Meta, OpenAI and Uber Just Taught AI Agents to Talk First. What About When to Stay Quiet?
https://www.marktechpost.com/2026/10/03/meta-openai-and-uber-just-taught-ai-agents-to-talk-first-what-about-when-to-stay-quiet/📌 當 AI 主動開口：Meta、OpenAI、Uber 都在賭「說話的時機」
TL;DR：主動式代理的真正難題不是回答什麼，而是何時開口、用哪個管道。
Chatbot 時代，使用者決定何時發問、問什麼；現在反過來，AI 代理主動找上門。問題是，訊息送錯時機，等於直接被使使用者滅音。
🤔 從 Pull 到 Push，介面反轉帶來的新難題
Meta 的 Muse、OpenAI 的 Dots、Uber 的司機助理，三者都採用「代理先開口」的設計。風險很直接：打斷太頻繁，使用者會靜音你；打斷太晚，熱潮早已退去或帳單早已逾期；管道選錯，再好的訊息也會失敗。LLM 能把訊息寫得漂亮，但它並不是決定「要不要送出」的合適工具。
🧩 把「要不要說」拆成可計算的決策
文章提出的核心框架是：每一則主動訊息都是一次賭注，只有當它對使用者的期望價值超過打斷他的成本時才該送出。價值由四個因素構成：賭注有多大、使用者行動的可能性、機會的時效性，以及這個價值究竟屬於使用者還是平臺。高價值且會過期的訊息立刻送出，中等價值的放進摘要（digest），其餘則不送。訊息價值越高，也就越有資格使用較具侵入性的管道——app 內卡片成本低，chat 成本更高，SMS 再高一級，語音通話則保留給緊急且使用者正在忙的情境。文章指出，通知與成長團隊其實已經用類似邏輯解決這類問題多年；Uber 的結果追蹤資料，正好提供了訓練這類模型所需的標籤。
📊 不寫文字、只回傳判斷的新模型類別
與 LLM 不同，這類「決策模型」不產生文字，只回傳有結構的判斷：一個選項、一個分數，或帶機率的是／否。文章提到 TypeSafe 的 Jev，以及 Supersonic Labs 開源的 Julia 1 屬於這一類，其中 Julia 1 可在 CPU 上執行，每次判斷約 33 毫秒。常駐代理評估的觸發次數遠多於實際送出的訊息量，因此可以先用這類便宜、結構化的問題過濾：這值得打斷嗎？有多重要、多急？現在送、延後、整批送出還是直接丟棄？該用卡片、chat、SMS 還是語音？通過這層篩選後，才輪到 LLM 出馬。
⚠️ 模型的侷限與商業化的信任風險
這類決策模型只能根據給它的上下文判斷，無法自行學會「這個使用者早上從不看訊息」之類的個人化偏好。同時，一個能主動開口的代理也是強力的推播通路：Meta 透露正在 Muse 中探索商務功能並已推出 Muse for Small Business；OpenAI 在發布 Dots 同時推出每月 500 美元的新方案；Uber 本身也有外送等服務可以推廣。但促銷訊息仍必須通過同一道「對使用者的價值」門檻——一旦使用者懷疑代理是在推銷而非服務，所有訊息的可信度都會受損。
🎯 實務啟示
打造主動式代理時，別讓 LLM 球員兼裁判。「是否打斷」與「選哪個管道」這兩個決策，應該交給專門、輕量的判斷模型處理，LLM 只負責把內容寫好。擁有多年通知資料的團隊，在這場「注意力判斷力」的競賽中已握有先發優勢。
🔗 來源
- 標題：Meta, OpenAI and Uber Just Taught AI Agents to Talk First. What About When to Stay Quiet?
- 作者／機構：Jean-marc Mommessin, MarkTechPost
- 連結：https://www.marktechpost.com/2026/10/03/meta-openai-and-uber-just-taught-ai-agents-to-talk-first-what-about-when-to-stay-quiet/
#AIAgents #ProactiveAI #DecisionModels #Meta #OpenAI #Uber #LLM #UXDesign #NotificationDesign #MachineLearning
原始資料 MarkTechPost · 收集於 2026-10-03
摘要原文
TL;DR: Meta’s Muse, OpenAI’s Dots and Uber’s driver assistant share one bet: the agent speaks first. That moves the hard problem from what to answer to when to interrupt, on which channel, and with what offer. Classic ML and new decision models can solve it. Chatbots were a pull interface: the user chose the moment, the channel and the question. Proactive agents invert that. Interrupt too often and users mute you. Interrupt too late and the surge has ended or the invoice is overdue. Pick the wrong channel and a good message fails. The LLM can write the message; it is the wrong tool to decide whether to send it. Every proactive message is a bet. Send only when its expected value to the user exceeds the cost of interrupting them. Value has four parts: how much is at stake, how likely the user is to act, how fast the opportunity expires, and whose value it is, the user’s or the platform’s. Value then sets both decisions. High-value, expiring messages go now; modest ones wait for a digest; the rest are never sent. And the more a message is worth, the more intrusive a channel it has earned. Notification and growth teams have solved versions of this for years: Channels have a price of admission. An in-app card costs little, chat costs more, SMS more still, and a voice call is reserved for urgent, hands-busy moments. Uber’s outcome tracking supplies the labels these models need. A new model class suits this job. Decision models such as TypeSafe’s Jev and the open-weights Julia 1 from Supersonic Labs return no text, only a typed judgment: a choice from a short list, a score, or a yes/no, each with a probability. Julia 1 runs on a CPU, from about 33 ms per decision. An always-on agent evaluates far more triggers than it sends, so each candidate can get cheap, structured questions before any LLM call: Is this worth an interruption? How valuable, how urgent? Send now, wait, batch or drop? Card, chat, SMS or voice? The limits matter. These models judge only the context they are given, so they will not learn that one user ignores mornings; An agent that speaks first is a powerful distribution channel. Meta says it is exploring commerce in Muse and has already launched Muse for Small Business; OpenAI announced Dots alongside a new $500 monthly tier; Uber has delivery and other services to promote. But an offer is just another message. It must clear the same bar, measured on value to the user. Once users suspect the agent is selling instead of serving, every message loses credibility. All three companies have strong models. The race will be won on judgment about attention: speaking rarely, at the right moment, in the right place, on the user’s side. Companies with years of notification data may have a head start. The post Meta, OpenAI and Uber Just Taught AI Agents to Talk First. What About When to Stay Quiet? appeared first on MarkTechPost .
由 claude-code/sonnet 自動生成