I want to start out by thanking Greg for writing this post. It is honest and critical without being at all shitty. It raises a number of valid concerns and proposes valid solutions without once getting self-righteous. I am grateful.

It turns out building the future is confusing, and some of this stuff is a lot more obvious in hindsight than it was when we started down this crazy path.

Let me be clear: I’m the guy at Axiom Zen who said “we should put cats on the blockchain”. I’m not some decentralization maximalist, and I don’t love Bitcoin in my heart of hearts. I don’t hate my bank, nor my government. But I do hate iTunes and think DRM as we know it is a farce. I don’t like the direction ownership in the digital age has been heading, but I don’t think the “right” answers are necessarily clear-cut either.

When we started working on CryptoKitties, we spent a lot of time arguing about what ownership means, what decentralization means, and why both are valuable. Eventually we settled on two things being absolutely paramount to a decentralized, trustless game: ownership and verifiability. Ownership: No one except the cat owner can decide what happens to a cat. Verifiability: It’s transparent and obvious to everyone why your cat is different (the genetics). These things had to live on the blockchain.

In fact, to make this work, we needed to create an entirely new proposal for Ethereum, ERC-721 (https://github.com/ethereum/EIPs/issues/721). It outlines a framework for handling non-fungible tokens in smart contracts, which are central to the CryptoKitties project, and hopefully to a vast number of other projects that will follow in the future.

Putting the art itself on the network was a much harder question. If you dig into ERC-721, you’ll see that we propose a mechanism for storing metadata (like Kitty images), but we didn’t get it built before launch. With ownership and genes on the blockchain, the art (which is directly derived from those genes) didn’t seem critical on day one.

What we got wrong — and, in retrospect, should have been obvious — is that the vast majority of users don’t feel any sort of connection to, or want ownership of, a 256-bit hex string on the Ethereum network. They want to own their damn cats, and to them, the picture is their Kitty. WHICH MAKES TOTAL SENSE.

The last two weeks have been a whirlwind. It would have deeply arrogant for us to assume we’d have this kind of success out of the gate. We thought we’d have a lot more time to figure some of these things out.

But let’s be clear: We want our users to “own” their cats, in all true senses of that word. My friend and coworker Dieter sent me an email the other day that captures this:

Let’s not lose sight of our vision here. CryptoKitties can and should be as awesome as we can make it, but what we can build on top of this foundation is even bigger.

We will set the norms for this new world if we play our cards right. And the future I want to live in is a world in which digital property rights look more like physical property rights, and less like copyright holders dictating how people use things they’ve paid a fair price for.

Unfortunately, copyright and legal terms of service are blunt instruments. We know that the way things are now is not good enough, but defining “good enough” takes longer than we’d like.

What I can do is make a firm commitment for early next year: We are going to store all Kitty images in IPFS, and store an immutable hash of that image in the blockchain. Once this happens, no one — not us, not you, not some black-hat hacker — will be able to change or destroy the image of your Kitties. Access to the image of your cat will not depend on our servers or our Amazon account. This is a top priority for us now, and was a part of our plan from the beginning.

We are also working to “change the terms to grant you a limited license to use the specific arrangement of graphic elements defined by your Kitty’s DNA,” just as Greg proposed. We’re not sure exactly what that license will look like yet, but we’re in total agreement with this approach, and are currently working to make it a reality.

If you (or anybody else) has thoughts or insights, you can always reach out to me directly — mack@axiomzen.co