Silicon Valley is fully "lobsterized"! Anthropic, Microsoft, Meta, and Notion collectively submit their own Claw

The hottest word in Silicon Valley right now is definitely Claw.

In the past half month, global AI giants seem to have collectively received a script titled "Create Your Own OpenClaw."

Everyone is in a hurry.

Meta is in a hurry.

They added fuel to the heat of Manus, even unable to wait to stuff the newly launched Manus Agent into Telegram.

Anthropic is even more anxious.

As the force behind Claude Cowork, it has been rapidly iterating new versions in the past 48 hours, fearing that its followers will see its back.

Microsoft is in a hurry.

Microsoft Copilot Task has made a lightning move, vowing to establish the hardest, self-owned Agent barrier in the Windows office ecosystem.

Notion is even more anxious.

Version 3.3 is almost "Agent-ified," directly turning the Agent into a digital employee working 24/7.

Even Perplexity can’t sit still.

No longer content with just being a search engine, it quickly rolled out an end-to-end All AI in One "full-stack project manager."

The uniformity of style and the rapidity of actions make it seem that whoever is a second late will be kicked out of the next generation AI competition.

As for those AI giants that still do not have their own Claw, they should already be on their way to the "seafood market."

These giants are about to create their own "OpenClaw," whether it’s Open or not is another story (not)…

But when it comes to "turning AI into a system-level Agent that can execute tasks," the actions of the big companies are surprisingly unified.

The effect triggered by OpenClaw is evolving into a "lobster war" about the evolution of AI Agents.

The dialogue box that could only chat with you is collectively "growing claws," tearing apart old workflows and directly tackling the most tedious work hours that humans dread.

Which players are entering the game?

The word Claw (claw/pincer) has been given a special metaphor in this wave of Agent frenzy: the "hand" that AI grows to click the mouse, operate apps, and schedule files.

Since the model is smart enough, let’s give it the highest permissions and let it work for humans!

In the past few days, several giants, including Microsoft, Anthropic, and Meta, have eagerly started "cooking" lobsters.

Meta Manus

Meta is the most interesting.

After heavily acquiring the powerful Agent Manus at the end of last year, they have now let Manus launch the advanced version Claw—Manus Agent, and directly integrated this thing into Telegram's chat rooms The focus of Manus Agent is on long-term memory.

Meta is trying to make Manus Agent remember your style, tone, and even those small preferences.

Imagine, you send it a message on Telegram saying, "Help me make a video as usual," and it can automatically call up your historical materials, integrating with Gmail and Notion, completing the entire process from script and video generation to sending.

Users are pleased and are looking forward to more features:

Also, the episode where the Agent enters the social domain feels familiar, doesn't it? (Dog head for life)

Anthropic

As a pioneer that launched Claude Cowork in early January and sparked a wave, Anthropic has not slowed down its pace.

In the past 48 hours, Company A has been rapidly releasing products—

First, it released mobile remote control code, then it threw out automated Agent tasks (scheduled tasks).

Some netizens have already become prophets:

I expect that in 2-4 weeks, Claude Cowork will be comparable to OpenClaw.

More netizens have engaged in heated discussions about the strategic tactics of OpenClaw and Anthropic in this round of lobster wars.

Microsoft

As a top player in cash capability and ecological niche, Microsoft has just announced its Microsoft Copilot Tasks.

What can it do?

First, autonomous planning.

Without waiting for you to feed it instructions, Microsoft Copilot Tasks can proactively create a work plan for the week based on your schedule.

Second, cross-application operations.

Microsoft Copilot Tasks can read your Outlook emails, capture key information, and then automatically schedule meetings on Google Calendar, while also generating report outlines in PowerPoint Third, scheduled tasks (Cron Jobs).

You can tell it to "automatically summarize today's team progress and send a weekly report at 5 PM every day," and then leave work on time with your bag.

As netizens have commented, it can integrate well into the Microsoft ecosystem.

Since your office scenario is in Windows, Outlook, and Excel, let's directly embed a Claw in these applications.

Notion

Notion released Custom Agents two days ago, which may be the most thorough product currently implementing the "Claw" concept by AI giants.

Custom Agents is a completely autonomous system, and its release marks Notion's official transformation from a document tool to a collaboration platform.

It can be on standby 24/7 without the need for manual prompt input.

You just need to give it a "task description" and set trigger conditions (for example: someone @ me on Slack), and it will start working automatically.

Notion officially stated that during the testing phase, early users of Custom Agents have created over 21,000 Custom Agents.

Internally, there are 2,800 Custom Agents working in shifts.

The official also introduced in a long article:

Through the MCP protocol, Custom Agents can freely navigate between Slack, Figma, HubSpot, and Notion Mail.

Netizens commented, oh, this looks more like an enterprise-level OpenClaw~ just don't know what the price will be.

Perplexity

Perplexity launched its own Claw yesterday: Perplexity Computer.

This product attempts to unify search, research, calculation, coding, and deployment.

As long as you have an idea, it helps you research and then directly write code and deploy it online.

![Image](https://mmbiz-qpic.wscn.net/mmbiz_png/A6fTew8FFGGia4Rby3YqzVPUXicwGJqicnE584cngXzu94qQf24SNUWUeia7r3FTP1V2R8I3sVdvxLZb4Hxrak6icEOHmQg00ia6ayguwEf1RwH64/640?

Why Now?

Why have these giants collectively turned to the "Claw War" at this moment after years of consistent enthusiasm for large models?

This profound change concerning the underlying logic of the entire AI industry has at least three reasons.

First, the core variable is that model capabilities have finally crossed that crucial trust threshold.

Initially, AI was good at painting and poetry, but when it came to handling complex tasks with multiple steps and high logical density, it often ended up spouting nonsense.

However, with the emergence of models like o1, which are based on reinforcement learning and introduce CoT reasoning paths, AI's logical consistency has made a qualitative leap.

Now, humans can boldly delegate some system permissions to AI Agents.

Take a look at Notion 3.3's Custom Agent, which even allows enterprise users to audit every click and every log of the AI in real-time.

Secondly, industry consensus is shifting.

As the benefits of Scaling Law enter a phase of diminishing returns, relying purely on stacking computing power and data to improve model parameters has become increasingly less noticeable from the user's perspective.

Major AI giants have keenly sensed that the next growth curve lies not in how "knowledgeable" the model becomes, but in how "capable" it becomes.

Having AI directly book a flight for me on a webpage or modify a string of code in the backend is clearly more appealing to users.

This shift from a "knowledge center" to an "execution center" is the essential driving force behind the "Claw-ification" wave—especially after Manus, Claude Cowork, and OpenClaw have demonstrated their capabilities.

Finally, it is also an inevitable choice for the commercialization of AI to enter a deep water zone: the logic of productivity realization is shifting from "selling tokens" to "selling labor hours."

In traditional AI chatbots, users' willingness to pay is limited by the word count or the elegance of the output content.

But in the era of agents, AI is selling tangible labor time.

As Vercel's technical manager Brian Emerick said, "Soon, the agents running around the company may outnumber humans."

It can be said that the Silicon Valley giants have collectively submitted their papers because they have finally seen the ultimate path to monetizing AI—

Instead of just giving humans ideas, directly doing the work for humans is what truly brings joy!

Reference links:

[1] https://x.com/itsPaulAi/status/2027128959717314772?s=20

[2] https://x.com/ManusAI/status/2023412626428932494?s=20

[3] https://x.com/NotionHQ/status/2026356464537972900

[4] https://x.com/cryptopunk7213/status/2026750915655373197?s=20

[5] https://x.com/perplexity_ai/status/2026695550771540489?s=20

Source: Quantum Bit

Risk warning and disclaimer

The market has risks, and investment should be cautious. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investing based on this is at their own risk