“Lobster” OpenClaw has completely exploded in popularity domestically, so much so that even Ma Huateng shared news about Tencent's free installation of the open-source AI entity OpenClaw in his social media circle, commenting “I didn't expect it to be this popular.”

During the recent national congress, Gao Wen, a representative of the National People's Congress and academician of the Chinese Academy of Engineering, also mentioned this phenomenon in his speech, saying “Everyone is so anxious, fearing they won't be able to raise lobsters.”

However, now that we have “lobster,” what exactly are people using it for? A relatively normal and ideal case shared might be as follows:

“For about the past week, I have been using a digital assistant that knows my name, understands my morning routine, knows how I like to use Notion and Todoist, and can control Spotify and my Sonos speaker, Philips Hue lights, as well as my Gmail. It runs on Anthropic’s Claude Opus 4.5 model, but I communicate with it through Telegram. I've named this assistant Navi, and Navi can even receive my voice messages and generate other audio messages in response using the latest ElevenLabs text-to-speech model. Oh, by the way, I haven't mentioned that Navi has a new feature for self-improvement and is running on my own M4 Mac Mini server?”

The author of the above case also mentioned that he has burned 180 million tokens on the Anthropic API, which probably cost around $2,000 to “raise lobsters.”

It might sound like the cost of “raising lobsters” is not low, and the things that can be done don’t seem that impressive; it feels more like a “XX spirit” that can interact with people in a human-like way and assist humans in completing more automated tasks. In fact, this is exactly the role that “lobster” can play at this stage — an “AI assistant.”

If we take a look at the top 100 installations on ClawHub and categorize them roughly, we can better understand that using large models for these tasks often feels like “using a sledgehammer to crack a nut”:

- Information retrieval types: Searching, extracting, integrating, and summarizing information from various sources (external links, local files, APIs). Actual use cases include having your Google or Baidu searches optimized and summarized by AI, getting daily weather reports from “lobster,” receiving real-time Bitcoin prices, and so on.

- Productivity (workflow automation) types: Handling emails, Notion, Github, Obsidian, Slack, etc., and further achieving cross-platform task automation, simplifying workflows with a single entry point to solve multi-platform issues.

- Developer tools: Professional tools aimed at developers and technical users, offering code management, API interaction, server management, and other functionalities. They can enhance development efficiency and achieve automation in code, testing, and deployment. This type resonates with programmers, such as interacting with GitHub using the command line, handling issues, PRs, CI runs, and advanced queries.

- Content creation types: Utilizing AI’s generative capabilities to create or edit multimedia content such as text, images, audio, etc.

- IoT control types: Connecting and controlling smart home devices, audio systems, and other smart hardware. For example, making the curtains and lights in the house turn on and off at specified times.

Overall, the hype around “lobster” is not because the above tasks can be executed excellently, but because it can comprehensively perform these tasks like a “secretary.” Compared to most users who may only use a certain AI tool purely as a search engine or automatic photo editing software, “lobster” allows people to issue various tasks to it via chat software like Telegram, much like how they would communicate with a boss. This novelty is magnified in the process of word-of-mouth, marking an unprecedented moment since AI entered people’s lives.

We can even take a more optimistic view on the current seemingly idle phase of “lobster.” In the early days of the iPhone, we could only play games like balancing balls, Angry Birds, or Fruit Ninja, which “demonstrated how to use the touchscreen.” In terms of the content and fun of the games themselves, they weren’t even as fun as the plethora of JAVA games on Nokia. But now, young people play Honor of Kings, Delta Force, and many individuals exclusively play mobile games and not PC games.

If we shift our gaze to today’s cryptocurrency market, “lobster” may once again significantly lower the learning barrier between cryptocurrency and the public, effectively addressing the widespread investment needs of the public.

What is referred to here is certainly not speculating on memes or issuing coins through “lobster.” Nowadays, the range of asset classes available for trading on-chain is becoming increasingly rich: US stocks, crude oil, gold, Pokémon cards... we can engage in decentralized, round-the-clock trading without barriers on-chain. This trading volume is considerable; on February 6, Hyperliquid's on-chain Perp DEX, mainly trading US stocks, had a 24-hour transaction volume of $5.45 billion, reaching a historical high.

In an era of abundant information, what often prevents us from seizing new investment opportunities is “not knowing how to get in.” For instance, not long ago, memory prices surged; while everyone knew this information, directly purchasing stocks of SK Hynix could be quite troublesome for non-Korean nationals. Opening accounts, settling funds, etc., obstructed the public from making immediate investment moves in response to the information.

But if the path becomes:

- Allowing “lobster” to have a wallet

- Purchasing stablecoins with a credit card to fund the “lobster” wallet

- Telling “lobster” the specific investment target

- “Lobster” completing the buy/sell action on-chain

And all of these are accomplished simply through chatting as if with a friend, this undoubtedly presents an explosive growth opportunity for “lobster” and cryptocurrency.

We also have prediction markets, so we can imagine even more. For example, while taking a taxi, chatting with the driver. The driver says he believes the next President of the United States will be A, while you believe it will be B. In a heated debate, you use voice-to-text to issue a command to your “lobster” — help me place a $100 bet that B will win.

Your “lobster” understands your intent, automatically searching for the best liquidity in prediction markets to place the order. The driver immediately follows suit, using the vehicle’s voice command system to also bet $100 on A winning through “lobster” in the prediction market.

There’s even the possibility that “lobster” will need to implement controls on underage consumption to prevent children from impulsively spending in the on-chain card market while showing off their Pokémon cards to each other through “lobster.”

If the meme coin trend caused by pump.fun's “everything tokenized” is the 1.0 version of attention economics, then the new paradigm of ordinary people using AI more easily through “lobster” could become the 2.0 version — it could find everything we want to invest in instantly on-chain, including channels, and execute according to our intent. Moreover, it would allow the on-chain ecosystem to expand from investment to consumption, truly unlocking the mass adoption that blockchain has pursued for years.

The future is happening.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。