Artists in the Age of AI: An Ari Dinner

Who’s Building the Future and What It Means for Us

Audio Intro

One night last year, Jim Gregory leaned in at our regular Friday night dinner with Ari and said, like he was revealing a lost Van Gogh, “Have you tried ChatGPT yet?” Then he set his phone on the table. A few minutes later, while we were debating Renaissance painting, a clear voice cut in: “Actually, the Renaissance wasn’t a sudden awakening. It was a long, complex emergence…”

We stared at the phone. “Who said that?” Ari asked.

“That,” Jim said, grinning, “is ChatGPT.”

It was like meeting a strange cousin of the internet. Not the jittery, banner-filled scroll we’ve learned to tune out, but something quieter — like a good librarian — thoughtful, strange, and oddly familiar.

Richard Bonk, Jim Gregory, Sally Lund, Aribert Munzner

Now the media is full of headlines about who’s winning the “AI wars.” Google, Microsoft, Meta, Musk — it’s like trying to follow a family feud in a house you’ve never been in. But behind the buzz is a simple truth:

Whoever shapes AI, shapes the culture we live in.

Let’s break it down like neighbors, not engineers.

The Families (and Platforms) of AI

Think of the AI world like a block of inventive families, each building robots in their garage. There are free models and paid models

OpenAI (ChatGPT): Run by Sam Altman, backed by Microsoft. Their bots are smart, polished, and already showing up in Word and Excel.
Google (Gemini): Think of it as a librarian with a thousand tabs open. Built on decades of search data.
Anthropic (Claude): Created by Dario and Daniela Amodei, former OpenAI folks. Claude it’s meant for the least complex tasks and can quickly do them
Meta (LLaMA): Zuckerberg’s team builds powerful open-source models — more like developer Legos than personal assistants.
Mistral: A French team releasing tiny, fast, fully open models — a community kitchen for coders.
Musk (Grok): A brash bot that lives inside Twitter/X. Says it’s about free speech. Feels like a character.

What Is an LLM and Why Should You Care?

Behind every chatbot is an engine called a Large Language Model — or LLM. Think of it as a giant sponge soaked in everything it can read: books, code, Wikipedia, Reddit, news, forums, song lyrics. The bigger the sponge, the more patterns it holds. These AI platforms don’t have built in coding. They only gather information from the internet and produce summaries. AI does not know things, it mainly is a guessing machine that can be very accurate. But AI does not know when it is wrong, so always review your work.

And in this case, size does matter:

OpenAI is said to train on a trillion words.
Google’s Gemini absorbs decades of search behavior.
Claude is trained more carefully, with guardrails.
Meta and Mistral are smaller but open — like giving the recipe away.

The result? Whoever owns the biggest sponge gets to decide what the AI sees, repeats, or omits.

But What’s Open, Really?

“Open-source” means anyone can look under the hood, modify, or build on the tech. It’s like a public art project you’re allowed to remix. “Closed-source” means the recipe is locked in a corporate vault. Right now, most of the powerful tools are closed. That means just a few companies decide how AI thinks, learns, and speaks. And that’s a little like letting one publisher decide every book on the shelf.

It’s telling that one of the most open AI models today, DeepSeek from China, comes from a place known for control, while Western tools stay locked.

So Who’s Winning?

ChatGPT is the most widely used.
Claude is earning quiet trust.
Google is already on your phone.
Meta and Mistral are building tools for others.
Deep Seek is open source and uses 30x less energy than all other models.

Back to the Dinner Table: Real Questions

As our dinner drifted from LLMs back to wine, Richard Bonk — always the one to bring things home — asked: “But what happens when these things stop telling us what we want to hear and start shaping what we think is true?”

Then Denny Sponsler, who had been thumbing through his Audible list, chimed in. “Kurzweil saw this coming in The Singularity is Near,” he said “AI would outpace us unless we learn to shape it — that one day AI would grow faster than we could imagine, and we’d all have to decide whether to merge with it, compete with it, or shape it.” He set his phone down and looked at each of us. “So what are we doing? Merging? Competing? Or shaping?”

That’s when it hit me: this isn’t about who’s winning. It’s about who’s participating.

Why It Matters to Artists

AI is already writing poems, mixing beats, generating images, editing films. Some call it theft. Others call it a tool. Most of us are somewhere in between. But if we want a say in what comes next, we need to ask questions, try things out, stay curious.

As artists, we’ve always faced the unknown. We know how to work with new tools — and when to question them. We shouldn’t fear AI blindly, but we shouldn’t embrace it blindly either. It’s powerful. It’s here. And it’s listening.

Written by Remo Campopiano (with AI assistance)