惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

D
DataBreaches.Net
罗磊的独立博客
M
MIT News - Artificial intelligence
G
Google Developers Blog
V
V2EX
D
Docker
博客园_首页
The Cloudflare Blog
人人都是产品经理
人人都是产品经理
Y
Y Combinator Blog
WordPress大学
WordPress大学
T
Tailwind CSS Blog
博客园 - 司徒正美
J
Java Code Geeks
L
LangChain Blog
博客园 - 三生石上(FineUI控件)
B
Blog RSS Feed
博客园 - 【当耐特】
小众软件
小众软件
Apple Machine Learning Research
Apple Machine Learning Research
大猫的无限游戏
大猫的无限游戏
P
Proofpoint News Feed
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
博客园 - Franky

Hacker News - Newest: "AI"

AI can't read an investor deck AI as an attorney? Student uses ChatGPT, Gemini to sue UW over alleged racial discrimination Hacking MCP Servers in AI Systems – The Rug Pull: Tool Changes After Approval GitHub - MeepCastana/KubeezCut: Free Web based video editor Can AI judge journalism? A Thiel-backed startup says yes, even if it risks chilling whistleblowers Coming soon: 10 Things That Matter in AI Right Now DARPA built an AI to fact-check enemy weapons claims What explains heterogeneity in AI adoption? When AI Meets Muscle: Context-Aware Electrical Stimulation Promises a New Way to Guide Human Movements - Department of Computer Science AI Changed How We Build. It Did Not Change What Matters. Linux rules on using AI-generated code - Copilot is OK, but humans must take 'full responsibility for the… Meta spins up AI version of Mark Zuckerberg to engage with employees Code Mode: Let Your AI Write Programs, Not Just Call Tools | TanStack Blog GitHub - Delavalom/graft: Go framework for building AI agents. Type-safe tools, multi-provider (OpenAI, Anthropic, Gemini, Bedrock), zero vendor SDKs. India's TCS tops estimates, says new AI models did not dent services demand Gen Z's fading AI hype Strong feeling: we are in a folded AI reality GitHub - machinarii/total-recall-catalog: A reference catalog of latest knowledge retrieval, memory & RAG systems GitHub - mensfeld/code-on-incus: Give each AI agent its own isolated machine with root, Docker, and systemd. Active defense detects and stops threats automatically.. Quantization, LoRA, and the 8% Problem: Benchmarking Local LLMs for Production AI Iran war: We spoke to the man making Lego-style AI videos that experts say are powerful propaganda Powell, Bessent discussed Anthropic's Mythos AI cyber threat with major U.S. banks GitHub - immartian/bellamem: Persistent belief-graph memory for AI agents. Retrieves decisive context by importance — not recency, not RAG, not /compact. recursive-mode: The Repo-Native Operating System for AI Engineering After the attack on Sam Altman's home, will AI CEO's go on the offensive? The biggest advance in AI since the LLM Opus 4.6 vs GPT 5.4 One Prompt Unity World Generation Test “AI polls” are fake polls Client Challenge Can AI be a 'child of God'? Inside Anthropic's meeting with Christian leaders
Anthropic rolls out Claude Tag, your new agentic AI cowor...
David Gewirtz · 2026-06-24 · via Hacker News - Newest: "AI"
Anthropic logo sticker
Chris Ratcliffe/Bloomberg via Getty Images

Follow ZDNET: Add us as a preferred source on Google.


ZDNET's key takeaways

  • Claude Tag puts an always-on AI coworker inside Slack.
  • Each Slack channel can get its own isolated Claude identity.
  • Claude can watch, remember, follow up, and complete tasks.

Today, Anthropic announced Claude Tag, an always-on instance of Claude that lives inside Slack channels. The idea is that it becomes another participant in a group Slack channel where it can read the channel context, participate in threads, and interact with other channel participants almost as if it were just another employee in the group. Anthropic says 65% of its product team's code is now created by its internal version of Claude Tag.

Anthropic says, "Claude Tag is an always-on Claude that works with your team in Slack the way you'd expect a teammate: reading the room, jumping into the thread, and moving work forward."

Also: How to troubleshoot your PC problems with Copilot or ChatGPT - effectively

Depending on how it's configured, Claude Tag can be solely reactive, or, if ambient behavior is enabled, it can proactively participate in the Slack channel. When users type @Claude in the channel, they can prompt Claude to take on a task, answer a question, or respond. Every user of the channel can see the responses just as they can see everything else in the channel.

Because Claude Tag lives in Slack, Anthropic is modeling it on the Slack environment, which is team-centric. It's a semi-public chat environment in that everyone in the channel can talk to Claude, and Claude can talk to everyone in the channel.

Multiplayer and agentic 

Anthropic calls this mode multiplayer. The idea is that there is a single instance of Claude interacting with the individual instances of each human in the channel. That one Claude instance will know what it said to Bob, what it said to Sue, and also what Bob said to Sue, and therefore can operate based on that common knowledge.

Claude Tag is agentic. It can break tasks into stages, work through them using the tools it has access to, and report back in a Slack thread.

Also: I tried Gmail's new Gemini Flows feature, and it's a huge filter improvement

When ambient behavior is enabled, Claude can sit and watch the entire channel context, learn from channel activity over time, and gather details and facts from wherever it's been granted permission to explore and work. Anthropic says it doesn't report from private channels.

According to the company, "Claude sets tasks for itself, allowing it to work autonomously." This is an asynchronous operation in that users can give Claude a task and then move on to other priorities while it continues to do its job behind the scenes. It's kind of like handing a team member an assignment in Slack and then waiting for them to come back later and tell you what they've accomplished.

This version of Claude learns over time, constantly keeping track of what is entered into Slack. Anthropic says, "This means that users don't need to explain things to Claude from scratch over and over again."

Claude can even act as a coach or a task master, depending on your perspective. If a thread has gone quiet or a task hasn't been completed by someone in the channel, it can follow up on those activities right in the channel. I'm not sure whether that intrigues me or kind of freaks me out. But I can definitely see the proactive productivity benefits for it, if it doesn't become too much of an annoyance.

Scoped permissions shape channel-specific access

Anthropic says that Claude Tag has been designed with teams and organizations in mind. The company is releasing it for Teams and Enterprise tiers only. The company says, "@Claude's access to sensitive data and task-specific tools can be very tightly scoped."

Also: Why Anthropic suddenly pulled Fable 5 and Mythos 5 for everyone

Each Slack channel gets its own Claude identity. Admins can scope Claude identities by channel and use case, and the individual Claudes don't share any information. This allows administrators to set up, for example, an @Claude for engineering and a @Claude for legal. The details between those two channels are not shared because each Claude is a separate identity and isolated from the other.

Any tools or information that Claude is allowed to access will also be scoped per channel. While one Claude might have been granted access to financial reports, for example, the Claude in Engineering's chat could be set up to not have any access to that data whatsoever.

As you might imagine, the second you start talking about multiple ambient Claudes, you start thinking about how much this will cost. If Claude is constantly "aware," it's going to be consuming tokens at a more prodigious pace.

Also: Why AI tokens will send your enterprise cloud bill sky-high again

However, administrators can set token spend limits for the organization overall, and for individual channels. While that won't limit Claude's consumption rate for tokens, it will limit the maximum spend, so financial types can at least predict spending levels for their new army of @Claude bots lurking in the company's Slack channels.

Anthropic is also enabling administrative oversight, so administrators can not only set the token limits but they can view a log of everything that Claude has done. They can even find out who requested the individual tasks. If Bob in accounting is asking for a farm subsidy report presented in conversational Klingon, the company's bosses will be aware of it. *Qapla'!*

Claude Tag is available today in beta for Claude Enterprise and Claude Team customers. Anthropic says its goal is to expand Claude Tag more widely over time, including beyond Slack. Claude Tag replaces the existing Claude in Slack app, and administrators can opt in within 30 days to migrate. Anthropic says it is issuing an introductory launch credit to eligible Enterprise and Team organizations.

Would you want an always-on Claude watching your team's Slack channels, or would you rather keep AI interactions one-on-one? Let us know in the comments below.


You can follow my day-to-day project updates on social media. Be sure to subscribe to my weekly update newsletter, and follow me on Twitter/X at @DavidGewirtz, on Facebook at Facebook.com/DavidGewirtz, on Instagram at Instagram.com/DavidGewirtz, on Bluesky at @DavidGewirtz.com, and on YouTube at YouTube.com/DavidGewirtzTV.

Featured