Unlocking the Power of Compute on L2
- GoL2 (based on the classic Conway’s Game of Life) is live on Starknet testnet
- It is a powerful demonstration of how compute-hungry dApps, that are impractical on L1, can thrive on L2
What is the Game of Life?
Going back to the 1940s, Computer Scientists were exploring the capabilities unlocked by Turing-complete programs, well before general purpose computers existed. Fast forward thirty years, and British mathematician John Conway had developed the Game of Life, proven to have the power of a universal Turing machine.
Compute on L1 is Expensive
Ethereum, by design, has limited blockspace which is in high demand. It is only natural that the ecosystem is trending towards Layer 2 native apps, which perform computation off chain, and use Ethereum as a settlement layer. We’ve written about this previously here. With Starknet, computation is trivially cheap. This is because Ethereum only needs to verify a single proof that guarantees all the individual computations it attests to were valid. Ethereum does not need to reexecute every individual transaction that is included in the proof. We see a world where computation intensive apps will be developed natively on Starknet.
Gaming on Starknet
Due to Ethereum gas costs, most blockchain games limit their on-chain footprint to minting and trading of assets. The actual gaming logic exists off-chain, often on trusted, centralized servers.
Starknet changes that, as it allows to make the game logic itself trustless and decentralized. GuiltyGyoza of Topology recently presented some gaming primitives that Starknet unlocks. In the Starknet implementation of Game of Life, every single interaction with the game — the game logic itself — is verified on the blockchain, without breaking the bank.
Game of Life is a basic example of on-chain gaming. There is much to explore in gaming, but what other compute-hungry apps are awaiting for their blockchain moment? The paradigm has completely changed. Starknet is the platform that will allow these to shine.