Author: Valeriy Zamaraiev, CTO of 1kx Network; Translation: 0xxz@金财经
1. The theme of this article is: Why AO may be the new real "Internet computer". I'll try to build something with it and then form an opinion, but for now it's just a summary of what I've learned so far in EthDenver.
2. It is more like a specification than an implementation. Of course, the code is already written, but that's not the point. The specification is key, once anyone writes their own CU, SU and MU, it allows anyone to interoperate. Compare this to hundreds of thousands of lines of code that few people can manage.
3. The execution environment can be any environment. No special assumptions are made. Currently there is wasm, but no one bans EVM. In fact, I bet someone is working on it right now.
4. Erlang model. The simplest way to explain AO to distributed systems people is: it's just Erlang interpreter state on top of Arweave. Add computing resources to improve the performance of this interpreter. Everything is asynchronous, you don't need a giant global computer to advance every step.
5. Heterogeneous computing based on economic incentives: Messages in the Actor model need to be processed, and the demand for this processing will attract exactly the required resources. Compare this to the high-end machine specs that some networks require to boot.
6. By using Arweave as storage, the team can rely on a battle-tested solution that has been running for many years. This makes the overall risk of AO quite low. If the current idea fails for some reason, it will be easy to move to a working model.
7. No permission required. Many protocols and public chains start with centralization. Thanks again to Arweave, AO has been decentralized from the beginning. If we find a centralized component, it's because no one has written a replacement. The key is that it doesn't have to be centralized at all.