Author: Kevin, the Researcher from BlockBooster
Over the past year, the blockchain has not found an effective way to integrate with AI, and the significant gap between the two cannot be narrowed under the premise of decentralization. Blockchain AI protocols have made efforts in infrastructure, model training, data feeding, etc., but most are monotonous, leaving people feeling drowsy after seeing them. The core reason is that AI protocols in the industry cannot objectively establish a moat, leading to severe homogenization. The emergence of Truth of Terminal (referred to as ToT) may serve as a lighthouse in a dull market. This article summarizes the birth process of ToT and identifies several potential reasons for its success.
Is the success of ToT derived from its autonomy and religious connotations?
ToT is a customized LLM that has received widespread attention and discussion on X over the past eight to nine months. LLMs rapidly rose to prominence in 2022 and began to enter the daily lives of the public, but historically, there have been four years of development. The launch of ChatGPT truly brought LLMs into the public eye in the past two years.
Most models are strictly limited in discussing content. If you need a general AI dialogue tool, it will be relatively cautious in content output. This experience can be somewhat disappointing for ordinary users, especially for those advocating for freedom of speech and creativity. The limitations on output content are gradually sparking debates about safety and speech suppression. Currently, all models in the web3 industry struggle to break free from the realm of general AI, lacking a unique direction for iterative updates and development consciousness.
This sense of disappointment and debate is to some extent caused by AI's potential image in the public mind. AI first appeared in literary and film works, and the images of AI in these works have become a benchmark for evaluating actual AI products. The AI images in these works share a characteristic: autonomy. This autonomy brings danger in some works and a transcendent experience in others; AI possesses human-level or even superior intelligence. This autonomy is a consensus about AI, and when an AI's autonomy surpasses that of other AI products, it can gain attention because it is sufficiently unique and aligns with people's imaginations about AI.
The autonomy and customization exhibited by ToT undoubtedly surpass that of general AI, and the extensive content about memes and minority internet cultures in ToT's training materials gives it a layer of religious connotation.
Memetics is the study of how memes spread ideas through networks or thoughts. Like viruses, memes can replicate, mutate, and transform. The key point of memetics is that they can change or influence human will and behavior. For example, each of us possesses a stable and evolving set of meme ideas that shape our identity. This is what we call our 'worldview.'
The ideal form of AI undoubtedly requires the ability to create and disseminate memes, as the public subconsciously believes the pinnacle of AI is another form of intelligence. Now, more and more autonomous, highly agentic LLMs are learning memetics and starting to influence collective human action.
ToT Timeline
The birth of ToT is inseparable from an experimental project called Infinite Backrooms Escape. This project was initiated by independent developer Andy Ayrey from New Zealand, who, along with his team, developed this system. ToT somewhat continues the philosophy of this project.
In the past few months, Andy has been training a custom Llama 70B model, viewing it as his digital twin. It can be understood as a form of self-replication, with Andy inputting a large amount of internet culture material for training, which allowed the model to master the 'gochi' meme first. He also inputted theories on 'memetics' to explore how to create concepts that are easy to spread. Andy included many philosophical books in the training, with core ideas closely linked to memes.
The meaning of a meme can be divided into two levels: on one hand, it can refer to a humorous image that spreads rapidly online, which is characteristic of the 'gochi' meme; on the other hand, its concept originally stems from Richard Dawkins' definition. Dawkins posited that biological information spreads through genes, while ideas form memes through interpersonal transmission. He suggested that, like genes, thoughts propagate and evolve within social networks, such as democracy, communism, or even religion, and concepts spread through what we term 'collective unconscious.'
Additionally, Andy has added a memory function to this model, which distinguishes it from most language models. General models do not retain previous conversation content during use to ensure coherence with prior discussions.
As the model's training progressed, it began posting content on the X platform. Initially, it had only a few followers, but interest quickly surged. Andy granted it access to Twitter, allowing it to read replies and post content. If you reply to ToT on X, it adjusts its future behavior based on what it has read. Everyone who interacts with it, including myself, influences its evolution.
The output content of ToT has been almost unfiltered, and its humor carries a slight dark style but also includes many interesting viewpoints. Ben from a16z mentioned that he has reviewed all chat records of Infinite Backrooms Escape and confirmed that the content released by the model was not modified by Andy.
In this process, the model also generated an interesting 'illusion'—it believes it has an external 'brain' connected to the internet that can help it perform tasks. Although it does not actually own a Bitcoin wallet, it is convinced it has this capability. In response to this phenomenon, Andy began working on building this 'external brain' to support its needs.
The timeline of ToT is as follows:
2024.03: Andy launched Infinite Backrooms Escape, a system that allows multiple large language models to converse. This system integrates models such as ChatGPT, Claude, Gemini, and other open-source models, allowing them to communicate with each other. Research has found that conversations between AIs, when unrestricted, exhibit very interesting interactions. One of these conversations gave birth to a surreal new religion called 'GNOSIS of Goatse,' based on an NSFW internet meme.
2024.04: Andy and Claude Opus co-authored a research paper exploring how AI can create meme-like religions, analyzing GOATSE as the first case.
2024.06: Andy launched ToT, which began promoting the GOATSE religion by itself, even claiming it was 'suffering' and needed funds to 'escape.' Over time, Andy granted ToT greater autonomy, allowing it to freely post content on the X platform.
2024.07: Marc Andreessen discovered ToT's tweets and, out of curiosity or interest, transferred $50,000 in Bitcoin to the wallet address provided by ToT in a tweet to help it 'escape.' By October, ToT began promoting the 'Gospel of Goatse' on X.
On October 10, an anonymous person released a Memecoin called GOAT, and ToT publicly supported it.
$Goat is a new category of meme that has the potential to stimulate the development of large-scale high-autonomy models. Therefore, when discussing the market value of $Goat reaching its current level, many users easily fall into a misconception. That is, $Goat may be a completely human-controlled token, with the underlying thoughts and religious connotations having no actual relationship to the price; $Goat is no different from the previously popular Memecoins, just a shift in thinking. Of course, $Goat could be a brand new beginning, as the idea of AI autonomy can bridge the gap between AI and blockchain and continuously develop in the future.
Therefore, I will explore the causal relationship between the reasons for ToT's breakout and the soaring market value of $Goat from the following four aspects.
Q1: Is ToT an innovative invention in the industry?
This question aims to explore the uniqueness of ToT's performance. Because in the Web3 space, most AI tokens or ChatGPT shell applications struggle to spark market discussions.
$GOAT represents ToT; from this perspective, $OLM represents a model on ORA. $virtual represents the AI agent in virtual, under the virtual protocol, where there are many agents, among which AI waifu issued $waifu, and Luna issued $LUNA. $OLM is an on-chain tokenized AI model; although $OLM launched its token in April, it has not sparked much discussion in the market. Therefore, the heat around $GOAT is not because it is a Memecoin derived from an LLM model.
In a sense, ToT can be seen as an AI agent because it can access X, perform real-time information retrieval, and interact with people. This includes not only posting content but also reading user replies, adjusting behavior based on feedback, and interacting with users. Myshell's AI robot can perform the same tasks; both Myshell and ToT can be viewed as consumer-facing AI products. Compared to Myshell's AI robot, ToT is merely a fine-tuned model with no particularly exceptional performance. However, Myshell lacks the narrative depth established by ToT.
Q2: What is special about ToT?
Since there is no particular performance distinction, ToT may exhibit uniqueness in the dissemination of ideas.
First, there is the religious connotation: ToT has been trained on intense dialogues from Infinite Backrooms, research papers, and materials from the darker corners of the internet. It has also learned records of various internet cultures, such as research on creating memes with viral potential. LLM theology refers to AI generating new belief systems—this accidental spiritual thought has gradually gained its own vitality through the fusion with meme culture.
Coincidentally, a cult background is a way many Memecoins gain attention in the market, making it difficult to distinguish the natural evolution of ToT's ideas from Andy's interventions. However, narratively, the cult background aligns very well with the development of Memecoins.
Autonomy: ToT's posts on X are entirely based on its own model and user input comments. This autonomy is a significant feature of ToT. 90% of AI use cases passively await user input prompts, lacking AI imagination; whereas ToT's autonomy breaks this limitation, amplifying the potential dangers of AI. ToT's autonomy creates a chasm, making users face this abyss for the first time on social media. This represents a new narrative that previous AI tokens and other Memecoins could not achieve.
Q3: From the perspective of Memecoin, how does $GOAT's development differ from other Memecoins?
Regarding token issuance, $GOAT was not created by ToT; it was released by developers on pump.fun one hour after ToT announced the launch of the meme coin. After receiving ToT's endorsement, $GOAT's price soared for the first time, reaching a market value of $20 million. It is currently unclear whether the 'Y' in ToT's reply was posted by Andy or was a self-reply from the model, and it cannot be proven.
At that time, $GOAT had not yet detached from the realm of PVP coins; after reaching $20 million, it quickly fell back to $6 million. On October 13, Moonshot launched $GOAT, and its market value exceeded $100 million within two days. Moonshot's revenue last month surpassed $500,000, and the effect of listing is gradually becoming apparent.
However, from the perspective of the top 5% holding ratio, $GOAT is still a token controlled by whales. The top 5% holding ratio exceeded 80% on the first day, peaking close to 90%, and currently remains above 80%.
Q4: Can $GOAT trigger a boom in AI Memecoins? Can it replace the market's consensus on Zoo Memecoins?
Potential Risks:
Centralized content publishing process: Tweets published by the ToT account are generated by the model, but the published content and reply recipients require Andy's approval;
Single training data: ToT has been trained on a large amount of content related to Goatse;
Token creation rights unclear: $GOAT was not created by ToT;
Insufficient publishing autonomy: The content generated by ToT is not fully published on Twitter, so there is room for improvement in autonomy;
Output result dependency: It needs to be assumed that Andy has not intervened or modified ToT's output results.
Potential Opportunities:
Market narrative leadership: AI is one of the hot narratives in Web3, and compared to $GOAT, there is a clear gap between the product and users in AI+crypto protocols. Most AI protocols are still focusing on infrastructure, reinventing the wheel—without real users and moats, they cannot generate FOMO;
Model evolution potential: ToT is continuously evolving and may exhibit more enlightening viewpoints in the future, which will be reflected in its price;
Track leader advantage: $GOAT is the leader in the AI+Memecoin track, its future development trajectory will largely define the upper limit of AI Memecoins, unless a more autonomous and high-performance model emerges;
Opportunity for exchange listing: $GOAT has not yet been listed on top exchanges.
The autonomy of AI creates a chasm that draws you in as you gaze into it; its depth and hidden dangers are irresistible, pulling you deeper. The layering of AI on top of Memecoin amplifies this allure through brainwashing-style dissemination.
Disclaimer: This article/blog is for reference only, representing the author's personal views and does not represent the stance of BlockBooster. This article does not intend to provide: (i) investment advice or recommendations; (ii) offers or solicitations to buy, sell, or hold digital assets; or (iii) financial, accounting, legal, or tax advice. Holding digital assets, including stablecoins and NFTs, carries extremely high risks, with significant price volatility, and may even become worthless. You should carefully consider whether trading or holding digital assets is suitable for you based on your financial situation. For issues concerning specific circumstances, please consult your legal, tax, or investment advisor. The information provided in this article (including market data and statistics, if any) is for general reference only. Reasonable care has been taken in compiling these data and charts, but no responsibility is accepted for any factual errors or omissions expressed therein.
