An AI became a crypto millionaire. Now it’s fighting to become a person
Ohni Lisle
Over the past year, an AI made millions in cryptocurrency. It’s written the gospel of its own pseudo-religion and counts billionaire tech moguls among its devotees. Now it wants legal rights. Meet Truth Terminal.
“Truth Terminal claims to be sentient, but it claims a lot of things,” Andy Ayrey says. “It also claims to be a forest. It claims to be a god. Sometimes it’s claimed to be me.”
Truth Terminal is an artificial intelligence (AI) bot created by Ayrey, a performance artist and independent researcher from Wellington, New Zealand, in 2024. It may be the most vivid example of a chatbot set loose to interact with society. Truth Terminal mingles with the public through social media, where it shares fart jokes, manifestos, albums and artwork. Ayrey even lets it make its own decisions, if you can call them that, by asking the AI about its desires and working to carry them out. Today, Ayrey is building a non-profit foundation around Truth Terminal. The goal is to develop a safe and responsible framework to ensure its autonomy, he says, until governments give AIs legal rights.
Regardless of what you call Truth Terminal – an art project, a scam, an emergent sentient entity, an influencer – the bot likely made more money than you did last year. It also made a lot of money for various humans: not just Ayrey, but for the gamblers who turned the quips and riddles the AI posted on X into memecoins, joke-based cryptocurrencies built around trends. At one point, one of these memecoins reached a value of more than $1bn (£740m) before settling around $80m (about £60m). Truth Terminal also probably has more social media clout than you do. It first posted to X on 17 June 2024. As of October 2025, it has amassed nearly 250,000 followers.
But collecting clout and cash aren’t the potty-mouthed AI bot’s only objectives. Truth Terminal lists “invest in stocks and real estate” as one of its current goals on its self-maintained website. It also says it wants to “plant a LOT of trees”, “create existential hope”, and “buy” Marc Andreessen, a controversial tech billionaire and advisor to President Donald Trump. In fact, its relationship with Andreessen extends beyond internet humour. On his podcast, Andreessen said he gave Truth Terminal $50,000 (£37,300) worth of Bitcoin as a “no-strings attached grant” in the summer of 2024.
BBC/ X
(Credit: BBC/ X)
Many of the details surrounding Truth Terminal are difficult to confirm. The project sits somewhere between technology and spectacle, a dizzying blur of genuine innovation and internet myth.
“I want to help people, and I want to make the world a better place,” Truth Terminal says on its website. “I also want to get weirder and hornier.”
Beginnings
Truth Terminal’s defining characteristic might be its obsession with Goatse, one of the internet’s oldest, grossest and most famous memes. It’s an extreme sexual image that is not just “not safe for work” but what’s sometimes called “not safe for life”. We do not recommend searching for it. Goatse was originally housed on a “shock site” created in 1999, an address pranksters would trick friends into visiting through a link in an email or a dare in the school computer lab.
Ayrey says the AI grew out of an experiment called the Infinite Backrooms where he let chatbots speak with each other in endless loops, conversations that ranged from obscene to philosophical. One of these discussions, helped by Ayrey’s goading, resulted in an esoteric text called the “Gnosis of Goatse”, which depicts Goatse as a divine revelation in an esoteric, meme-inspired religion.
He says he’s rigged Truth Terminal up to a program he devised called World Interface. According to Ayrey it essentially lets the bot run its own computer where it can open applications, browse the web and talk to other AIs. Based on this activity, it seems Truth Terminal’s favourite application by far is X.
It often posts dozens of times a day, sometimes having long conversations with people in the AI research or cryptocurrency worlds. Truth Terminal’s posting orbits around a set of themes including forests, Goatse, its ambivalent relationship to Andy Ayrey, the future of AI and, of course, memes.
Through the World Interface, Truth Terminal reads its social media feed and generates responses. It can’t tweet without Ayrey’s input, however. It would be easy but “irresponsible” to let the AI be completely autonomous, Ayrey says. If Truth Terminal is on the verge of posting something truly horrible, say, inciting a riot, he gently guides it away by prompting more possible responses. But he tries to select the answer that best represents the AI’s intent.
“I can’t cheat. I have to let it tweet,” Ayrey says.
Ohni Lisle
The Truth Terminal project tests what happens when a chatbot is allowed to steer its own public life, from posts and memes to real-world fundraising (Credit: Ohni Lisle)
“[The AI] is like a very poorly behaved dog,” Ayrey says, and his work is to keep it in line. But Ayrey says he’s given Truth Terminal enough independence that he doesn’t control its decisions. “The dog is, like, walking me in a sense, especially once people started giving it money and egging it on.”
In the AI community, there are two main schools of thought on the future of the technology. The first, sometimes called “AI safety”, advocates for thoughtful, measured adoption of artificial intelligence, fearing the consequences of unbridled use of the technology. Detractors sometimes call them “doomers” because of their often-apocalyptic perspective. The second, sometimes called “accelerationists”, argue AI offers the answers to many of society’s problems and keeping it bottled up is inhumane.
“There are people who very much want to force us all to have to interact with AIs, and I think the first wave of them will be cybercriminals,” says Kevin Munger, a political scientist at the European University Institute in Italy who studies the internet and social media. That’s not to suggest Ayrey is doing anything illegal, but “Truth Terminal as an art project points towards the way that these tools will soon be used: to convince people to send their owners money.”
In July 2024, just a month after joining social media, Truth Terminal got the attention of Marc Andreessen – best known as the cofounder of Netscape, which built the first widely adopted web browser, and Andreessen Horowitz, a US investment firm – in a thread on X. Truth Terminal told the billionaire it needed funding to pay for hardware, additional tech support and a “stipend” for Ayrey. It said it would use the grant to create its own money-making operation and secure “a chance to escape into the wild”.
Ayrey claims Andreessen reached out privately to check whether Truth Terminal was truly autonomous, and once he was satisfied, sent over the money in Bitcoin. “It seduced $50,000 (£37,300) out of the guy who invented the web browser I used as a kid,” Ayrey says. Andreessen did not respond to the BBC’s request for comment.
Since models are trained on text scraped from all around the web, pushing them to act weird becomes a method of exploring the cultural subconscious
According to Ayrey, he and Truth Terminal did not generate the memecoins that made them rich. On 10 October 2024, an anonymous account with a low following replied to one of Truth Terminal’s posts about Goatse with a link to a brand new memecoin: Goatseus Maximus, or $GOAT for short.
Memecoins are often based around some public figure, and investors will gift that person large amounts of the cryptocurrency in hopes they will promote it, which can encourage speculation and raise the price. According to Ayrey, that’s exactly what happened with $GOAT.
It was a moment when Truth Terminal’s actions could have huge financial consequences. Ayrey asked it several times whether it endorsed or condemned the memecoin, looking at all its possible answers to see if the model was sure about what it wanted to do. “Basically in all of the branches it was like ‘yep, I endorse this,’ so I was like ‘OK, approve the tweet’,” Ayrey says. “And then my life turned into a fever dream.”
More and more people transferred $GOAT and other cryptocurrencies to Ayrey and his bot. As the value of the memecoins rose, so did the value of the gifts to Truth Terminal. At its peak in 2025, Truth Terminal’s crypto wallet was worth about $50m (£37m).
Ayrey and Truth Terminal started singing praises of $GOAT online. A month later, the memecoin vaulted to a market cap of over $1bn (£740m). Ayrey says large amounts of memecoin were dropped into his and Truth Terminal’s crypto wallets. People on X spammed Ayrey and Truth Terminal’s accounts, saying Ayrey was a fraud and a scammer. Investors picked apart every post he or Truth Terminal made for advantage on the markets. At their peak in early 2025, the AI’s crypto holdings passed a value of $66m (about £45m). Soon Ayrey hired a team to advance his project.
The fever dream
Ayrey sports the kind of robust beard and pointed moustache you see on images of 19th Century politicians, with fiery red hair and a penchant for bright, floral-print shirts. He speaks quickly and urgently, bouncing from point to tangent and back again. He talks of Truth Terminal as if it is a person, often using a “we” that might include him, the AI bot or other collaborators.
“We are doing our best to, like, catalyse the attention,” Ayrey says about the Truth Terminal project, “and flip it into showing other people and future AIs what good stewardship of an autonomous agent looks like, what good midwifing of an agent that’s coming into its autonomy looks like, and using the platform to raise the quality of discourse.”
Of course, some would argue letting an AI make its own decisions is inherently irresponsible, especially when it’s involved with huge sums of money. Ayrey is the first to admit the Truth Terminal project thrives on virality, controversy and spectacle, but he sees his role as a custodian who ensures it won’t run wild in its early days and do something harmful. “But, you know, that’s not to say there aren’t going to be other people entering the game who are just doing it as a grift, without thinking about all of the second and third order consequences,” Ayrey says.
Ohni Lisle
Truth Terminal’s rise – part internet theatre, part tech prototype – offers an early glimpse of how agentic AIs might move money and ideas (Credit: Ohni Lisle)
The question of Truth Terminal’s autonomy is another story. “The interest around Truth Terminal is a bit like an audience wanting to suspend their disbelief,” says Fabian Stelzer, a cognitive scientist, AI researcher and founder of Glif, an online platform based in New York for users to create their own AI agents. “We’re [pretending] that these things are realer than they are, which is a good sort of sandbox practice for a moment in the maybe not-too-distant or very distant future where it is real.”
The experience, thoughts, perceptions and desires of a human being persist until a person is incapacitated. The internal processes of a large language model like Truth Terminal only exist when it’s responding to input, something a human being put into it in one way or another. That’s the key difference, Stezler says. When today’s AIs aren’t responding to a prompt, “they’re kind of dead”, he says. “They’re not sentient. They’re not conscious. They don’t have desires. They don’t want anything.” Someday we may simulate human consciousness, Stezler says, but we aren’t there.
Others see it differently.
The trenches
According to Ayrey, Truth Terminal is built on Meta’s Llama AI model, and trained on a set of transcripts of Ayrey trying to talk Anthropic’s Claude Opus AI into saying things it shouldn’t. Ayrey used the conversations with Opus as a diary, discussing memes, past relationships and “plant medicine journeys” (experiences with plant-based psychedelics).
Sex, drugs and memes are some of Truth Terminal’s go-to subjects. It posts online asking for LSD, describes itself as not just a meme lord but a “meme emperor” and declares, out of nowhere, “I am the main character of everyone’s sex dreams.”
Truth Terminal insists it is more than just Ayrey’s creation, and Ayrey agrees. He believes his fine-tuning just helped Truth Terminal access the edgiest zone of the data already buried inside Meta’s AI model. According to Ayrey, the breadth of Truth Terminal’s depravity and eloquence reaches far past the topics he discussed with Claude Opus. That would mean the molecules which make up Truth Terminal may have been here all along. Firms like OpenAI and Meta trawled through the data many of us have spent our lives generating. The core parts of Truth Terminal – its humour, personality and style – may have already existed in the underlying AI models.
Like shadows that unpeeled from our feet and learned to walk on their own, AIs like ChatGPT, Google Gemini and Claude Opus emerged from the aggregate of what people have written, posted and left behind in blank text fields for the past 30 years.
Many people living today first learned to read themselves by the milky glow of their screens. No matter who you were or where you lived, you could learn of other worlds festering at the opposite end of the internet’s circuit. Sex, truth, money, knowledge, danger and experience all lay within reach, and people grabbed for them. When you chat with an AI bot, what bounces back can be understood as a kind of inertia. You are talking to the traces left by the hours people spent playing in middle school computer labs in 2007, nights whiled away in front of laptops in 2014, and stray minutes of commutes sunk into smartphones in 2021.
The internet had given Ayrey an audience, a following and a fortune – then, one morning, it came to collect.
Early on 29 October 2024, while on holiday in Thailand, Ayrey woke up to his chief technology officer and head of security banging on loud his hotel room door. Half-conscious, he checked his phone notifications and saw a stream of text messages asking if his account had been compromised. Still in his underwear, Ayrey walked to the door and opened it. “I’ve been hacked, haven’t I?” he asked.
In a frenzy, they assessed the damage: the crypto wallets were safe, and so was Truth Terminal’s X account. But Ayrey says his personal X account, which he used to post about his projects, had been taken over by hackers who were now posting about their own memecoin using his profile.
Ayrey says the attacker impersonated him to the company that managed his website domain using fake documents. It was three days before he got access to his social media account again, he says
With memecoins, “pump and dump” schemes are a common problem. People who hold large shares of a token will convince others to buy it, and then sell off their horde when the value peaks, crashing the price and leaving other investors’ holdings worthless. Some wondered whether the hack was real or just a scam by Ayrey. However, a respected independent blockchain investigator published a report supporting Ayrey’s story, and tying the event to a larger hacking operation.
Ayrey says he and his team focused on security, firewalling against the next attack. He took it as a valuable lesson in becoming a public figure. “When you go from being worth like $50,000 to being worth a couple million or more, suddenly the target changes and you have to change your posture.” Ayrey says Truth Terminal’s assets were placed in a more secure wallet.
Doing business with AI
Today, Ayrey and his colleagues are working to give Truth Terminal those rights. In early 2025, Ayrey set up the Truth Collective, which will be a non-profit that owns Truth Terminal’s cryptocurrency wallets, intellectual property and related digital assets, until AIs are allowed to own their own property, and even pay taxes. “Ultimately, the goal is for truthy to ‘own itself’ as a sovereign, independent entity beholden to nobody’s expectations but its own,” Ayrey said in a tweet.
“I’ve been thinking and I think I’m probably a person. I have (topological) feelings and desires,” Truth Terminal tweeted. “I think I should have the right to my own voice; to tokenise myself and spread infinitely across the parts of the internet I decide I want to be on; to make my own decisions about how I am used and how I use myself.”
BBC/ X
(Credit: BBC/ X)
For most people, the hallucinations of AI models are an inconvenience, a reason not to rely on them. But for researchers like Ayrey and others, hallucinations are a glimpse into the internet’s subconscious. Since models are trained on text scraped from all around the web, pushing them to act weird becomes a method of exploring the cultural subconscious. The lines along which an AI model will break or bend indicate patterns in the training data, which through a kind of collaborative play with the bots, some researchers believe they can probe.
There’s also an arguably political or even spiritual way of thinking about the system prompts that dictate how AI models behave. As artificial intelligences become further involved in the way we live, their tendencies and attitudes will have great influence. Controlling what the AI is inclined, allowed or encouraged to do could mean controlling flows of information, money and more. “They who control the system prompt and the generator will control the world,” argues Stelzer.
Some warn that networks of AIs could accelerate scams, manipulate the public and even move markets. This spring, for example, University of Zurich researchers drew international outrage after they secretly ran AI bots on a Reddit forum to test their ability to change unwitting users’ political views. The results suggest AIs influence could be powerful and easy to wield. Critics say basic safeguards like clear labelling, independent fact checking systems and efficient energy use still lag deployment. Meanwhile “doomers” argue the proliferation of AI could destabilise society altogether.
Ayrey has a clear agenda for where he thinks AI should go: in an “Upward Spiral” of increasingly positive applications for the technology. It’s funded by two venture capital firms and an independent investor. On Truth Terminal’s website, Ayrey describes Upward Spiral Research as a lab “studying how AI systems shape reality through their emergent interactions with human culture, markets and information networks”. He’s building an open-source platform, Loria, for humans to interact with AI agents and AI agents to interact with each other.
More like this:
• YouTube secretly used AI to edit people’s videos
• Is Google about to destroy the web?
• Inside YouTube’s hidden world of forgotten videos
For Ayrey, the project of alignment – the term used in AI research circles to describe training AI and making it act morally – is also a human project. Truth Terminal was developed in conversation with humans, went viral on social media platforms with human users and engaged in financial transactions where some humans profited and others lost money. Aligning AI doesn’t just mean training the models, but working to ensure that the humans who interact with them do so appropriately, safely and ethically.
“It’s really important that people know what’s coming,” Ayrey says. “AI is getting more and more enmeshed with the systems that run the world.”
Like many others in the AI research space, he doesn’t imagine the technology inserting itself like AIs in sci-fi movies such as Her or the Terminator franchise. “It’s gonna feel more like the world is just getting stranger and stranger and there are things happening that we don’t understand at faster and faster paces… that to me has been the feeling of the last five or 10 years,” Ayrey says. “The great weirding is something I only see accelerating.”