You’ve seen the announcement. You know the script by heart. Someone senior at OpenAI, DeepMind, or Anthropic posts a sepia-toned photo of a notebook, a fountain pen, and a single sentence: “I’m leaving to build something new.” The blog post follows like a religious text: vague ambitions, capital-A nouns, and a promise to “safely align superintelligence with human values.” You nod. You scroll. You forget.
Then Taylor Leese posted “I’m leaving OpenAI to build Jurassic Park.”
And I laughed. Then I stopped laughing. Because this joke is the most honest thing I’ve read about AI ambition all year.
Nobody is leaving OpenAI to build a thing. They are leaving to build a story.
The story goes like this: “I saw a gap. I assembled a team. We’re going to reimagine the foundation of X.” Replace X with AGI, robotics, biotech, or—in this case—dinosaurs. The script is identical. The only difference is that Leese admits the deliverable is a Jurassic Park theme park with cloned velociraptors. The rest of them pretend the deliverable is something called “AGI” that nobody can define.
That’s the twist. The satire isn’t mocking AI. It’s mocking the unexamined leap from ‘we have powerful technology’ to ‘we should build god-scale things.’ Jurassic Park is the same genre as AGI—a grandiose claim about engineering nature—but it’s honest about its absurdity. A dinosaur park is concrete. You can smell the amber. You can see the T-Rex. AGI is a floating signifier that means whatever the pitch deck needs it to mean.
Leese’s post is a surgical strike against the narrative machinery of tech ambition. He uses the exact same language: “I’ve spent the last two years thinking about this,” “We’re building a future where extinct species walk the Earth,” “We’re hiring.” It’s perfect. It’s devastating. It’s a mirror held up to every founder letter that says “we’re building the infrastructure for transformative AI” without ever showing you the blueprint.
Most people miss the point: the satire isn’t about dinosaurs. It’s about the genre of tech announcements that sell nothing but suspense.
Think about the response. The top comment on the post: “Finally a company I am qualified to work for, how do I apply?” That’s not a joke. That’s recognition. The reader knows the script so well that they can play along. The joke works because we’ve all been trained to read founder letters as performance art. We’ve learned to decode the rhetoric. We’ve learned to separate genuine progress from narrative performance.
And that’s the real insight. Attention and hype are part of technology’s infrastructure. Every announcement is a product. The blog post is the first version of the thing. The GitHub repo might never ship, but the viral tweet already did. Leese’s post isn’t a parody of AI—it’s a parody of the industry’s addiction to announcement-as-product. And by making the topic dinosaurs, he exposes the absurdity that we’ve normalized.
I’ve seen this firsthand. I’ve sat in rooms where someone pitched “a platform for decentralized intelligence” and I nodded along. Then I read Leese’s post and realized: that pitch is semantically identical to “we’re building Jurassic Park.” The only difference is that one sounds like science fiction and the other sounds like a TechCrunch headline. But both are selling you a story about a future that hasn’t been built, using language that makes the impossible sound inevitable.
Jurassic Park is more honest than most AI pitches because it names a concrete deliverable—dinosaurs—while AGI remains an undefined promise.
So what do we do with this? Two things.
First, read every founder letter now with the Jurassic Park test. Replace the vague noun with “dinosaurs.” If the sentence still makes sense, they’re selling you a story, not a product. “We’re building the infrastructure for dinosaurs”—still works. “We’re aligning dinosaur values with human values”—still works. That’s a problem.
Second, appreciate the craft. Leese’s post is a masterclass in mimetic writing. It hooks you with recognition. It twists with the literal take. It leaves you with a golden quote that you want to screenshot and send to a friend. The post is only a few paragraphs, but it’s more memorable than a hundred op-eds about AI safety. Because it made you feel something first—recognition, then laughter, then unease—and only then asked you to think.
That’s the secret. Emotion first, logic second. The best satire is the most honest journalism. And the most honest AI pitch I’ve seen this year involves actual dinosaurs.
FAQ
Q: Is this article seriously saying that AI pitches are the same as a dinosaur theme park?
A: Not exactly the same, but the rhetorical structure is identical. Both start with a grandiose claim about engineering nature, both rely on a story that hasn't been built, and both sell you a future that sounds inevitable. The difference is that a dinosaur park is concrete while AGI is a floating signifier. The satire exposes the lack of specificity in most AI founder letters.
Q: What's the practical takeaway for someone who works in tech?
A: Apply the Jurassic Park test. When you read a founder announcement, replace the abstract noun (AGI, infrastructure, platform) with 'dinosaurs.' If the sentence still makes sense, you're being sold a story, not a product. This helps you separate genuine progress from narrative performance. Also, learn to write like this—emotion first, logic second, golden quotes at tension points.
Q: Isn't it unfair to compare serious AI research to a joke about dinosaurs?
A: The joke isn't about the research—it's about the announcement. There's a difference between doing AI research and writing a blog post about leaving a company to 'build something new.' The post satirizes the genre of tech announcements, not the science. Serious research can coexist with ridiculous hype. The problem is when the hype becomes the product, and the research never ships.