Ethereum co-founder Vitalik Buterin laid out some views on a modern computing architecture trend that separates tasks into “business logic” and “expensive work.” This structure’s purpose is to make efficiency better through a “glue and coprocessor” approach. It will also make the performance and security become better.
Recently, Vitalik Buterin published an article in which he talked about how this trend is transforming different computing areas including blockchain technology, artificial intelligence, and cryptography.
The model “glue and coprocessor” separates computing tasks into two types. First, “business logic,” which concerns the less intensive computation but the flow of information is needed. Second is “expensive work,” which needs the implementation of high efficiency. This division makes it possible to have better performance without losing the generality of the computing processes.
Examples of task separation in Ethereum and AI
According to Buterin, these serve as instances of Ethereum Virtual Machine (EVM) in today’s trend. The majority of computations focus on organized, strenuous tasks. These include things like reading and writing storage, cryptographic functions, and so on.
In AI and programmable cryptography, developers divide the operations between high-level programming languages like Python for “business logic” and highly optimized code for “expensive work,” such as CUDA or custom-built ASICs.
The article furthermore emphasizes the fact that the deployment of this architecture across different areas has become commonplace. In AI, for example, targeted matrix computations are excellently performed by either GPUs or chips specially designed for such purpose.
Likewise, the use of zk-SNARKs and such techniques is an illustration of this difference as it relates to general-purpose computing versus specialized hardware for tasks like hashing or signature verification.
Buterin emphasizes a trend in computation moving towards more modular and specialized approaches. This trend leads to better results by using specialized modules for intensive tasks while maintaining a flexible core for general use.
Such an approach not only enhances supercomputing performance but also allows people to use computers easier and more comfortably. This scenario suggests a future with efficiency, security, and openness working together for the support of more resilient and decentralized computing systems.
However, the computing sphere is an ever-changing one, and hence the critical equilibrium between efficiency, safety, and accessibility will be of great importance. This tape and coprocessor model is a decisive step in that direction by endorsing a more modular and thus secure computing environment.
