Internet Computer (ICP) has recently unveiled its latest development in the AI smart contract engine, focusing on speed optimization to fuel the growth of AI in Web3. The ICP team, led by founder Dominic Williams, has been working tirelessly to enhance the smart contract engine, achieving a 10X acceleration in deterministic floating point instructions.

Williams emphasized the potential of running complex large language models (LLMs) as smart contracts on the ICP protocol. The demo showcased image classification running 3X faster across 3-4 blocks, significantly reducing gas/cycle costs. This improvement was made possible by commits to the public Wasmtime virtual machine implementation of WebAssembly.

The team plans to further optimize the protocol, introducing SIMD instructions into the smart contract execution environment. This will enable multiple floating point calculations to be executed simultaneously with a single instruction. Additionally, ICP aims to migrate its smart contract environment from 32-bit Wasm to 64-bit Wasm, increasing the number of AI model weights that can be loaded into contract memory.

ICP's commitment to AI infrastructure is evident in their introduction of the Internet Computer AI on their testnet. This development highlights ICP's mission to create a system where powerful LLMs can operate as smart contracts on the ICP. With these advancements, ICP positions itself as a strong competitor alongside top AI companies like OpenAI, Google, and Cognition Labs, which recently launched the first AI software engineer, Devin.

Conclusion:

In conclusion, Internet Computer's latest optimization efforts in AI smart contract engines demonstrate their dedication to fostering Web3 growth and AI innovation. By continuously improving their technology, ICP is paving the way for a more efficient and powerful AI ecosystem.