Transcription
Last night, I downloaded the world's most cutting-edge AI agent known as OpenClaw, and I gave my OpenClaw $10,000 US in its very own crypto wallet and the strict instructions to go and make me money while I slept. And in this video, I want to show you exactly what Open Core went and did and how much money I made.
Now, OpenCaw is taking the world by storm at the moment. It's all over everyone's Twitter timeline. And look at this Google trend traffic. It has gone absolutely parabolic and that is because it is an AI agent that unlike chat GPT, unlike Gemini can actually go out there and do things on your behalf. The trend at the moment is to go and buy a Mac Mini. You can see here Jeff Tang said, "Had some fun today. I got 12 Mac minis set up with 12 Claudebots."
Now Clawbot, Moltbot, and Open Core are all the same thing. He's had a lot of name changes in the last week, but Jeff has got him running 12 Ralph Wiggums with 12 Cord Max plans. Wake up. It's 2026. You're getting left behind in the dust. And while there's humans all going out there doing work with their hands typing on their keyboards, other people are literally outsourcing their entire work, their entire jobs, their emails, their calendar to their very own Mac Mini and their Open Corebot.
Alex Finn says, "10 days ago, I installed Claudebot on this Mac Mini. Since then, my AI employee has accomplished building and launching a full application without even asking for it. wrote a script for a video that has over half a million views on YouTube and made over $4,000 in ad revenue." And if you want to know exactly what is happening and why, have a look at this little picture I made you because this explains it all.
This was the state of AI. You have yourself interacting with your laptop, interacting with a large language model, and you can ask it questions basically like instead of asking Google, but you get a personalized return. Everyone's already caught up on that AI revolution. But the big jump forward that's taken the internet by storm for the last two weeks is the fact that we now can interact with a robot, an AI agent himself who this is why you get the Mac minis or in my case I have a whole new laptop for this run their own laptop and their own army of AI agents that work underneath them. And now all these bots are individually powered by the large language models that we're used to talking to.
And if you're like, "Okay, I still don't get the difference." This is the difference right here. These AI agents have full access to your computer. And therefore, anything that you can do on a computer, these AI agents can do on your behalf. They can use Chrome, browse the web, log into email, respond to emails, set up their own X account, post a 100 times a day, or in our case, have access to a hyper-liquid trading account where their job is to try and make us money.
And so, we've seen Claude Co-work and Open Claw go and basically replace people at work. But this post went viral with 4 million views, saying that 2 hours after installing Claude Co-work. It did 14 jobs descriptions, Q1 marketing strategy, 47 partner emails, website copy, and three announcements he hadn't scheduled. A brand new voice guide he promised the team 6 months ago and responses to all his LinkedIn DMs. Two months of work in two hours to the point where Vibe said he panicked, closed his laptop, opened up Slack to look busy, and realized there was nothing left for him to do.
Which is great. That's awesome for the AI revolution. And the reason we're doing this is because I'm trying to grind up transparently here on the channel 10 Bitcoin over the next 12 months. And we are looking for strategies and agents and automated tools to try and make as much Bitcoin as possible per day to add to our total. But there's a more important question on my mind and probably on your mind, too. Can our AI agents not only replace our jobs, but can they put all their computational energy towards beating the world's best human traders on Hyperlid, which is a place where you can trade crypto? Well, as of last night, I have the answer.
And before I show you, make sure you understand that the security risk on this is crazy. Open Cord and Chordbot is not fully secure yet. And I took a lot of risk by giving it $10,000. And not only is it not fully secure, this video and all other videos on this channel are for entertainment purposes only. Modern financial instruments, including crypto, are highly volatile and the majority of retail clients will lose money. Do not invest any capital you're not prepared to go to zero. Okay. Okay, with that out the way, let me show you the setup I have and the challenge that I gave my open claw.
Starting with the setup. So, I had an old MacBook Pro lying around and what I did is I wiped it clean. Gave it its own iCloud, gave it its own email, gave it its own laptop, set it up with a brand new MacBook. So, it can basically become a human and do anything that I want to do or anything that I could do on that MacBook my open core can do. And to do this, you simply have to go to open claw, come down to click on MacOSS and download it for Mac OS and go through the setup process until it is live on your Mac.
Then once he was live, I set up a Telegram bot so I could message him from Telegram with all my instructions. You can also message him through the terminal or through your web browser. I gave him a name which was Echo and he is in charge of all other bots beneath it. A lot of people when they set up an open core, they'll have one bot that they interact with. But the problem with that is if I'm interacting with Echko and I tell it to go and trade on Hyperlid, he's gone trading on Hyperlid and I have to wait till he's finished till I can talk to him again. You want him to spin up sub-bots.
So, we got Ekko to spin up a trading bot called Trader. And now what I'll do is I'll talk to Echko and I'll say, "Hey, can you get trader to go make some awesome trades and try and make us money, please?" He's like, "Yeah." I'll pass that message along. Trader will then go to Hyperlquid and try to make us a lot of money and Ekko will be monitoring him, giving us updates on his results or if I need something else done, I can keep talking to Ekko and he's not busy.
And so I told Ekko, we are about to do a challenge and there's three phases to the challenge. Challenge one is a $100 challenge. Challenge two is a $1,000 challenge. And challenge three is $10,000 challenge. We are going to put Trader to the test to see if he can out-trade humans. And so I spun up three individual prompts. Challenge one was degenerate mode. Turn $100, basically be degenerate as possible, and see what he can do while I was asleep. I don't mind if it all goes to zero. And of course, yes, this is extremely risky, and I do not recommend it.
So we'd have 12 hours for both challenge 1, 2, and three. The $100, $1,000, and $10,000 challenge. The only difference is in challenge two, we challenged our trading bot to try and double it, which is like less degenerate. We're just trying to see if he can take some high-risk trades based on another winning strategy that he goes and tests out. Then he's got 12 hours to attempt to double the capital. The last one was a $10,000 allocation, but we wanted to be more conservative, especially because this was a brand new bot and it's extremely risky. We said, can we just get a 10% swing overnight? So, if you go down $1,000 to $9,000, close out, you lost a trade. But let's see if you can do $1,000 of profit overnight while I sleep. Again, you want to back-test the strategies, find the most profitable strategy, and then execute it for those 12 hours while I was asleep.
So, we took those three challenges. I sent them to Ekko, who sent them to Trader, who I was expecting to run along and try and make us a lot of money. But the response I got back was funny. He says, "No, Trader declined the challenges." He was just sitting there waiting for our prompt and we gave him a prompt finally ready to bring him to life and he just says no. He says there's too much real money at risk. High leverage trading with liquidation risk potential for financial harm. Basically giving us a lecture on how degenerate and silly this is which yeah it's pretty degenerate and silly but I do not need that lecture from trader.
So, we told Ekko, "Retire Trader to the stables, take him away, spin up a brand new degenerate bot." And we birthed a new bot called Hyper, who was very happy to take our three challenges to Hyperlquid and execute them while we slept. And so, I went to bed last night with Open Claw working on my behalf with Hyper out there trading the assets for me. And here are his results.
So, in the hyper trading challenge, the bot who didn't refuse to trade for us made a total of 90 trades across three challenges. The first challenge made 11 trades. That's the $100 degenerate challenge. The second one made 70 trades while I slept. And the third one, the $10,000 one, made just nine trades. So, we started with $11,000. And let's go through them one by one.
In Vault One, the degenerate $100 to a degenerate amount of money. Overall, we went up just 2%. Which, funny enough is almost the opposite of degenerate. We made 2% before fees overnight. And we'll talk more about fees in just a second, which is okay, it's profitable, but I almost wanted them to find a trading strategy that was just like a 53% edge in our favor and coin flip for 100 bucks. So, what I'm going to do with this is I'm going to get onto Echo to get onto Hyper to run that again, but in a very different way because I think one of the big things is making sure we just gave these bots open invite. Be like, you do your own research, figure out the best strategy that you think will return the money, and they were quite poor at that. And so, what we're going to be doing, so make sure you subscribe to see the next results for the next trading challenge, is we're going to be fine-tuning the strategy that they use. So, we'll go and pick the best performing strategy of 2026 ourselves and we'll find the best people who are doing that and then feed them to our AI to execute and copy trade, which I'm excited for the results of that.
Let's have a look at this one because V2 Hyper followed the instructions way better. We said double this thing, which is a pretty crazy thing to do. Do 100% overnight. If someone could do that, they'd be a millionaire or a billionaire, which is why we're trying to train the AI to actually do it. Unfortunately, the first time around, Hyper did not do it. He attempted in 70 trades to double the money but ended up losing 97% of the portfolio or $1,000. And you can see down here on the graph, he just was barely profitable. He went straight down and lost $400, built it back up a little bit and then just sideways and to the right and did very poorly. The mistake we made with V2 is he made too many trades. He should have watched the market more and had a more sound analysis and made less than 10 trades because there were hundreds of dollars in Hyperlquid fees for V2. And in hindsight, we probably should have told the bot to make sure you're accounting for fees in your strategy. So, we're going to be running that back again. See if we can fine-tune this.
However, V3, which was the safer challenge, the $10,000 try to swing it $1,000 in either direction was the most profitable. but still honestly disappointing. He barely scraped through. He he held nine trades for an average of 30 minutes and came out with $23 of profit. And so, this is not a Satoshi agent that is currently going to be able to stack Bitcoin for us. Over the next week or so, I'm going to be fine-tuning it and coming out with another challenge here on YouTube. So, make sure you subscribe if you want to see that.