新一代消費級智能體即將亮相

2026年9月29日 00:00
站內 AI 整理稿

AIReportAnalysisOpenAI’s AI agents need to catch upA rumored new ‘Aeon’ platform could be entering a crowded market of 24/7 assistants.A rumored new ‘Aeon’ platform could be entering a crowded market of 24/7 assistants.

by Hayden FieldSep 28, 2026, 6:45 PM UTCImage: Cath Virginia / The Verge, Getty ImagesHayden Field is The Verge’s senior AI reporter.An AI beat reporter for more than five years, her work has also appeared in CNBC, MIT Technology Review, Wired UK, and other outlets.

OpenAI popularized the modern generative AI chatbot, but as its 2026 DevDay event approaches, it’s fallen behind in one of the industry’s hottest categories: continuously running, consumer-facing AI agents.On Tuesday, it will likely try to capture the lead in that race.

Rumors abound that OpenAI will release its own AI agent, dubbed Aeon — its answer to Meta’s Muse, SpaceX’s Grok Bot, OpenClaw, and many more.Aeon could have a shot at success, but to beat the competition, it would need to pull together the best features of several different agent platforms.

The ideal system would be as useful as the open-source OpenClaw, as cheap and easy to set up as Meta’s Muse, and as simple to chat with as the up-and-coming Instinct.

Conversely, it would aim to solve the serious security risks that have plagued agents since the rise of OpenClaw and only continued with Muse.

AI agents — platforms capable of booking travel, buying groceries, scheduling appointments, helping with work presentations, and more — have been an industry goal for years.

OpenAI demoed an agentic API feature at the 2024 DevDay and released early versions of consumer-facing AI agents, focused on computer use and research, in mid-2025.It was only late last year with OpenClaw, however, that agents went from proof of concept to practical success.

Since then, the race has only accelerated, while OpenAI’s visible progress around agents has lagged.At this point, OpenAI’s primary competitor for agentic AI is likely Meta.

Since its release earlier this month, Muse has topped the App Store charts and garnered 600,000 daily active users in the US, according to Apptopia.Early reviews put it in the category of useful but slightly creepy.

Another contender is Instinct, whose eponymous agent has made waves for its ease of use — including the fact that users can communicate with it over iMessage and many other chat apps.

And then there’s Gemini Spark, Google’s own horse in the race, which is also a 24/7-running, consumer-facing agent, with upwards of 30 external service partners, like Dropbox, Uber, and Spotify.

Competing AI agents are, according to some accounts, filling out paperwork for doctor visits, canceling subscriptions, organizing group trips, booking activities, setting up DMV appointments, paying bills, and more — paying off the expectations tech companies set years ago.

But security concerns abound, particularly when agents are given more autonomy.

At least one known (now patched) vulnerability allowed attackers to take control of accounts, and users have complained that Muse gave out a home address without permission and read private messages, though individual anecdotes can be hard to verify.

As of now, little is known about AeonAs of right now, little is known about Aeon.The rare thing we do know is that OpenAI currently employs the creator of OpenClaw, Peter Steinberger, who is very likely lending his expertise.OpenAI has set expectations high for its DevDay announcements.

The company has been hyping up its latest frontier model, GPT-6 Astra, as the moment around which artificial general intelligence arrives — “I think it’s not unreasonable to feel that we are now in the AGI era,” OpenAI president Greg Brockman said during a press briefing earlier this month.

Astra would likely underpin whatever agent it announces, which should ostensibly make something like Aeon more powerful at cybersecurity tasks, software engineering, scientific research, and computer use.But OpenAI is at a disadvantage in some other ways.

Muse integrates with other Meta services used by billions of people, for instance, and Gemini Spark connects to Google’s wide range of enterprise and personal services.

At OpenAI’s past annual events, it’s tried to bridge the gap — like last year, when OpenAI heavily pushed the new ChatGPT Apps SDK, allowing controls for other apps, like Zillow, Spotify, and Canva, to surface within the chatbot experience.

Aeon could be simply one of several things announced at OpenAI’s event.Other rumors include potential updates on the AI hardware device(s) the company is building with famed ex-Apple designer Jony Ive, a potential advancement of GPT-6 (like Astra 6.

1, for instance), a hardware update, and maybe even more AGI claims.A continuously running AI agent would tie a lot of this together, which is why it’s likely OpenAI’s next big bet, particularly as the company says goodbye to “side quests.

” The question is how it’ll stack up against OpenAI’s cheap, heavily integrated competitors’ agents, particularly under the immense pressure of its upcoming IPO.Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

Hayden FieldAIAnalysisOpenAIReportMost PopularMost PopularBose’s first wired earbuds in 11 years add noise cancelingTrump finalizes rule to make cars less fuel efficientVolkswagen replaces ID.

4 with all-electric TiguanThe smart home graveyard is getting crowdedOpenAI keeps bulldozing mathematiciansAdvertiser Content FromThis is the title for the native adMore in AIAMD is acquiring AI company World Labs in a deal worth more than $8 billionAI is supercharging hacking, and your local hospitals and banks aren’t readyFlorida seeks a ban on ChatGPT acting like a personOpenAI keeps bulldozing mathematiciansWalmart won’t hike prices based on your shopping history, CEO saysThe SaaSpocalypse that wasn’t, with Atlassian CEO Mike Cannon-BrookesAMD is acquiring AI company World Labs in a deal worth more than $8 billionJay PetersSep 28AI is supercharging hacking, and your local hospitals and banks aren’t readyHayden FieldSep 28Florida seeks a ban on ChatGPT acting like a personStevie BonifieldSep 28OpenAI keeps bulldozing mathematiciansRobert HartSep 28Walmart won’t hike prices based on your shopping history, CEO saysEmma RothSep 28The SaaSpocalypse that wasn’t, with Atlassian CEO Mike Cannon-BrookesNilay PatelSep 28Advertiser Content FromThis is the title for the native adTop StoriesSep 28OpenAI keeps bulldozing mathematiciansSep 28Volkswagen replaces ID.

4 with all-electric TiguanSep 28AI is supercharging hacking, and your local hospitals and banks aren’t readySep 28Bose’s first wired earbuds in 11 years add noise canceling

Related

相關文章

智東西AI Agent

蹲守北京黑客松現場:機器人主持婚禮、變臉,60支極客小隊連夜攢局手搓Demo

作者 | 程茜 編輯 | 漠影 60支隊伍、40+小時,一場充滿想象力的極客盛宴在中關村上演。 現場圍觀機器人主持婚禮、表演變臉;戴上眼鏡能實時識別飛鳥,也可以用AI眼鏡輔助聽障群體教學實操;各類Agent即將奔赴外貿、教育場景落地試用。 這些腦洞十足又緊貼真實場景需求的作品,都來自剛剛收官的Junction China 2026國際極客嘉年華。 作為歐洲極具影響力的老牌國際黑客松品牌,Junction此番首次落地北京,共設置了AI、機器人和可穿戴設備三條賽道。

2 小時前
智東西AI Agent

為Agent全生命週期量身定製!海光掏出全家桶方案,打通Agent落地每一層

作者 | 程茜 編輯 | 漠影 一場由智能體驅動的應用變革,正在撬動AI底層算力體系重塑。 今年9月,IDC發佈的《2026年中國人工智能計算力發展評估報告》給出一組關鍵預判:全球AI算力需求滿足率自2024年79%持續下行,2027年跌至71%的谷底,直至2030年才回升至77%。 有一個簡單邏輯可以概括算力消耗的變化:Token消耗量≈多智能體協同×單智能體資源消耗×單任務執行次數。智能體的持續運行模式,正在持續放大AI基礎硬件的負載壓力。

2 小時前
雷峰網AI Agent

靈睿智芯完成數億元Pre-A+輪融資,產業資本密集押注

本文作者: 陳悅琳 2026-11-13 09:52 導語:目前,基於P100的智能體CPU研發正緊鑼密鼓推進,並與國內頭部系統廠商達成合作意向。近日,上海靈睿智芯計算技術有限公司(以下簡稱“靈睿智芯”)宣佈完成數億元Pre-A+輪融資。本輪融資由美團戰投領投,京東關聯基金、乾杯基金、力合算芯基金(東方算芯CVC)、鈞犀資本、商湯國香資本、尚頎資本(上汽CVC)、深信服科技的產業投資基金琥珀資本等頭部產業資本和眾多老股東跟投。

11 小時前
智東西AI Agent

B2A崛起!Agent 14大應用與Infra賽道一文看懂

編譯 | 江宇 編輯 | 漠影 11月12日消息,近日,硅谷知名風投機構a16z官網刊文《Welcome to the B2A Economy》(歡迎來到B2A經濟時代),提出新判斷:AI Agent正在成為互聯網的主要用戶,企業未來不僅要爭奪人類消費者,還要爭奪AI Agent帶來的流量和訂單。

1 天前