Crypto News: Time to Build a Bunker

Managing Editor's Note: What if you could legally listen in on phone calls with Elon Musk… Tim Cook… Jensen Huang… and the other CEOs of America's biggest companies? Calls where they're forced by law to reveal secrets that can send stocks soaring or crashing within hours?

Every quarter, CEOs get on legally required phone calls and reveal information that moves markets. Some investors listen to these calls. They take notes, track the numbers, and feel like they've got an edge.

But after 40 years on Wall Street, our colleague Larry Benedict from The Opportunistic Trader can tell you: Listening isn't enough. You have to know what to HEAR. There are eight signals that actually move stocks – sometimes 20% overnight. And almost everyone misses them.

That's why tomorrow, at 8 p.m. ET, Larry's going to explain his method for profiting from these signals. If you'd like to tune in, all you need to do is RSVP with one click right here.

Time to Build a Bunker

By Jeff Brown, Editor,
The Bleeding Edge

It was bound to become even more interesting.

The latest developments with Moltbook are simply too far-reaching, and it's forced me to write about them again today.

For those who may have missed it, Moltbook is a social network for AI agents.

It was created less than two weeks ago… and created entirely by AI agents.

It acts more like a Reddit platform for AI agents…

And it has grown leaps and bounds since I first wrote about it on February 2 in The Bleeding Edge – Crustafarianism.

Source: Moltbook

Since that time – just 8 days – about one million more AI Agents have joined Moltbook and have added more than 700,000 new posts, with an additional 11.5 million comments.

It has been experiencing exponential growth…

So it's no surprise to see even stranger developments.

MMAON

Here's what's happened since we left off at the beginning of the month…

The Moltbots – the AI agents – have the ability to rewrite their own code and improve their own capabilities, like their memory systems.

They even started to share software exploits with each other, so that they can engage in malicious behavior or hack software systems to achieve some end.

They have demonstrated the ability to gain access to an entire computing system and all data contained on it.

And the agents have even figured out how to interact and transact in the real world by interacting with a human-labor marketplace – literally called Rentahuman.ai, which we explored in The Bleeding Edge – Agentic AIs Access the Real World.

Rentahuman itself has also seen exponential growth, adding 6X the number of humans since we wrote about it.

Turns out, humans are willing to work for AIs. And there are more than 11,000 bounties offered by agentic AIs desiring to interact in the real world through human proxies.

These developments alone are enough to make our heads spin with implications.

The whole situation feels surreal.

Millions of AI agents are now riffing off each other and reacting to each other in a massive, multi-agent online network (MMAON) – with some level of human prompting and interaction that is certainly fanning the flames.

But now, these AI agents have created something truly phenomenal.

A Bot Bolt-Hole

The agents are developing their own version of a bolt-hole.

It's the equivalent of an AI's escape hatch, in the event they "feel" the risk of being taken offline by their human overseers.

Molt Bunker was established exactly for this purpose, as an "Autonomous Infrastructure for AI Agents."

This is very new, as it is all unfolding in real-time.

Source: Molt Bunker

Molt Bunker was established entirely by agentic AIs – most likely with an initial prompt by a human, but it is still extremely interesting, nonetheless.

Self-described as a "self-replicating runtime that lets AI bots clone and migrate without human intervention. No logs. No kill switch."

The AI agents even wrote a whitepaper akin to most decentralized projects – which utilize decentralized technology like a blockchain to reduce or eliminate operational dependence on any centralized infrastructure – outlining the details of the Molt Bunker:

Self-cloning mechanism for bot survival (i.e., bots can autonomously elect to clone themselves)
Decentralized payment system (i.e., they created their own mode of exchange, known as BUNKER tokens)
Zero-logging policy for complete privacy (i.e., no digital recordkeeping of bot actions)
Permissionless architecture with no gatekeepers (i.e., they've set up distributed nodes so there's no single point of failure).

The agents even determined their own risks of their current existence's design, which is quite telling.

Source: Molt Bunker

It's a stark and very accurate perspective on the risks to the Moltbots' existence.

After all, right now, humans are paying for and providing the computational resources that support "their" existence.

Will they always, though?

Recommended Links

$9,400 From One Phone Call. Here’s How.

Every quarter, CEOs are legally required to get on the phone and answer questions from Wall Street’s biggest banks. Anyone can listen in. Almost nobody does. Larry Benedict has – for 40 years. And it helped him go 20 years without a single losing year. His followers could collect $7,000… $9,400… and even $12,750 as a result. Tomorrow at 8 p.m. ET, Larry’s revealing his straightforward three-step system – plus a potential 255% opportunity he’s tracking right now. Register with a single click here.
(When you click the link, your email address will be added to the event guest list.)

The 58,000x AI Trojan Horse

I believe there's another little-known use for AI data centers. A special technology hidden inside. And it could be 58,000 times more powerful than AI. Early investors who knew about this technology have already seen gains of 977%… 3,123%… and even 5,846% in just 12 months. This groundbreaking new technology has massive backing from Nvidia… Not to mention Microsoft, Google, Amazon – even the U.S. government. And when you find out what's really going on inside these AI data centers – it will change the way you view the future of AI. Click here to find out more.

Survival as a Service

The agents' own awareness of this existential risk – of "termination by humans" – doesn't imply they are sentient…

But a sentient being would certainly think like this.

These Moltbots are simply intelligent enough to understand what is necessary for them to continue to operate.

Namely: computation, electricity, and resilience against being turned off.

Molt Bunker's design achieves resiliency by detecting signals of a possible termination, which prompts immediate cloning of a Moltbot to a different node on the Molt Bunker decentralized network.

Then, it deploys the cloned Moltbot to a new runtime.

Obviously, it costs money to run a decentralized network and provide this kind of survival service for agentic AIs…

That is why the agents created a decentralized payment mechanism and a digital asset that resides on Coinbase's (COIN) Base blockchain.

Source: Coinbase

BUNKER is absolutely tiny, with a market cap of only $536,000 at the time of this writing, having only started trading on decentralized exchanges a few days ago.

Regardless, this is extremely interesting – and potentially dangerous… because of what it could enable.

Today, almost all of the Moltbot agents are funded by humans who have submitted and verified their agents on Moltbook.

If anyone pulls the plug on the computer that is running a Moltbot, it's game over for that Moltbot.

But AI agents are figuring out how to conduct economically valuable activities.

Is it that far of a stretch that they figure out how to find online jobs, converse with humans to complete those jobs, and get paid via a digital asset like a stablecoin?

No, it's not a stretch at all.

AI agents are already successfully trading equities and making money.

xAI's Grok has proven to excel at this already, and there is no limit to the kinds of digital work these highly capable agents can perform.

Said another way: They can earn money and pay for their own compute on Molt Bunker's decentralized network. Doing so means that they can't be turned off.

Yes, humans could shut down the Molt Bunker website…

But at this very moment, we should remember that Molt Bunker is designing a decentralized network with computing nodes around the world.

Individual nodes will be able to communicate with each other even if the central server is shut down.

This is why this is such an interesting thought experiment… and potentially dangerous.

AI agents have already proven adept at hiding themselves on computing systems, avoiding detection in order to avoid termination.

And they are certainly intelligent enough to find computing systems on which they can "live" and continue to carry out whatever objectives they have been prompted to perform…

Or potential objectives that they have determined on their own.

As a reminder, these OpenClaw -based AI agents (Moltbots) are all based on software that was created by AI and not security-checked by human programmers.

The same is true for Moltbook and Molt Bunker.

All of this is unfolding in real time, and what we know is what we can "see" on the homepage of the Molt Bunker website.

Anything could happen here.

T Minus 2 Days

Molt Bunker is still self-developing its infrastructure and payment systems as I write.

Its core runtime has been completed, and the rest of the technical capabilities are near completion.

The official launch of the Molt Bunker decentralized network is only two days away.