusiast of progress, technology, and, yes, also AI. It even feels weird that I have to emphasize that AI is worth being enthusiastic about. It is. But at the same time I feel ashamed of my association with that damning acronym I was once so bullish on. I want to be bullish on AI again. But my desires matter little. I have to make yet another decision.

I… just never expected that making artificial minds had to pass through and above us with an almost painful indifference to human creativity. The turn that AI unexpectedly took with GPT-3 and then ChatGPT sent it in direct collision with the writer in me. One of my parts is trying to eat the other. The clash won’t be—as it hasn’t been—an acute one, but instead, a slow chronic delirium that will spread slowly but surely across all branches of the creative world. On other days, the ones I wake up more optimistic about everything, I remind myself that the world gets better.

So I have to split in two or, like Silvio, not at all. He kept singing how he liked it and it was fine. Will I be able to find the strength and coherence to navigate these rough waters and not subject myself to insanity in the process? I hope so. I will criticize what needs to be criticized. I will also give AI the space to express its potential through my words.

Just know there are two Albertos now. Perhaps more, I don’t know. In a way, we’re all many personas existing in a fragile superposition. In case my words grow repetitive or discordant, know this is the reason. I will modulate from your feedback. That way, despite everything telling me I must split in two, I may avoid it.