I'm leaving OpenAI to build Jurassic Park

130 points by j11g


harrigan

Congrats on the launch. This is exactly the kind of work the field needs right now. I'm excited to see where the roadmap goes from here, and looking forward to follow-up posts such as "Why we're not worried about emergent reproductive capability" and "Jurassic World: scaling to meet customer demand".

dbushell

they put the Velociraptor in the "dim" quadrant, I'm doubting their expertise in this venture

lake

I have to say upfront, I am critical of the current trends in the AI industry, and the tech industry broadly. However, I am engaging in good faith, and hope that you will as well.

In short, I think much of your announcement, and, frankly, your startup's idea, relies on a very cherry-picked definition of "dinosaur." You write

dinosaur wins against pretty much everything

and

most scientists agreed that a dinosaur could totally destroy a datacenter if it really wanted to.

I am uncertain of these scientists' expertise, and whether their opinion was free of influence and bias, for one simple reason:

We do already have dinosaurs, right now. Hummingbirds, sparrows, warblers, finches, and the black-capped chickadee are all around, and I don't see them winning against "pretty much anything", especially datacentres (except sparrows, they are known to be undefeatable, so I think with them it's a matter of will).

I do agree about energy efficiency.

I don't have a question.