我一開始並沒有去研究 Newton Protocol,是因爲我當時在尋找另一個 AI 項目。真正讓我被吸引進來的是一個小小的問題:它在我讀完該方案的設計之後一直縈繞在我腦海裏。我不斷思考,真正的決策到底在什麼地方結束,而基礎設施又從哪裏開始。起初,我以爲這種界限會很清楚。但事實並非如此。 我看得越多,就越意識到我一直把 AI 當成主要敘事。這個假設並沒有持續太久。 和許多關注 Web3 的人一樣,我見過不少項目承諾用 AI 實現自動化。大多數討論最終都會圍繞模型有多“聰明”,或者它們能多準確地做出決策。我的第一想法是:Newton Protocol 正在與同一領域裏做同樣的事。我原本以爲重點會放在爲交易、策略執行或自動化的金融管理構建更聰明的代理(agents)上。
I went into Newton Protocol expecting another project using AI as the main selling point. That was my first impression, at least.
The more I looked into it, the more I felt the AI narrative was almost distracting from what it's actually trying to build.
AI strategies will eventually become abundant. Better models will appear, cheaper models will appear, and whatever feels unique today probably won't stay unique for long. What seems harder to replicate is the infrastructure those systems rely on once they start managing real assets and making real decisions.
That's why the secure rollup caught my attention more than the AI marketplace itself. If autonomous agents are going to execute trades, coordinate capital, or interact across protocols, the environment they operate in may end up being more valuable than the agents themselves.
I think that's where Newton becomes interesting. Not because it promises smarter AI, but because it's asking what the foundation for AI-native finance should look like.
I'm still skeptical, but I also think the market might be spending too much time debating the intelligence of the agents and not enough time asking who ends up owning the layer they all have to trust.