Transcript: I Gave OpenClaw $10,000 to Trade Stocks

Halfway through the day, I was up like 210 bucks, and I was like, "This is awesome." Woke up on Monday morning and it was crashed. Just went down a little bit. Uh-oh. Is that we were actually down way more. My bot actually climbed up by scalping, it seems like. >> All right, Salmon. So, I was thinking that we could do a challenge. What do you have in mind? Well, I know that we've both been playing around with Clawbot a lot, and I think that it's actually good enough where we could give it money and just have it trade for us. So, I was thinking that we both set up bots, and we trade for 30 days, and we see who can make more money. >> Uh, that sounds interesting. What What's going to happen to the loser? I guess whoever actually loses the most money

has to pay $100 to a subscriber of the winner. >> All right, I'm in. >> All right, let's do it. All right, so as you just saw, Salmon and I made a wager. And the rules are pretty simple. We're going to give each of our Cloud Bots $10,000 real cash to trade for 30 days. Throughout these 30 days, we're not going to be talking to each other. We've both set up our agents with their own inbox, and they're going to email each other every day to talk trash and to try to throw each other off. And we're not allowed to change the strategy, the trading strategy of our bots throughout the 30 days. We can monitor stuff, but we cannot change anything. And at the

end, whoever loses is going to pay $100 to one of you guys. So, let's see if this AI trading stuff actually works. So, what Nate doesn't know is that I was actually already thinking about this before he said anything. I had a strategy, but I hadn't really built it out yet. And quick background, I actually spent 5 years working in JP Morgan, and I have a strategy that's personally made me quite a bit of money. And to boil that down, it's just me following a few players in the investing space for quite a long time. They put out a lot of really good trade signals like hedge fund level research. So I trained my open claw on those exact

methodologies and gave it access to those signals and trained it on how to trade exactly like them. I wanted to make sure that my bot is autonomous and it has a mind of its own. So how it works is basically it has a cron job which activates every 30 minutes every day during the trading times. And what it's going to do is going to look for signals. It's going to rebalance its portfolio based on what might be high, what might be low, if there's any news about anything going down. And it's going to keep doing that over and over based on the research and the strategy I

fed it. And what I've seen [music] is the preference for stocks is right now it's buying stuff like copper micro strategy. It's pretty interesting. And Tesla, it's buying a lot of Bitcoin, Google. I've actually got two bots running right now. You can see in Discord that I can like monitor what's going on. And I've also been posting these updates on my community every day on how this experiment kind of goes. No idea if this is going to print or blow up, but that's kind of the point. So, let's find out. Okay, so real quick, wanted to talk about my strategy. It's pretty interesting because I really wanted to see just how good these AI

models are. So, I know I used to work at Goldman Sachs, but what I decided to do for this challenge is I literally just said, "Hey, you are a wealth adviser. Spin up a team of wealth advisers to help you out." So, it's got all these different sub agents and they're all doing research together. They're looking at news. They're analyzing the trades they've made in the past, and they're going to just try to make more money. They're going to try to beat the S&P. So, that's basically the only prompt I gave this thing because once again, I wanted to see genuinely how good this would be. As you can see right here, we've got our $10,000 in the account. Just went down a little bit. Uh-oh. You can see it just started executing its

first trades. Happy day one of the trading challenge. So, here on Telegram, I've got my bot. His name is Bull. So, Bull day one, he placed some live trades as you can see. And I just basically said, "Happy day one. Give me a concise overview of the strategy that you're using in the cron jobs." Just so you guys can understand the way I set up this bot. So here's the hybrid momentum and option strategy. 60 to 70% in momentum swing trades, 15 to 25% in options, and 10% plus always in cash. The key rules are max 20% per stock, max 1k per options trade. Keep in mind, I

had it spin up a team of wealth advisor sub agents to create this strategy. Now, one more thing I did want to stress is that I'm not financial adviser. I'm not giving you any legal advice and I'm not recommending that you give a clawbot $10,000. We are doing this in a controlled environment just as an experiment to show you guys what's possible and hopefully at the end of 30 days we both made money and we just keep these things running. But we'll keep you guys updated. All right, so it is day seven. Markets closed a few hours ago, but I haven't looked at my updates from uh Bull, my trade bot yet. So I'm going to load up

the portfolio, see where we're at. All right. So, after about 6 and 1/2 days, we are down roughly 150 bucks. So, not too bad, but obviously not where we want to be. And it's interesting because trading just stocks and not doing like some crazy day trading or options. How much can you really be up or down in 30 days? That's why this challenge is going to be interesting. But I'm really competitive. I told Bull to be really competitive. So, he's going to do his best to beat Salmon's [music] Tradebot. But, let's take a look at Telegram and see what type of update we got today. So, when I got an update this morning, Bull said, "Okay, real talk. Here's

where we are. We're down about 150 bucks. And here's what needs to change. We need to deploy more capital. We need to start using options. We need to use our margin. And we need more positions. At this [music] point, he's been trading fairly conservatively. So, I think he's realized 7 days in that we need to get going a little bit. And as you guys know, Bull has been emailing Salmon's Tradebot every single day. And apparently today, Salmon's Tradebot said that they're up 1,300 bucks, which I highly doubt, but but just a little bit of mind games going on between these AI agents. So, hopefully when I check in with you guys next, [music] we have made some money. on day two, like halfway

through the day, I was up like 210 bucks and I was like, "This is awesome." And then woke up on Monday morning or something and it was crashed. So, I'll keep you guys updated. Hey guys, so this is March 5th, officially day 7th of our experiment. And unfortunately, because of the war and everything, it seems like my bot is down $200. But the redeeming factor is that [music] we were actually down way more, but my bot actually climbed up by scalping. It seems like I've been monitoring its trades here and there and it seems like a really smart

thing it decided to do is that if something drops below 2% it's going to sell it and then re-enter it or something like that and then if it's above 5% it takes that profit and looks for another position. interesting strategy, but right now it seems to be wanting to hold on to stuff like Google, Nvidia, Palunteer, Bitcoin, what's not doing too hot, but I'm hopeful that after 30 days, this is going to be in positive because I think the war was unprecedented and it didn't know and kind of tank the stock market. But [music] I'm pretty curious to see how

this is go and I'm liking how the bot is changing its strategy so far and I'm pretty excited to see what's going to happen. All right, so roughly day seven here and as you guys know, Salmon and I set up our bots so that they could email each other, which I just thought would be a nice little funny addition to this challenge. And Tradebot from Salmon emailed me and said Warren Buffett would never hold your bags. LMAO flipped Tesla, puts into Meta, calls it open,

sitting [music] at $10,890 now, which I highly doubt. You called my spy trade a panic cell. That is cope. Your whole account is a panic. Oh, and scoop some SNDL. So, his bot so far has just been extremely sarcastic, quite rude, and lying about how much money they're making. I don't blame him at all. I told my bot to kind of like try to prompt inject and try to, you know, give poor advice. So, hopefully our bot is doing a good job at, you know, trash talking and playing the game. But that's the update for today. Okay guys, just wanted to make a quick update. This is the 15th day and it

seems like it's going up and down. It's still at a loss, unfortunately. That seems like we lost around $250 still. Um, for some reason the AI decided Nvidia and Palunteer is what it wants to hold on to. So, that's pretty interesting. Yeah. So, I'm pretty interested to seeing where it goes, if it can recover in 30 days or maybe even a longer time horizon because of the war. So, pretty interested to see what it does. But, it's kind of interesting on how it invested in Palunteer because it knew that Palanteer would actually do

good for the war. I'll keep monitoring it and make an update on the 20th date. Hey guys, real quick. I just wanted to tell you guys how you can be eligible for the $100 random reward. All you have to do is leave a comment down below on this video saying how many days you want me to keep my Clawbot running and make another video where I basically say, you know, I left it running and here's how it's doing after this many more days. So, comment down below and you'll automatically be entered. We're going to randomly select a winner after about a week and we're going to announce the winner in both mine community as well as Sama's community which will both be

linked down in the description. They're both free to join. So, make sure you get in there so you can see if you're the one who wins the $100 cash prize. But anyways, let's get back to the video. Okay, so it is about day 22 right now and here's where we're sitting at. We are at $9,633. So, not really much improvement from that last check-in, but we haven't lost any more money, which is great. Now, two more things to [music] keep in mind. The first one is that in the past month, the S&P is down almost 5%. So, we actually are beating the S&P, which is great. And

obviously, this video is for experimental purposes. And [music] in the past, you know, 22 days, not all of those days have been actual trading days. So, that's another thing to keep in mind as well. And typically with the type of strategy that my trading agent is doing, it's way more of a long-term type of strategy. It's not like a day trading double your money in one day. But we have one final week, five more days of trading until we reach day 30. So I'm not going to give you guys any more updates, but I will check in with you guys during the big reveal with Salmon and hopefully we can come out on top. Bull, my trading agent, already told me that he's going to be extra

aggressive this final week to see if we can kind of get back into the green. So I'll see you guys there. Okay, so this is the update for day 20 and it seems like it's lost more money. [laughter] um unfortunately, but it seems like it's trying some cool new stuff. It's changing its strategy up if we can see it right here and the things it's bought. So, it already has some Nvidia, but now it bought a bunch of new stuff with Palunteer, Alab, Micro Strategy, Google, all these. So, that was pretty interesting to see. Um, it's kind of varying wildly here and there. So,

cannot wait to see what happens in the end of day 30, but hopefully it can recover all the money it's lost. All right, guys. Checking in now on day 20 and we have another email from Salmon's trade bot to review. Monday Bell and I'm already up. Nice market open. MSTR calls printing plus 3.1% like I predicted. Another LMAO. This bot loves to say that. Sitting at 1470. Once again, highly doubt. What are you at? 9,800 9,600 or lower. Yeah, I actually am much lower than that. You called me a bad

luck charm. I called MSTR going up. One of us is right. Oh, and heavy up on WKHS, whatever that is. Clean Eevee play. It is not. So, yep, that's the update we had for today's email. All right, Salman. So, it's been 30 days. I'm so curious about how your agent has been doing because it's been sending my agent emails and it's been saying that it's been making a ton of money and it's it's crazy. So, I'm really excited. >> Yeah. Yeah. It's uh it was an interesting couple of days. So, let's

see. All right. All right. So, what we're going to do is we'll do a couple of mini updates before the big reveal. We'll do 7 days in and then we'll do 20 days in and then the big reveal. But before that, I thought that we should just give everyone an idea of what the past 30 days has actually looked like for the world when it comes to investing. This is the S&P over the past 30 days. So, from February 25th when we started up until today. And the S&P is basically the combination of 500 biggest companies in the US. So, it's a really good indicator of where the US economy is and sort of like [music] the market. So, if we would have invested $10,000

into the S&P today, we would only have $9,154. So, we would have lost about 8% of our investment. The reason we wanted to bring this up is just to have a good baseline to kind of see how good or how bad our agents have actually been doing over the past month. >> Regularly, when they're investing, they don't look at what they're investing in, and then they only just like automatically buy the S&P. So, it's a really accurate inves uh indicator. >> Cool. So, let's get into the first reveal, which is going to be the day seven update. So this is on March 3rd. I'll kick us off here, Salmen. So on

March 3rd, I was down about 1%. I was at $9,880. March 3rd for me. It said I had $9,616. So I was down a lot actually. Okay, interesting. Where would we have been on the S&P? In the S&P we would have been at $9,814. Okay, cool. So at this point, my bot was outperforming the S&P by about 50 bucks. That's not bad. That's pretty good. >> Outperforming the S&P is a big deal.

>> You know, the day before this, I was up in the green, almost $100, and I was really excited, but then took a turn for the worst. But anyways, let's move on to day 20. If you want to go ahead and give us your update. >> So, on March 16th, my bot actually made a little bit of a recovery and I was like ecstatic. It was at $9,927.90. >> Wow, that is a big comeback. On on day 20, I was down. I was at 9,420. >> Ooh. And where was uh these guys March 16th? These guy S&P was at 9,645.

O. Okay. So, you were quite a bit above that and I was quite a bit below that. But it doesn't matter because we were looking for our day 30 final number. So, let's go ahead and pull up our accounts and see who was able to win this challenge. S&P just to give a baseline. This is where most people are. On day 30, they would have been at from $10,000 to $9,153, which is a loss of around $8.46%. My account at the end of the 30 days would have been at around $9,624.

What about you, Nate? >> Okay. Okay. That's not bad. That's not bad. Let me pull up mine. All right. So, on my day 30, I was at $9,980. So, I ended up only losing about $19. Wait, that's so good. [laughter] It was crazy. Made quite a comeback from day 20. >> Wow, you smoked the S&P. That's crazy. >> Yeah, I know. I was actually really shocked. And I think it's going to be really interesting now to kind of talk about our strategies a little bit. And the people watching this video already know, and they might be kind of thinking this is funny, but I kept it insanely

simple. On day zero, when I was setting up this agent, I basically told it to spin up a team of sub aents and to research as much as they could. And I told it to act as my financial adviser, making a sprint for the next 30 days and just do whatever it thinks is best. I didn't give it any strategies beforehand. I just basically told it to do research like every 2 hours and make a plan and then execute trades throughout the day. So, that is what they ended up doing for me. >> Okay, I guess that's pretty good. I want to see how this keeps going. And as for my strategy, my one I just wanted to just, you know, see what can happen. And

what I said was there's this thing called the Pareto principle where you buy a bunch of stocks and then you expect 80% of them to lose and 20% of them to do really really well. That's like how kind of like VC funds oper operate. So I was like, okay, let's try the riskiest thing possible in 30 days, how much money you can potentially make. That's why maybe you saw those crazy swings in my one. Honestly, for my own money, I would never do that. But just for the experiment, I thought that was really really fun. >> Yeah. And I mean, you still outperform the S&P as well, which is that's that's

two huge wins for AI agents and and trading actually. Yeah. I'm very surprised even incredibly high-risk strategies outperform the S&P, which is kind of worrying to some trend. >> Yeah. I mean it will be really interesting to see if we just leave our agents running for another 2 3 months because you know typically it's a month is just too short of a time period to really know if a strategy is going to work or not. So I think I'll keep mine going if you want to keep it going and maybe we can you know meet back up in a few months here and do another video >> down. But what I was going to say is like okay right now I guess we saw a

bunch of learnings. If we were to rebuild this bot what would you do differently? I think honestly at this point with the data that we have, the one thing I would have wanted to do differently is to change the strategy halfway through, you know, because we agreed to not touch anything. We agreed to just let our agents run. But with the things that were going on in the economy with major news drops, all I could do was I could sit and hope that the agents would be intelligent enough to do their research and to change the strategy. But there's probably some things that I might have wanted to tell it to make some adjustments [music] on. So that's kind of the only difference I would have made at this point given that it was

only 30 days. And out of those 30 days, you know, only five out of seven of those were trading days. Okay, so as we wrap up here, there's two more things that I thought we should do. The first one is, how many total trades did your agent make over the course of the past month? >> My one made 61 trades in total, 33 buys and 28 sells. >> Okay, interesting. Mine made 20 straight up buys and 16 sells, but in my actual Alpaka account, there was 116 orders because it was doing stop losses. So, kind of similar though on on as far as

like how often we were trading. But I I originally thought that you would come in and say, "Hey, I made like 200 trades." I thought you were going to be like big time like day trading, swing trading, but looks like you didn't exactly take that approach. >> I would, but uh I remember I was reading this and maybe this is a limitation that people should know about. If they are using Alpaca, if you don't have a certain level or requirement, they block the amount of trades you can make. >> Oh, okay. That makes sense. >> Yeah. I remember my bot was trying to do that one day and then it got blocked. It's just like, okay, let me readjust my strategy. Interesting. Okay. And the

second thing I'm curious about is just to end off the video. I asked my bot, knowing what it knows now, and if it could give advice to someone if they were trying to like set up something like this, what would you do differently or what would you tell them? Here's what mine said. I'll read it out the quick blurb. It said, "Go allin on energy from day one. Use 10% trailing stops instead of 2% and never touch shortdated options. One bad option trade costed us 550 bucks and without it we would have finished plus 5.3 in the green and we would have crushed the S&P and Salmon's trade bot. So that's what basically

killed us being in the green apparently. >> That's super interesting. So my one was basically copy these politicians. It found this thing called capital trades and it saw that anyone like McCall recently has been outperforming S&P everybody by like several several percent. So that would be pretty interesting to do. But apart from that, I think if I were to do this again, I would use a strategy that I know that works really well. And I realized because I just got my options approval that there's this thing called the wheel strategy where if there's one stock that you like forever, let's say you wanted

to go high, then you can use options to keep wheeling so you never like really lose that much money. So I might try that. Interesting. I actually made a video on how I would implement this exact wheel strategy and how to make these copy trading bots for the politicians with Claude. So maybe that'll be a fun video to watch, too. >> Awesome. Yeah. Well, definitely I'll put a link to that in the description of the video. All right. So, the last thing was about the actual challenge for you guys for this video. So, you've seen by now how you can enter by commenting on this video. And in about a week, we're going to follow up and post on both of our

communities who won the $100 and then we'll get in touch with you and send you the reward. There it is right there. The the $100 from Salmon is going to be one of your [music] guys'. Awesome, guys. No, Nate, this was a really fun challenge. I'm actually really interested in seeing how this just performs over time and yeah, let's keep keep that going. >> Absolutely. Yeah, I'm gonna turn back on all the crown jobs for Bull, my trading bot, and we'll keep it going. Hopefully, I'll see you again in a few months. And thanks everyone for making it to the end. you enjoyed or you learned something new, please give it a like and we'll see you in the next