We’re thrilled to be backing David, Aaron, Phil, and Hal by investing in Altered State Machine’s US$11m strategic funding round! Altered State Machine is a new web3 protocol to embed, train and own an AI as an NFT.
Making NFTs smart
NFTs have exploded in popularity over the past two years, with trading volume growing ~400x from US$33m in 2020 to >US$13bn in 2021. To date, most of the growth in NFTs has been within art and digital collectibles (e.g., Cryptopunks, Bored Ape YC, Meebits etc.). At their core, most of these projects are images/characters with limited applications outside of use as an avatar or display picture. While we love these use cases as much as the next person, we do believe that NFTs have the potential for much broader use cases.
Enter Altered State Machine (ASM). The ASM platform provides web3 applications with the ability to create AI agents that users can embed and own within an NFT. The key idea behind ASM’s vision is to allow the creation of NFTs that 1) learn and evolve over time based on their environment and experience, and 2) allow users to port AI models between different web3 applications via an NFT.
It sounds complex, but I'll bring it to life by explaining how the ASM protocol would work within their initial area of focus - web3 gaming.
I’ve been into video games as early as I can remember. Some of my best memories growing up involved sneaking onto the computer late at night to play everything from World of Warcraft to Counter-Strike with my mates.
My favourite game is Team Fight Tactics (TFT). TFT is similar to chess, except you build your board through a combination of champions & items, and take turns battling other opponents' boards. You don’t control your team in battles, and champions & items have a fixed set of attributes. So in theory, you should be able to discover the optimal strategy to play for each and every game.
But, playing the same strategy in each and every game doesn’t sound like much fun. So how do the TFT game developers keep things interesting for players?
Patches contain changes to the inputs of the game (champions, items, traits), which in turn change the optimal and dominant strategies. The idea is that if you patch the game frequently enough, you never leave enough time for a completely dominant strategy to emerge. As a result, players are incentivised to play the game in a more flexible manner, which is a whole lot more fun than forcing the same strategy in every game.
However, when it comes to gaming in web3 - patches don’t work very well. Many web3 games implement semi-regular, minor patches. However, it’s often difficult to implement drastic changes that would cause a material shift in strategy since governance decisions are often decentralized amongst token holders and players. And, given the ‘play to earn’ component, it doesn’t make much sense for players to approve a major patch that could significantly impact their earnings. The frequent end result is deterministic gameplay where players force the same strategy in each and every game.
ASM aims to solve this problem by embedding an AI model within the input of a game (e.g., character, items). The idea is that as the AI model evolves with player input, so does the game input. As a result, the optimal strategy constantly evolves without the need for patches. Further, by housing the AI model within an NFT, players also gain an in-game asset that they own and can port between games that utilise the ASM protocol (e.g., imagine having an NFT of a boxer and porting it over to a mixed martial arts game).
To prove out their technology, the team plans to first launch their own fantasy-ball-style game - AIFA - over the coming months. After AIFA, the team will work with their pipeline of game developers and content owners to launch additional games that utilise the ASM protocol.
Eventually, the team plans to open up the underlying primitive widely, so that any content owner / developer can build their own games, as well as web3 applications outside gaming (e.g., DeFi bots, metaverse assistants) that leverage the ASM protocol.
Meet the team
The core founding team - Aaron, David, Hal, and Phil - possess many of the traits we look for in teams we back. They previously worked together as execs at a leading blockchain development studio and have a wealth of experience in the web3 space. They also have a massive, global vision for the potential applications of NFTs that tie in with other leading NFT projects they’re affiliated with (e.g., Non-Fungible Labs, FLUF World, and Universe XYZ).
We’re stoked to be supporting them on their journey as part of our new, dedicated Web3 fund. If you’re keen to learn more about or potentially join the next great Kiwi Web3 project, check out their website and Discord.
At AirTree, we are actively looking to support more web3 projects. If you’re building something in the space, please do get in touch. Our DMs are open :)