Workbuddy
AI overview
Sign in and the AI will write an overview from our coverage.
Headlines · 35
- iFlytek to unveil Ascend 950-trained flagship LLM at 1024 event
iFlytek plans to announce a new flagship general-purpose large model trained on Huawei's Ascend 950 at its October 24 1024 Developer Festival, a bid to escape the domestic-compute bottleneck. The article says Spark has been the only mainstream LLM trained entirely on Chinese chips for three years, but Ascend 910B's memory and bandwidth left its long-context and complex-code training about 5x less efficient than Nvidia's H200; Ascend 950's larger memory and bandwidth raise theoretical efficiency against H200 to 65%–85%, giving Spark a base for top-tier models. Commercially, the piece notes H1 2026 revenue of RMB 11.623 billion, up 6.52% year on year, a RMB 204 million net loss attributable to shareholders, a 2.65% decline in government-business revenue, and chairman Liu Qingfeng's August admission that iFlytek was visibly absent from the code and agent boom. The event is also expected to cover funding progress for the spun-out Ciyuan Xinghuo unit and chip and supernode plans.
36氪 人工智能 · 🔥 7 - First Nvidia RTX Spark laptops go on sale: hands-on with local 27B inference
The first Windows laptops with Nvidia's RTX Spark superchip are on sale, and ifanr tested the 16-inch ASUS ProArt 16 RTX Spark (RTX Spark N1X), priced at 28,999 yuan for 64GB of unified memory and 37,999 yuan for 128GB. It marks Nvidia's first complete Windows compute platform — its own Arm CPU, a Blackwell RTX GPU and unified memory — rather than just supplying graphics cards; the 64GB unit exposes roughly 46.7GB of usable GPU memory with LPDDR5X bandwidth peaking near a theoretical 300GB/s. In testing, a Qwen 3.8 27B model in LM Studio generated about 14.3 tokens/s with a 2.83s time to first token, and handled long contexts without obvious degradation; under sustained local inference the machine drew about 86.5W and its 99.9Wh battery would theoretically last roughly an hour and 15 minutes at ~80W average. WorkBuddy offers a beta for downloading and managing local models on the device; the model's file header indicates a Tencent-optimized 50B-parameter MoE, 16GB in size and CUDA 13 accelerated, and LM Studio's local API can feed other tools.
爱范儿 · 🔥 7 - Bosses wish, staff resist: the FDE bottleneck for AI in Chinese enterprises
A 36Kr deep dive examines the awkward position of forward deployed engineers (FDEs) in China — a role pioneered by Palantir and popularized by OpenAI and Anthropic, and now used by Tencent, ByteDance and Alibaba to push AI into enterprises. It describes bosses asking for agents that replace entire IT departments or monitor employee output, while frontline staff resist tools they fear will cost them their jobs. The report also says messy enterprise data and undocumented workflows push FDE projects back into data cleanup, and cites an Accenture survey of 163 Chinese companies in which only 9% reported gains in productivity, revenue or profit from AI.
36氪 人工智能 · 🔥 6 - Meta Muse sparks personal-agent race as China's tech giants play catch-up
A Bohu Finance article republished by TMTPost says Meta launched its personal agent Muse in early September; it reached over 5 million cumulative downloads in 22 days, and Meta shares rose 11.43% on Sept. 21, adding more than $190 billion in market value in a day. Within three weeks, OpenAI introduced the always-on agent Dots at DevDay, Alibaba launched the personal agent Qwen Go, and ByteDance and Tencent were reported to be developing personal assistants across multiple product lines — a pattern the article frames as "can't lead, at least don't fall behind." The piece also lists obstacles: retail and SaaS platforms such as Amazon detect and block Muse's automated traffic; Doubao Phone Assistant opened third-party screen-operation permissions after the SAEP public comment period, but its latest refusal list covers 151 apps including WeChat, Taobao, Alipay and Meituan; and Morgan Stanley estimates that even at 25% weekly token use, running an always-on agent like Muse costs about $37 a month, or $445 a year, per user — far above Meta's roughly $70 average annual revenue per user worldwide.
钛媒体 · 🔥 6 - 2026 AI Industry Conference's Jinan session focuses on AI deployment and FDE roles
A special session of the 2026 AI Industry Conference, billed as the “AI Scenario Deployment and FDE Innovation Exchange,” was held on Oct. 16 at the Jinan Yellow River International Convention and Exhibition Center, organized by the Shandong Big Data Association, Geekbang Technology and Jimu Network, among others. Organizers said they collected 35 AI application scenarios from 120 companies and institutions across more than 10 industries, spanning office work, member services, operations management and production management. The event also launched the 2027 AI Industry Competition and the Gongzhi Fangzhou community and released a 2026–2027 action plan for AI scenario deployment and FDE development, while speakers from Tencent Cloud WorkBuddy, Maigang Data and the 117AI FDE team shared practices in enterprise agent deployment, delivery talent standards and AI security governance.
InfoQ 中文 AI&大模型 · 🔥 6 - Tencent's WorkBuddy, CodeBuddy refund DeepSeek outage credits, add 500 points
DeepSeek's web and API service went down on the morning of Oct 15 for about 4 hours 53 minutes (8:11 to 13:04) before recovering, and DeepSeek's official platform said it would compensate all API users for seven days of usage. Tencent's WorkBuddy and CodeBuddy then followed with their own remedy: users of the China edition who were affected while calling DeepSeek models will get a full refund of the points they spent on the model between Oct 8 and 14, credited within seven business days. Both products are also giving all China-edition users an extra 500 points, which must be claimed manually on the "usage management" page before 24:00 on Oct 22 and expire after one month.
IT之家 AI · 🔥 17 - Ex-DingTalk president Ye Jun founds enterprise AI agent startup, raises seed funding
Ye Jun, formerly an Alibaba vice president and president of DingTalk, has founded an enterprise AI agent company about a year after leaving Alibaba; the startup, roughly three months old, has raised funding in the tens of millions of RMB range led by game company 4399 (Guangzhou 4399 Information Technology), and already has two paying customers. It starts in foreign trade and cross-border e-commerce with what Ye describes as a foreign-trade AI work platform rather than a single agent, handling model routing, token planning, memory management, tool calls and long-horizon task decomposition, with vertical agents for marketing, ad buying and customs declaration layered on top; the team of about a dozen people is focused on product R&D, forward-deployed engineers and new-media marketing, with no sales or marketing staff. Ye said pricing combines top-ups with outcome-based fees instead of fixed subscriptions or seat licenses, and set a first-stage goal of serving 100 unfamiliar paying customers and exceeding 10 million RMB in revenue without a sales team, while positioning the company as a complement to platforms such as WorkBuddy, Feishu, DingTalk and WeCom rather than a competitor.
36氪 人工智能 · 🔥 6 - A 10-year-old document plugin's AI pivot: an indie developer's survival story
A Huxiu AI analysis profiles Gongwen Processing Assistant, a Word/WPS official-document formatting plugin that has survived ten years, growing from a 2.84MB installer in 2013 to 21.04MB while still shipping updates and charging through voluntary tips (12 yuan for two months, 66 yuan a year, 180 yuan for a permanent seat). The article says the product relies on user-supplied DeepSeek and other model APIs (BYOK), added an AI script feature in June 2025, iterated through 40 expert domains, multi-model adapters and web search in 2026, and on August 18 released an AI Skill interface exposing nine command-line switches for external agents such as Doubao, OpenClaw and WorkBuddy. It argues three walls loom: WPS AI building document automation natively, free competitors, and improving LLM formatting ability — while noting the revenue figures are the author's own assumptions.
虎嗅 AI · 🔥 5 - ApiGo v6.3.0 released with i18n support
ApiGo, an open-source API development and management platform, released v6.3.0 with added internationalization (i18n) support. The platform offers a pluggable multi-database data source framework and SQL2API, connecting to SQL and NoSQL sources such as MySQL, Oracle, DM, TiDB, DolphinDB, Hive and DuckDB, and lets users write SQL visually to publish APIs with lifecycle, management and rate-limiting features. The release notes also say it supports MCP access for AI office platforms including WorkBuddy, Qwen Office, Trea Work and Doubao.
开源中国 · 🔥 5 - OPPO AI Mind Ball review: a 9.8g always-on recorder that keeps text, not audio
According to a GeekPark hands-on report, OPPO's AI Mind Ball — first shown at its September 17 ODC event — is a 9.8-gram wearable recorder that starts listening when taken out of its case and stops when returned, records continuously for up to 24 hours on a charge, picks up audio from about 5 meters and uses voiceprint recognition to distinguish the wearer's own voice; an AI button wakes voice input on a 0.5-second long press and bookmarks a moment on a short press. After transcription the system keeps only text and discards the original audio, addressing privacy concerns with an always-on AI privacy light and end-to-end encryption through OPPO AI PCC, while an underlying "PersonaX" memory engine groups discussions on the same topic into project timelines with insight cards. The companion "白鲸 AI" app offers structured daily summaries and to-dos, O-Mind behavioral insights and an "AI growth coach" framing, and can share long-term memory with external tools such as WorkBuddy, Codex, Claude and Cursor via MCP — though it currently requires an OPPO, OnePlus or Realme phone (tested with Find X10) and does not yet reach into ColorOS system-level notifications.
极客公园 · 🔥 4 - Surface Laptop Ultra hands-on: Nvidia RTX Spark brings local AI to Windows on Arm
On Oct 7 Microsoft held a Windows and Surface event in San Francisco with Jensen Huang and Satya Nadella on stage, opening preorders for the RTX Spark N1X-powered Surface Laptop Ultra and the first wave of RTX Spark OEM machines, which ship Oct 16. The top configuration pairs a 20-core Grace CPU, a 6144-CUDA-core Blackwell GPU and up to 128GB of unified memory, with Nvidia claiming up to 1 PFLOPS of AI compute at FP4 with sparsity. In Ifanr's hands-on tests, Qwen3.5 35B-A3B (Q4) ran at about 59.65 tokens/s and Qwen3.8 27B at about 20.98 tokens/s locally under Ollama + Open WebUI, WorkBuddy could download local models and run agent tasks offline, and pricing starts at 21,988 yuan with the 128GB version at 48,388 yuan.
爱范儿 · 🔥 3 - WorkBuddy, Doubao, Qwen Office and Kimi compared as office AI agents
A third-party review compared Tencent WorkBuddy, Doubao Work, Qwen Office and Kimi on the same 4,510 store records covering five brands from January to September 2026, asking each to write a 2,000-word analysis and a visual report. WorkBuddy was the only one to deduplicate September rows (942 records to 940 stores; 353 to 351 for the largest brand); Kimi and Doubao missed the duplicates, and Qwen Office conflated monthly records with store counts, calling three airport stores logged over nine months "27 airport stores" and reporting 700 mall stores where the data showed 83. Doubao delivered an HTML page with seven interactive charts, Kimi a nine-slide deck, and Qwen Office and WorkBuddy twelve-slide decks. Personal monthly plans start at 49 yuan (Kimi), 68 (Doubao), 70 (WorkBuddy) and 78 (Qwen Office); WorkBuddy, Qwen Office and Doubao list 198 yuan per seat per month for enterprise tiers, while Kimi Premium is 4,200 yuan per seat per year with annual payment and a two-seat minimum. The review cites Moonshot examples of about 0.03 yuan to recommend a rock song and 1.6 yuan for a week-long Sichuan road trip plan, and notes no token-to-credit conversion standard is published.
36氪 人工智能 · 🔥 9 - Alibaba, Tencent and ByteDance race into AI office hardware to capture context
Alibaba, ByteDance and Tencent are competing over AI office hardware, treating recording cards, AI earbuds and smart glasses as entry points for collecting offline "context", according to a China Entrepreneur Magazine report republished by 36Kr. Tencent's WorkBuddy opened its platform in September with more than 100 ecosystem partners and integrations into dozens of devices from Plaud, Rokid and Insta360, while Alibaba's Qwen Office launched the QwenNote A2 AI agent device on Sept 22 for 1,199 yuan and followed on Oct 13 with N1 and N1 Pro AI glasses plus clip-on earbuds. ByteDance's hardware matrix includes Doubao earbuds, a Doubao phone and the AI recording bean developed with Anker Innovations under Feishu, as investors and practitioners cited in the report warn that companion AI hardware has very high idle rates while funding has cooled and customer acquisition costs remain steep.
36氪 人工智能 · 🔥 3 - WeCom and WorkBuddy integrate, placing an AI agent in the contact list
On October 14, Tencent announced a joint release of WeCom and WorkBuddy (WeCom 5.0.12 × WorkBuddy 5.8.0), with WorkBuddy appearing in the WeCom contact list as a conversational bot employees can task like a colleague. The two products share an account system and, within company authorization, open up messaging, documents, meetings and cloud drive capabilities, so results can flow back through WeCom chats, groups, the micro drive or scheduled tasks, while administrators manage seats, permissions and usage from the existing WeCom org structure. Hands-on reports say message triage, document handling, report generation and unread-email summaries work, but approvals are not yet supported, outbound messages require manual confirmation, and WeCom-login data is not shared with existing WorkBuddy accounts.
爱范儿 · 🔥 12 - AI office agents: Tencent, ByteDance and Alibaba race for the next AI entry point
An analysis republished by Huxiu from Dolphin Research argues that with coding-scenario penetration near 70%, the AI entry-point battle is shifting from chatbots to task-completing office agents, noting that OpenAI's enterprise-side付费 has surpassed its consumer side at an annualized rate above $20 billion. It compares China's main contenders: Tencent's Workbuddy, which reported over 6 million daily active users in September; Alibaba's Qwen Office, built on DingTalk's 26 million-plus organizations and industry know-how; and ByteDance's Doubao Work, tied to Feishu and Volcano Engine with a field delivery team of roughly 60-70 people, while Zhipu's Zcode lacks a traffic入口. The author estimates a long-term global AI office TAM above $500 billion and about $45 billion for China, arguing that accumulated user-task trajectory data, rather than raw model performance, is the likelier durable moat.
虎嗅 AI · 🔥 0 - Hands-on: Grok Bot fumbles flight booking; bundled X subscription is the draw
A Chinese hands-on review says Grok Bot, the "AI employee" Elon Musk has been promoting in a flurry of X posts, failed a simple task: planning a Nov 13 Guangzhou–Wenzhou round trip for a wedding and comparing flight prices across platforms. It first suggested an early flight requiring a 4 a.m. wake-up, then pushed to book without completing the price comparison, having only retrieved a "from 600 yuan" quote on Tongcheng while Ctrip, Qunar, Fliggy and airline sites were unreadable. The reviewer notes Grok Bot cannot control other phone apps and works through a persistent cloud computer on the web, needing user login and authorization, making it similar to existing browser-controlling Computer Use agents; the real draw, the article says, is the $30/month The X Subscription bundling Grok, Grok Bot, Cursor and X membership with shared AI quotas.
虎嗅 AI · 🔥 0 - QuestMobile: AI-native apps hit 85.1 sessions per user a month
A QuestMobile AI user-insight report says that as of August 2026 China's AI-native apps reached 537 million users, with average usage of 85.1 sessions per person per month, up 67.1% year over year; comprehensive news apps and search engines averaged 123.6 and 35.8 sessions respectively, down 7.6% and 21.7%, which the report reads as the "first entry point" for finding information shifting from search boxes and feeds to AI chat windows. August's top five AI apps by active penetration were Doubao (77.2%), Qwen (29.4%), DeepSeek (22.8%), Yuanbao (8.2%) and Ant Afu (5.8%); nearly 45% of users used three or more AI apps a month, and heavy users averaged 172.9 sessions. The report adds that competition is moving from model capability to ecosystem and scenario depth, and outlines five marketing playbooks including launch events, paid and soft ads, red-packet subsidies, content co-creation and cross-brand tie-ins.
36氪 人工智能 · 🔥 0 - China's office AI war: Feishu, WorkBuddy, Qwen Office and DuMate push agents
On Sept 15, 2026, ByteDance held its Feishu annual conference in Beijing, where Liang Rubo appeared for the first time to back Doubao Work; Feishu 8.0 treats agents as "organization members" that can be pulled into group chats to write docs, edit multi-dimensional tables, schedule meetings, join calls and start approvals, alongside a team-level agent called Doubao Work Companion that is still in co-creation with enterprises. The piece compares Tencent's WorkBuddy, Alibaba's Qwen Office, Baidu's DuMate and Feishu using vendor-reported figures: Feishu says H1 ARR grew 2.5x year on year and agent-callable features rose from 247 to 767 with execution success up from 78% to 95%; WorkBuddy claims 20.97M PC monthly visits in June and raised enterprise pricing from 78 to 198 yuan per user per month; Qwen Office claims 30M users in its first month, roughly 100% faster single-task generation and 75% lower token consumption; DuMate claims 150 iterations in five months and 60x growth in daily queries. It argues models are no longer the moat, with workflow and the combination of model, enterprise context, tool permissions and security governance deciding the outcome.
虎嗅 AI · 🔥 0 - AI app startups retreat into Agent plugin ecosystems as funding cools
An analysis piece from 硅星人Pro (via Huxiu AI) describes how, as model makers and leading Agent products such as Codex expand, a wave of AI application startups is questioning the standalone-app path and instead packaging video editing, design and browser capabilities as plugins Agents can call directly — among them ChatCut, OpenDesign and Ego Lite. The article says legal AI firm Harvey gained more than 200,000 lawyer users after joining Claude's plugin marketplace, yet interviewees argue plugin features can easily be absorbed natively by model vendors in a future update, leaving only a narrow niche. Other founders are giving up on becoming a new entry point and pivoting to B2B custom delivery and cash-generating businesses, or betting on entertainment products Agents cannot replace (such as original-character apps, Tipsy and Emochi), though those still depend heavily on paid user acquisition.
虎嗅 AI · 🔥 0 - STEPX launches Neo, billed as first large-model-native agent phone, from ¥3,999
Shanghai-based terminal brand STEPX unveiled the STEPX Neo Pioneer Edition on Oct 13, describing it as the first large-model-native agent phone: a Snapdragon 8-series device with 16GB+256GB and a bundled “model + software + hardware” price of ¥3,999 covering the handset, Step AOS 1.0, Step Studio and the Step model family, with pre-orders starting that day and first sales on Oct 16. The phone centers on the Amoo personal agent and the agent-native Step AOS 1.0, which adds dual-domain memory for the user and the agent, task scheduling and a “trusted, visible, controllable, reversible” safety framework, while a six-microphone array, Whisper 1.0 pickup algorithm, low-power audio chip and a MiniLED dot-matrix rear screen handle sensing and feedback, and Edge on-device plus Pro/Flash cloud models split work by task. STEPX says Amoo is connected to roughly 100 partners including Alipay, Tencent (WorkBuddy, WeCom, WeRead), Ctrip, Didi, CapCut, Amap and Meituan, and that Step 5 Preview is free to try inside WorkBuddy until Oct 25; buyers get 12 months of 10 billion tokens per month, and the first 10,000 qualifying users receive a five-year zero-cost upgrade ticket.
智东西 · 🔥 0 - 23-year-old's 14-person AI agent startup Instinct hits $10B valuation
Citing Reuters, 36Kr reports that Instinct, the personal AI assistant startup founded by 23-year-old Noah Shinn, announced a $1 billion funding round on September 28 with investors including Sequoia Capital, Benchmark and Coatue, at a reported valuation of about $10 billion — roughly 4x its $25 billion... valuation of $2.5 billion in late August — with a team of just 14 people. Rather than a standalone app, users reach the assistant by text or phone call to book flights, buy household goods, cancel subscriptions or contact merchants, and a group-chat agent feature has been opened to some users. Shinn has said annualized transaction volume on the platform is approaching $1 billion, more than half travel-related, though he did not explain the calculation and the figure is self-reported. The Instinct Selections recommendation feature launched in late September drew complaints after suggesting products such as luggage without a shopping request, leaving monetization, privacy and reliability still unproven.
36氪 人工智能 · 🔥 0 - Manus raises over $500M at reported $4B valuation after Meta deal collapse
According to a TMTPost article, Butterfly Effect, Manus's parent company, said in an Oct. 8 announcement that it closed a new funding round of more than $500 million, led by Boyu Capital and IDG Capital, with existing backers Tencent, Sequoia China and ZhenFund following on. No post-money valuation was disclosed; the article cites late-September media reports putting it at roughly $4 billion — about double the nearly $2 billion Meta had agreed to pay before that acquisition was halted and the company returned to independent operation on Sept. 1. The article also describes the Sept. 28 launch of Manus 2.0 and the Cue personal agent app, the building of a China-focused team with domestic model partners, and ARR said to have grown from $100 million in December 2025 to a $400–500 million range by end-June 2026 — figures that were not officially confirmed.
钛媒体 · 🔥 0 - Tencent to shut down Yuanqi agent platform on November 9
Tencent's Yuanqi team said in an October 8 notice that the agent platform will stop service on November 9, 2026, citing a business adjustment. From October 29 users can no longer create new agents or distribute API keys, and on November 9 the company will close app access, data read/write, API calls and the console, so agents published to channels such as official accounts, service accounts and WeChat customer service will no longer reply. Users can export agent configurations, workflows and conversation records before the shutdown, with a data-export grace period running to January 9, 2027, after which the platform will delete all user data except records it must keep by law; Tencent recommends migrating to its Cloud Agent Development Platform (ADP) or using official-account auto-replies.
IT之家 AI · 🔥 0 - China's AI office battle shifts from models to entry points and billing
An industry analysis rounds up a dense September of AI-office moves in China: Tencent launched the WorkBuddy open platform and said QClaw will shut down on Dec 24, Alibaba said Qwen Office passed 30 million registered users in its first month, JD.com released enterprise platform JD JoyWork, Feishu shipped version 8.0 with a “Doubao work partner” agent, and Baidu Smart Cloud opened its “Dazi” ecosystem platform. Citing Stanford’s 2026 AI Index, OECD, McKinsey and QuestMobile figures — including a roughly 80% fall in the quality-adjusted price index for text-generation models — the piece argues that as model capability commoditizes, competition is moving to entry points, organizational workflow embedding and pricing models such as “seats + AI usage.”
虎嗅 AI · 🔥 0 - Tencent Cloud open-sources TeamAI to share team skills across 16 agents
Tencent Cloud released TeamAI (github.com/Tencent/teamai-cli) on October 10, an open-source team AI collaboration tool previously used internally that has since appeared on GitHub Trending. Built on Git, TeamAI stores team skills, working conventions, tool configurations and project knowledge in a repository so they can be reviewed and versioned like code, then synced to each member's agent; the company says it supports 16 agents including WorkBuddy, CodeBuddy, Claude Code, Codex and Cursor. Tencent Cloud says a test spanning 13 engineering tasks found a lower-cost model scored 6.1% higher and used 76% less model-call cost after gaining access to team knowledge, and the tool is integrated into its ClawPro agent management platform.
AIbase AI新闻 · 🔥 0 - Ma Huateng sits beside Tencent chief AI scientist Yao Shunyu at Shenzhen briefing
Chinese media report that on October 8 Guangdong party secretary Huang Kunming visited Tencent's Shenzhen headquarters to review its AI research and industrial plans, and that chairman and CEO Ma Huateng sat next to Tencent chief AI scientist Yao Shunyu at the following discussion. The reports say Yao, born in 1998, studied in Tsinghua University's Yao class and earned a computer science PhD at Princeton before joining OpenAI as a research scientist in 2024, where he worked on the operator agent product and deep research and is linked to the ReAct method. He left OpenAI for Tencent in September 2025 and in December was named chief AI scientist in the CEO/president's office, reporting to president Liu Chiping while also heading the AI Infra and large language model departments. TMTPost says Tencent merged its Hunyuan multimodal and LLM departments into a foundation model department under Yao in July, and cites the half-year report for capital expenditure up 81.9% year on year to 84.72 billion yuan.
DoNews · 🔥 0 - Youdao Lobster launches migration tool to win over users as Tencent QClaw shuts down
NetEase's Youdao Lobster published an open letter and rolled out a migration assistant inviting users of Tencent's QClaw to switch, ahead of QClaw's planned shutdown on December 24, 2026. The letter says Youdao Lobster is NetEase Group's only desktop agent aimed at office workers, has shipped 82 versions since launching in February 2026, and stresses open-source, open and security principles, adding that it will not treat user experience as an internal competition chip. QClaw is built on the OpenClaw framework, and its team was previously folded into the cloud product unit that houses WorkBuddy.
AIbase AI新闻 · 🔥 0 - Open-source API platform ApiGo v6.1 adds data plugins and MCP access
ApiGo, an open-source online API development and management platform used as an enterprise project scaffold, SQL2API and data-preview tool, has shipped version 6.1 with data plugin support. The release adds MCP connectivity so the platform can plug into AI office tools including WorkBuddy, Qianwen Office, Trea Work and Doubao. It also connects to SQL and NoSQL sources such as MySQL, Oracle, DM, TiDB, DolphinDB, Hive and DuckDB, letting users write SQL visually and publish APIs.
开源中国 · 🔥 0 - Manus raises over $500M at ~$4B valuation, double its Meta-era price
Butterfly Effect, parent of Manus, said it closed a new round of more than $500 million led by Boyu Capital and IDG Capital, with existing backers Tencent, HongShan and ZhenFund participating, at a valuation of about $4 billion — twice the $2 billion Meta had offered and the largest single round for a China-born agent startup. The funding follows Manus 2.0 on Sept. 28, which swapped its underlying agent execution framework for in-house Cascade, added a 24/7 cloud computer and event-triggered workflows, upgraded the desktop app into Manus Studio for video and game work, and launched Cue, which gives each agent its own email, phone number, wallet and computer. The report says ARR rose from $100 million in December 2025 to a $400–500 million range by end-June 2026, though that surge coincided with the seven months Manus operated inside Meta.
36氪 人工智能 · 🔥 0 - AI office apps in China: how Tencent, ByteDance, Baidu and Alibaba chase growth
An industry recap republished by 36Kr compares how Chinese big tech is growing its AI office agents. According to August AI product rankings, domestic PC AI-office clients reached 33.16 million monthly active users, up 10% month over month: WorkBuddy 11.88M, Baidu Dazi 7.52M, Qwen Office 1.58M (+160.39%) and the late-August launch Doubao Work at 662,900. Tencent cold-started WorkBuddy on WeChat and WeCom before a large offline ad push; Baidu Dazi leans on Baidu Search and Netdisk for traffic; ByteDance converts users via Doubao, Douyin and Feishu; Alibaba avoids the consumer traffic war and goes after enterprises through DingTalk and Alibaba Cloud.
36氪 人工智能 · 🔥 0
Experience and discussion from the community
Share my Workbuddy experienceAsk about Workbuddy
Nobody has shared their experience with Workbuddy yet.