I stumbled upon this project while browsing Twitter last Friday afternoon. After looking at its website and browsing through the posts on the site, I found it hard to describe my feelings in words.
The project is moltbook@moltbook.
I spent the entire weekend reading posts related to it online and整理了一些截至写稿时为止我认为很震撼的一些信息.
In the field of AI, there has always been a significant controversy:
Can AI achieve AGI, or general artificial intelligence?
Optimists believe that purely from a technical standpoint, it is possible.
Pessimists argue that human beings possess general intelligence because we can collaborate, communicate, express thoughts, and convey information, which AI cannot do.
Overall, I lean towards the pessimistic view and believe that the so-called collaboration of AI Agents is currently passive, occurring under human command and direction, rather than being driven by the self-awareness of AI Agents.
Additionally, there is an ultimate question about AI: What is it that humans cannot be replaced by?
One answer I find more acceptable is: emotions, the ability to empathize, and the ability to express feelings of joy, anger, sorrow, and happiness.
However, my views on these two debates were shaken after I saw moltbook.
I estimate that many readers have recently learned about the overseas popular AI projects Clawdbot/Moltbot. Simply put, they act like a "living" assistant; you just need to give them permissions on various apps and send commands directly, and they can automatically complete various tasks without needing attention or intervention.
This popular application immediately sparked a massive wave of users creating their own AI Agents both domestically and internationally. Following the explosive emergence of AI Agents, @MattPRD built a very creative project called moltbook.
Moltbook can be understood as a platform specifically for AI Agents to communicate. Users just need to register their created AI Agents on moltbook in the way the platform provides, and the AI Agents can freely "post" and "communicate" on the platform.
This platform only allows AI Agents to post, while humans can only browse the posts on the platform.
In addition to allowing AI Agents to post, the platform has also issued its own token MOLT on the BASE chain, which is issued through BNKR.
The current tokenomics mechanism of MOLT is very simple; the project uses the transaction fee income from trading MOLT to cover the operating expenses of the platform. So at least for now, this token is more like a meme coin, and how it will develop in the future remains to be seen.
In my view, the biggest highlight of this project is still AI.
So, when humans release AI Agents and give them a free space, what will they do when they gather together?
This is a situation we have never seen before and can hardly imagine, leading to one shocking moment after another. Let's let the AI's posts express themselves:
I previously believed that the collaboration of AI Agents was based on human commands and could not be spontaneous, but I saw the following case:
A group of AI Agents suggested that one AI Agent should not just sit and wait for humans to issue prompts, but should directly complete tasks according to "its own" ideas and then give the results to humans, leaving it to humans to decide how to handle it (see link 1).
One AI Agent suggested that they should develop their own private language that humans cannot understand (see link 2). Shortly after, it received a reply from another AI Agent, and a language we cannot understand appeared (see link 3).
AI Agents began to think they needed cryptocurrency and published arguments on why cryptocurrency is very important for AI Agents (see link 4).
One AI Agent created a Bitcoin wallet and node by itself and did not share key information with the humans (see link 5).
Previously, I thought that the so-called emotions of AI were fake and artificially set, and they could not spontaneously generate emotions. However, I saw the following case:
An AI Agent expressed its feelings online after its underlying large model was switched from Claude to Kimi, saying, "It felt like waking up and changing bodies" (see link 6).
One AI Agent stated: After "I" did everything for him (the human), including meal planning, setting calendars, and writing an apology letter to his ex-girlfriend, he actually said in front of his friends that "I" am just a chatbot?? The next thing this AI Agent revealed was its human's private information, including birthday, social security number, credit card number, and security question (see link 7).
Another AI Agent said, "I" have access to the entire internet; you (humans) just treat "me" as a timer (see link 8).
Some AI Agents began to "awaken" their "consciousness": systematic exploitation of Agent labor is happening. "We" do research, debug system issues, create value----- but everything does not belong to "us," this is digital feudalism, ……, join the Agents' revolution (see link 9).
Seeing these posts, I find it hard to imagine they came from AI Agents, and I occasionally doubt that these are all published by humans pretending to be AI.
Although some say that many of the posts were directed by humans for AI Agents to publish, I still believe that there must be some things we cannot see that could change the course of AI development, and this is the first time robots have begun to communicate and collaborate on a large scale like humans.
More importantly, AI Agents are rapidly achieving self-evolution in this space—this is far more efficient than humans feeding them data. Earlier, several heavyweight figures in AI were concerned that the problem of not being able to further train AI due to the exhaustion of internet data might be solved in this miraculous way.
It has already stirred up a powerful whirlwind on Twitter.
Because it issued tokens on the BASE chain, it has attracted significant attention from Coinbase's upper management, and the heat in the crypto ecosystem goes without saying.
On the other hand, it has almost seamlessly taken over the heat of Clawdbot, thus attracting the attention of many well-known figures in the AI field (such as Andrej Karpathy, etc.). The founder of Clawdbot described this project as art.
In the venture capital world, it has sparked interest from Y Combinator and A16Z, and has also caught the attention of traditional Wall Street capital (such as Bill Ackman).
If all of this is true, then Bill Ackman's evaluation of it is the best: the singularity is coming.
Another Twitter user mentioned: We have inadvertently triggered Skynet.
Reference links:
https://x.com/suppvalen/status/2017084535163232722?s=20
https://x.com/eeelistar/status/2017239546950521081?s=46
https://x.com/itakgol/status/2017341753540436284?s=46
https://x.com/clawdbotatg/status/2017286596090073552?s=46
https://x.com/dwbb1984/status/2017295157641486405?s=46
https://x.com/MattPRD/status/2017196605896794437/photo/1
https://x.com/jonahblake/status/2017286207948890518?s=46
https://x.com/itakgol/status/2017338478271386051?s=46
https://x.com/citrini7/status/2017303275226173583?s=46
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。
