WHAT NEW IN X11.04 WED

AI collections from X · 22 collections · 66,612 posts

jevARTICLE

Building a risk agent without a frontier model using Jev for typed judgments

The setup uses You.com for search, Jev for typed judgments and Qwen for proposals, integrated via MCP, at about $0.02 per sweep.

Edward Irby
@edwardirbyArchived 2026.10.02Originally posted 2026.10.02

What this article says

文章给出了一套不依赖前沿大模型的风险 agent 架构:用 You.com 做检索获取外部信息,用 Typesafe 的 Jev 承担类型化判断(typed judgments),把需要结构化决策的环节交给决策模型而不是通用生成模型,用 Qwen 负责提案生成与综合,最后通过 MCP 把这些组件串起来。作者强调这套组合能把每次扫描的成本压到约 0.02 美元,同时保持可靠性,因为判断环节被约束成类型化输出而非自由文本。结论是:风险类 agent 不必依赖昂贵的前沿模型,把检索、判断、生成拆成不同职责的组件,用决策模型处理判断,就能在低成本下获得稳定可用的结果。

How to build a reliable risk agent without a frontier model ($0.02/sweep):

• @youdotcom search • @typesafeai Jev for typed judgments • @QwenDevs for proposals & synthesis • MCP for integration

Full architecture & cost breakdown below 👇 https://x.com/i/article/2105364784459997184

Edward Irby @edwardirbyOriginally posted 2026.10.02
Open original on X
x.com/edwardirby/status/2105695667486355848Archived 2026.10.02 10:03