“It’s going to get ugly before it gets sexy.”
Mark Rydon, the co-founder of Aethir, a behind-the-scenes powerhouse fueling the current AI agent craze, is calling it like he sees it: things are about to get really messy.
Singapore-based Aethir, as Rydon tells MONIIFY, is providing “compute” to most agents at the moment and is talking to many more to come on board.
Compute enables AI to process data, make decisions, and, in the case of crypto, instantly create agents to trade, promote, or launch coins.
And he acknowledges that the “majority of agents in crypto at the moment are actually chat bots.”
Messier times
Rydon says the barrier to launching AI agents or their tokens is about to hit rock bottom. Platforms like Pump.fun already make it laughably easy to spin up a token-toting AI.
Soon, anyone will be able to connect an agent to social media, launch a token, and cash in, he says.
But this ease of entry will flood the market with low-value projects, creating what he calls “on-chain pollution” — worthless coins and useless agents cluttering blockchains.
That is the ugly. The “sexy” bit?
That’s still a few steps away, Rydon predicts. When the market matures enough to sift the signal from the noise, we’ll see AI agents with real utility usher in what he calls the “golden age of crypto and AI.”
With Trump 2.0 set to start, and his decision to appoint David Sacks as the White House’s first czar for both AI and crypto, Rydon is very bullish about AI agents in the future.
Bullish enough that Aethir’s also got a $100 million fund to support builders who will eventually need access to its compute offerings.
Enter Vitalik
Earlier this week, Ethereum co-founder Vitalik Buterin weighed in, suggesting a global “soft pause” button for industrial-scale AI development.
Reducing worldwide available compute by 90% to 99% for a couple of years at this critical period, he says, will buy more time for humanity to prepare for the risks that AI moving at warp speed will bring.
He doubled down on this on Friday: “AI done wrong is making new forms of independent self-replicating intelligent lift” which could “risk permanent human disempowerment.”
Rydon, while agreeable to taking the risks of super intelligent AI ending humanity seriously, is skeptical.
He puts forth a conundrum: who’s actually going to stop? “If Microsoft pauses, will Open AI? If the US pauses, will China?”
Everyone will still be working on AI in secret, he says. “In the basement of the Pentagon and underneath the Forbidden City,” things will carry on.
The real deals?
But is there gold at the end of this chaotic rainbow? Amid the chaos, how do you pick winners?
“That’s the trillion-dollar question,” says Rydon. But he has a few favorites for MONIIFY’s readers to consider.
Rydon points to the likes of Virtuals and Eliza (backed by ai16z, but not the a16z), as projects that combine technical depth with credibility. Creator.bid, with its ability to launch agents boasting $15 million+ market caps, also gets a nod.
Then there is the Roastmaster9000, a Solana-launched agent with a $50 million market cap, and MoMate, which Rydon says is “underpriced” with its 500K+ user generated agents and “cool” focus on livestreams and games.
Full disclosure: Aethir is already providing compute for these projects, according to Rydon.
Conflict of interest? Yes, sure. But Rydon’s bets are worth watching as the AI agent frenzy builds towards something more substantial.
The AI agent space might be a meme-fueled-free-for-all right now, but if Rydon is right, the best of the space is yet to reveal itself.
Edited by Ankush Chibber. If you have any tips, ideas or feedback, please get in touch: talk-to-us@moniify.com
