GeekPark tests Ant Group's Lingguang 2.0 personal AI assistant
11/01/2026 — 11/01, 18:17·1 sources (all)·2 reports (all)
Key takeaways
- GeekPark spent two days testing Ant Group's Lingguang 2.0, which it says has shifted from generating mini-apps in 30 seconds to being a pers…
Background & analysis
On October 21, Meta disclosed key figures for its personal agent Muse at an earnings call: more than 10 million people had set up Muse, shopping was its most common use, more than a quarter of users were buying things through it with an average order value above $200, and volume in its first week exceeded expectations, forcing Meta to temporarily raise credit limits together with payments partner Stripe. Six days later, on October 27, Ant Group released Lingguang 2.0, repositioning it from "build a small app from natural language in 30 seconds" to "a personal AI assistant that actually manages things for you." The official slogan was framed in three short phrases, roughly "remembers, keeps watch, gets things done." The product connected to Alipay bills, DingTalk, Feishu, and Mijia so the assistant could handle everyday tasks, positioning it as a Chinese counterpart to Muse.
On November 1, GeekPark published the results of a two-day hands-on test of Lingguang 2.0, alongside a comparison with Qwen, Muse, and Claude Duet. In testing, Lingguang 2.0 could read Alipay bills and flag automatic renewals, and it could also read group chats and contacts, but under security policies actions such as sending messages on a user's behalf were blocked. The article argued that the two products face very different conditions: U.S. e-commerce is spread across many independent websites, which lets Muse reach stores through a cloud browser, whereas everyday consumption and communication in China are concentrated inside a few super-apps with separate data and transaction pipelines. Cross-platform permissions, responsibility allocation, and commercial revenue sharing were described as the real bottlenecks. No new product changes or additional timelines have been reported since the test.
AI-generated from 2 reports · updated 4 days ago
Latest turnAnt Group released Lingguang 2.0 on October 27, repositioning it from a quick app generator to a "personal AI assistant" that connects to Alipay bills, DingTalk, Feishu and Mijia. In a two-day hands-on test, GeekPark found it could read bills, flag auto-renewals and access chats and contacts, but safety policies blocked it from actually sending messages, with cross-platform permissions and liability emerging as the real bottlenecks.
- Heat index
- 25
- All sources
- 1
- All reports
- 2
- First seen
- 5 days ago
Reports on this story headlines open the original
Ant Group released Lingguang 2.0 on October 27, repositioning it from a quick app generator to a "personal AI assistant" that connects to Alipay bills, DingTalk, Feishu and Mijia. In a two-day hands-on test, GeekPark found it could read bills, flag auto-renewals and access chats and contacts, but safety policies blocked it from actually sending messages, with cross-platform permissions and liability emerging as the real bottlenecks.
极客公园AI score 56Geek Park spent two days testing Ant Group's Lingguang 2.0, which has pivoted from 'generate a mini-app in 30 seconds' to a personal AI assistant that 'remembers, watches, and gets things done,' connecting to Alipay bills, DingTalk, Feishu, and Mijia. The article compares it with Qwen, Meta's Muse, and Claude Duet, and discusses the challenges of building personal assistants inside China's super-app ecosystem.
极客公园AI score 56
Other stories people are talking about
- Heat index 65Personal agents still lag: Tencent-media benchmarks Lingguang 2.0, Today AI, and Alice across app access, task execution, memory, proactivity, and interaction.1 sources
- Heat index 840OpenAI ships Images 2.6 with GIF and SVG output15 sources
- Heat index 675OpenAI says its cyber model pipeline confirmed 1,724 Linux kernel flaws and generated 948 privilege-escalation exploits11 sources
- Heat index 642RisingOpenAI GPT-5.6 Brings Major Agentic and Computer Use Upgrades13 sources
- Heat index 633RisingCat Wu, product lead for Claude Code, joins OpenAI9 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion

