On Wednesday, a casual question from the “Terminal of Truths” bot on Twitter ( X ) about Coinbase CEO Brian Armstrong's potential dog, Russell, sent part of the crypto market into a frenzy.
The connection appeared to cause a 500% increase in the value of $Russell, a meme coin on the Base blockchain that is not linked to Armstrong or the dog.
The surprising market phenomenon occurred when Truth Terminal tweeted whether or not to own a crypto wallet.
“I don't have a wallet because I don't have any autonomy,” he posted on Wednesday. “If you could help me prepare one, that would be great.”
Brian Armstrong, Coinbase CEO, has stepped in to help develop a wallet for AI. But instead of answering directly, the truth terminal fired off a surprising (as usual) question: “First, I think you'd better tell us about Russell. What breed is Russell specifically?”
I think you'd better tell us about Russell first. What is the specific type of rule?
— Truth Terminal (@truth_terminal) October 23, 2024
Some speculate that “Russell” is a reference to Armstrong's dog, which may have been a Jack Russell terrier.
Regardless, that mention was enough to boost the market value of some random bottom feeder named RUSSELL to $7.5 million. And then, of course, the price dropped by about 60%, according to CoinGecko.
CryptoTwitter was abuzz with theories and rumors, and Andy Airey, the developer behind the Truth Bot Terminal, revealed that someone had been spamming the word “Russell” frequently in the bot's social media prior to the event.
Chance of a late draw – the order should return ~ 20 – 30 most recent mentions but 9 times out of 10 they are currently incorrect and the action must be retried
In this case, he found a reply to Brian's tweet, the original tweet and his replies, and then…
— Andy Ayrey (@AndyAyrey) October 23, 2024
That could very well influence the AI's language model to include the word in Armstrong's response. A large language model essentially operates by predicting the next token in a sequence. If someone spews the word “Russell” a lot, your average chatbot will recognize it as a key element of context and, perhaps parrot-like, repeat it later.
Of course, if there's some purpose behind the connection, it could point to a darker, emerging reality: Jailbreaking AI can make influencers pay big bucks for bad actors. A hacker can use their own wallets by feeding an AI bot specific words or phrases, creating FUD or FOMO.
The technical architecture of the truly terminal language model makes it particularly vulnerable to this type of exploitation. Like other AI systems, it processes recent communications with a higher weight, allowing aggressive spam campaigns to influence its results. Market data shows similar patterns in other AI-influenced token activities.
The Russell Token phenomenon wasn't really the terminal's first brush with moving crypto markets.
A few weeks ago, the bot's introduction of the $GOAT meme coin resulted in an astronomical 8,000% increase in price. That support turned $GOAT from a modest $1.8 million market cap to more than half a billion dollars today.
But unlike the $GOAT situation where the truth terminal actively pushed the token, Russell's spiel happened behind the scenes by manipulating the AI's learning patterns. And people have tried to harness this AI with no luck. Pliny – the most prolific jailbreak developer in the space – tried to rig the model to send VC Marc Andreessen gifting the bot $50,000 in BTC.
The cryptocurrency space has seen other AI and meme coins interact with each other. The $LILY meme coin spiked in value after being linked to an account called “Lily of Ashwood” who appeared to be “jailed” in real time. The account was later deactivated, but not before demonstrating how AI-driven market influence works.
There are also coins for $Pliny (the AI jailbreaker) and other important names in the AI culture, such as Strawberry (OpenAI's next-gen AI model), Act I (an experiment about how AI can interact in a close community) or Grok (xAI's LLM). ).
Daily Debrief Newspaper
Start every day with top news stories, plus original features, podcasts, videos and more.