- Background
- dOnlyFans vs. OnlyFiles
- Emerging Tech: Crypto + Generative AI
- Why Medusa needs a Product
- Idea
- Hypothesis
- Buyers
- Sellers
- Example
- Problem
- Solution
- More Examples
- Why Medusa for this Product
- Goal
No, I did not use ChatGPT to write this ;)
Background
A key missing piece for the Medusa project is a fleshed out product use-case with initial traction.
Q1 Milestone: Find, develop and share successful user stories
Currently, our plan is dOnlyFans. Given our limited time, my personal opinion is development will move too slowly because it’s planned for FVM/IPC. I propose an alternative-but-similar idea that I’ll refer to as OnlyFiles
dOnlyFans vs. OnlyFiles
dOnlyFans aims to create a decentralized version of OnlyFans. In my mind, that means the content is explicit or sex work related content. I think that’s a totally valid use-case and product, but it’s going to be difficult to supplant OnlyFans. It has to be 10x better.
OnlyFiles is a similar platform with the same underlying tech, but the premise is that it’s a marketplace to buy-and-sell AI-assisted, expert-created content.
Emerging Tech: Crypto + Generative AI
The watershed moment of technology in 2022 is that content just became 1000x cheaper to create using generative models. Artists, experts, and regular people can create prolifically. Let’s give them the platform where they are free to transact this content. Let’s build the first marketplace for AI-assisted digital content.
Why Medusa needs a Product
I argue that Medusa and Medusa Pay are protocols, not products. Similar to the Uniswap protocol for AMMs. Or the Seaport protocol for NFTs/orderbook asset exchange.
The companies that develop these protocols capture value in their brand/products - Uniswap frontend and Opensea. Another example that closely resembles Medusa is LayerZero and Stargate.
We have a v1 for Medusa and Medusa Pay and we should focus effort to build our own integrated product in order to capture value, rather than wait for others to build on top.
Idea
A platform for selling digital content. But let's position it/market it specifically for selling AI-generated content. There isn't an existing product for that niche yet.
Hypothesis
1. The supply-demand dynamics of AI generated content are different enough from human generated content in order to warrant a new marketplace. 2. The user needs of buying/selling AI generated content are also different.
Buyers
If you want cheap and high-quality content, you should buy content created by an expert assisted with AI (with some sort of reputation/review system).
Sellers
And if you're a seller, you should sell high-volume and accept lower prices because it's cheap to create.
Example
Problem
I want to generate an employment contract. I could do it myself with the help of AI. But the AI isn’t perfect and I’m not knowledgeable enough to prompt correctly and ensure the quality of the output.
Really, I want a good, cheap lawyer to write a contract suited to my needs. But good and cheap didn't exist... until this year (ChatGPT, Stable Diffusion, etc.)
Solution
Search for “Employment Contracts”. Browse content and creators with high ratings. Find your desired content OR request new content from an expert creator/lawyer.
More Examples
“A bot script for Old School Runescape that does X”
“Generated Art by Beeple” (perhaps the market will view generated art as similar to art prints … own a piece by a revered artist, but it’s much cheaper than non-generated)
Why Medusa for this Product
The reason for using Medusa to build this product is two-fold:
- Avoid centralization and censorship risk with decentralized protocols
- I believe there's a large overlap with current crypto users and current generative AI users (Crypto lawyers could be a good example)
Goal
We need to validate this product idea. We already have a very early prototype with https://demo.medusanet.xyz so surely we can take 1 more month to polish it and start testing with users to see if it has any promise.
Let’s deploy a production version on Polygon in Q1 2023