Alex Stone

中文参考

Typesafe发布了一个不生成文本的模型，它打破了代理循环的成本底线 jev是他们所谓的系统一模型。你向它发送一个状态和一组问题，它会返回带有概率的类型化答案。没有令牌，没有采样，没有顺序解码。 路线|决定|守卫|下工具|观察|完成？ 它们报告称，分类任务的推理速度比同类LLM快20- 200倍，成本低40- 400倍。请求中的每个问题都是并行评估的，因此添加问题几乎不会改变延迟。 在两圈循环中，这意味着八个决定。需要计划的两个人仍然使用聊天模式。其他六个-选择模型、屏幕工具调用、检查任务是否完成-总共花费0.06秒。 要点很简单：停止为选择支付发电价格。经纪人决定的大部分是选择、得分或是/否。

原帖全文 / EN

TYPESAFE RELEASED A MODEL THAT DOESN'T GENERATE TEXT, AND IT BREAKS THE COST FLOOR OF THE AGENT LOOP jev is what they call a system one model. you send it a state and a set of questions, it returns typed answers with probabilities. no tokens, no sampling, no sequential decoding. route → decide → guard → run tool → observe → done? they report 20-200x faster inference and 40-400x lower cost than comparable llms on classification tasks. every question in a request is evaluated in parallel, so adding questions barely moves latency. in a two-turn loop that means eight decisions. the two that need a plan still go to a chat model. the other six - pick a model, screen a tool call, check if the task is done - cost 0.06s combined. the takeaway is simple: stop paying generation prices for choices. most of what an agent decides is a pick, a score or a yes/no.