[00:02] I'm just going to hit claim. Yep, we got it. 37 added to our account. Perfect. Okay, so today I have a pretty cool video because I have hooked up my kind of my AI agent or I didn't hook it up. I just learned
[00:14] it kind of to navigate Polymarket. So I set up the skill MD under polymarket/skillmd. And now we can actually do trading on Polymarket using our browser-based automation system with Claude code Opus
[00:27] in the background. So I had a lot of fun with this. So I thought I just want to share kind of how I set this up and how it works. So basically now I just been testing out the 5-minute Bitcoin trades. So we're going to kind of look at a bit
[00:40] in Altmine what kind of strategy I picked so far. We can track the trades we have done. Also set up this Binance web socket so we can kind of always see the pricing and the liquidation. And yeah, like I said, this is going to be
[00:54] like browser-based automation like I done before in my previous videos. And what I thought we could do today is just do like 1 hour of trading. Of course I'm not going to record a full hour, but we're going to do highlights from the
[01:06] hour and we're going to see how much can we make in like 1 hour with this very simple Claude code strategy here on Polymarket. So yeah, let's just head over to the Mac Mini and I'm going to show you how this is set up. Let's just
[01:22] start and see what happens. Okay, so first I'm just going to talk a bit about going to spend too much time on it because yeah, this is not a proven strategy or anything like that. This is something I set up just to try this out.
[01:38] So the core strategy is to kind of do like this front-loaded bets. We're not going to bet in the 5-minute window. We're going to bet in the next window. So basically we can see we have some signals.
[01:50] So we have like seven different signals. Price versus target. We have the Binance web socket versus target, sidebar consensus, short trend, and we have crowd
[02:03] positioning. This is pretty good. So, this is kind of the psychological part. So, if you have like seven downs in a row, people think I must go up now, right? You get the point. So, and we have the sidebar shift direction, too.
[02:18] So, basically, that is the strategy. So, we're going to front load into the next window. If you haven't seen Polymarket for, you will see in the in the when we do the trading now how this works and how the agent can operate
[02:32] So, I'm going to be running this on Cloud Code, right? And the setup I have. But, I think it could be also very very cool to try this on like open-source models, maybe running this locally.
[02:46] >> So, if you want to try that, I think today's sponsor is very interesting for you. So, I do travel a lot, so this could be flights where I have no Wi-Fi. I go to go to mountain cabins in Norway. There's no connection, of course, there.
[02:59] So, I kind of always run into the same problem. Let's say I need AI on my I'm offline or I'm dealing with maybe some sensitive stuff that I want don't want to send via the API. So, lately, I've been testing out this laptop that
[03:12] has the AMD Ryzen AI Pro chip. This kind of combines NPU, GPU, and CPU all in one chip. So, this means that you can run your AI workflows locally, right? And as you can see here, I kind of run my local LLMs, let's say GPT OSS 20B, and I get
[03:28] around 50 tokens per second. That's kind of fast enough for my use case. I even tried like Coder Qwen Coder 30B. That's a really good coding model, and that works great, too. You can see also here I'm using kind of
[03:41] the Qwen 3 Vision model. I can analyze documents, screenshots. I even run some agentic workflows here on Open Code. That was super useful, too. So, if you care about security, you want to do things locally, I think this laptops
[03:56] with the AMD Ryzen AI Pro is perfect for you. So, AMD also has a free loaner program that you can check out if you just want to to put a link in the description if you want to check that out. So, a big thanks
[04:10] to AMD for sponsoring this part of the video. Now, let's go back to the project. Okay, so let's do it now. So, I said we were going to do 1 hour of trading. I think the time now is like Yeah, 13:32 or something. So, let's try
[04:24] to do 1 hour of trading. So, I'm just going to do a very simple command here in my Claude Code browser. So, I'm just going to say, let's do 1 hour of 5-minute Bitcoin bets with the Polymarket skill. So, when we launch
[04:37] trigger our skill that we have kind of honed into do this. Use the skill honed into do this. Use the skill Polymarket. Yes, I think I forgot to do Let me just check this because I think I want to I need dangerously skip here.
[04:53] want to I need dangerously skip here. So, let me run in this mode let's just do the same 1 hour. Okay. So, now we should be able to do this and this already should Yeah, we loaded the skill and it should know how to
[05:06] navigate now Polymarket and everything. So, it's probably just going to open up Polymarket. Yeah, you can see Bitcoin 5-minute up or down here and it's just trading. So, for the first trades now, I think
[05:20] I'm just going to stay here with you and we're going to try to go together to see what happens here. And you can see here is where we bet, right? So, you can see we have up or down here. So, basically we're just going to guess the market is
[05:32] it going to go up or down in the next 5 minutes and after 5 minutes we know if the price is up or down. So, what we are trying to do is actually front load this. We're going to try to skip to the next 5-minute window to get
[05:46] a better price or not to be too uh implemented by the price that is kind of late in the window. Okay, so we are coming up on 1 minute now and I think the sleep should be finished soon. We're going to check it. Hopefully this time
[06:01] we move on to the next window slot. We place a bet and just wait for the next place a bet and just wait for the next one again. Okay, that was an insane bet. I don't know how it did that, but uh you can see we made $9 on that bet.
[06:15] That was just crazy. It put $1 on up at the last second and the market flipped the last second and the market flipped and we made $9 on a $1 bet. 960% gains. So that was pretty crazy. Uh I haven't seen that before, so that
[06:31] was insane. So uh let's see price crashed uh below target. Up is going to lose. Uh it thought it was going to lose, but it flipped around at the end and yeah, that was an insane
[06:45] return here. 900% Uh I don't know how we did that. So now slot and let's see if we're going to make a bet here now. Again, we are a bit behind. I think I need to update the workflow here a bit. Okay, so now you
[07:00] can see we are kind of early. Now we're going to go into the 8:05 window and I think we're going to place a $3 bet here on up. Uh I don't know how that's going to go, but let's see how this uh is turning out
[07:13] now. One. it three times? Yeah, we're going to do one three times. I don't know why, but
[07:25] it's fine. So our cost uh will add up to three here I guess anyway. Yeah, we made three $1 bets. That's fine. Uh I guess it could change the price a bit, but uh I'm not going to worry too much about that.
[07:39] So So so far we have been doing very well just because um uh of that 900% bet. And I'm just going to turn off the camera a bit and just uh come back if we have some interesting
[07:55] trading here. >> [music] hasn't gone very well. I think we missed uh like three in a row or something. Uh
[08:12] just want to see how this goes. So, we are going to go to the next window and make a new bet. And yeah, let's just wait and see how it's going to do that. Uh yeah, we're going to go to the new
[08:26] window. Perfect. And how much are we going to bet? We're going to do one one on down this time. Uh I think let's see. Yeah, only one down. So, we place that one. Uh now we are almost into the new window. So, the
[08:42] system has been working pretty good. That's what I was uh yeah, that's what I wanted to show to off today. Uh the strategy is not good, right? There's no strategy here. It's just uh mostly about fun and kind of trying to hone in the
[08:57] system to follow what you want to do. So again, we are starting a new sleep. So now we are going to aiming in the sleep to come back at 1:30 before close to get ready for the next front loading in the new window. So again, I'm just
[09:12] going to wait for this and let's see how this bet goes. So again, this is looking like a loss uh like yeah, it's probably going to be a loss again. So, this is not looking too good. Uh when I tested this earlier today, uh we had much more
[09:27] this earlier today, uh we had much more of success. But uh this just shows that this strategy I'm using here now has no edge, right? It's just basically gambling. I would say that. Uh but the fun It's really fun though. I
[09:41] really like to see and watch the agent makes the bets and it kind of makes the reason why it's doing this. So, maybe over time we can find something that is a bit more profitable, but I don't think so. I don't think this way of doing this
[09:57] could be profitable, not at least in this 5-minute up and down poly market But, uh let's make a few more bets before we end the video and see if we can kind of have some win at wins at the end here now. So, now we're going to
[10:12] end here now. So, now we're going to place $1 on up or we're going to do $2 on up. So, let's do one more and that should be it. And again, winning bet here now. Okay, so again, this is looking pretty
[10:27] bad. So, we can't just seem to catch a break here now. So, what I'm going to do out, I thought we can try something different. So, we have been trading for an hour now. Please take the learnings and create a new trading strategy with
[10:40] more risk for the next 30 minutes as a test. Think hard and creative, use the data you have gathered. So, I just want to see uh if we can just come up with something different now for the last uh
[10:53] 30 minutes just to test it out. And maybe we can do something a bit more different. So, let's see what kind of strategy we come up with here. Okay, so here is the new strategy. Fade the swing. The core thesis in a high
[11:06] volatility whip saw markets, bet against the current windows trend. The crowd piles into the obvious direction, odds gets expensive, and the reversal uh is free money. Okay, that seems risky. So, uh we're going to do the opposite. We're
[11:20] going to do sidebar inversion, aggressive sizing. Start at three, then double up or not double, but do five, eight, and then and see what happens. Speed or analysis, Binance confirmation, and the dead cat exception.
[11:37] Ooh, okay. So, let's just go for it and see what happen. So, I'm just going to yes, let's do it, okay? do it, okay? Uh let's just see what happens.
[11:57] Uh and if that doesn't work, I guess we're going to do five on down. This just seems like straight-up gambling. Uh but let's see how it turns out. I have no idea.
[12:11] So, again, uh let's just let this run out and I'm going to take you back. Okay, so it looks like this trade at least is going to go pretty good because least is going to go pretty good because we are 250 up, so that's good. So, that
[12:23] was a winning bet, so we made uh how much did we make? Yeah, $2 or think we're just going to move straight into the next one. Uh what is that going to be? Let me see. That's going to be $3 on
[12:38] So, this just seems random. I guess we didn't adjust the price setting, but let's just see what happens. So, now we're going to do three on down. So, the we're going to do three on down. So, the strategy is basically just alternate.
[12:53] Uh did we make that bet? Let's see. Yeah, we did three on down. and yeah, let's see what happens. So, now we basically this time we lost, so now we lost everything we won in the previous bet. So, now we are going to
[13:08] break even, and I think this means that now we're going to do like a a $5 bet on um a window. I think that was the strategy, let's see. So, that means that we should probably
[13:23] do up on $5, I guess. Yeah, Now we're going to do up five in the next window. 3
[13:37] 4 and 5. So, yeah. Probably going to lose this one, too. Uh but yeah, let's just wait for it. And uh we lose, then we're going to part watch the final $8 bet
[13:52] was another loss. Just a stupid strategy, to be honest. Uh we have moved on. Did we move on to the next window? I don't know what we did here, but yeah, I think we are moving into the $8 bet. This is just going to be the last one,
[14:06] this strategy is just um because this strategy doesn't have anything to do with It's just gambling, right? But basically, let's just run
[14:19] this out and then we're going to do a simple conclusion. Okay, so it looks like we actually end on a high here. We have $4.3 in uh yeah, return. So, that That was at
[14:31] least good. Uh but the price has also dropped, so our portfolio hasn't dropped, too. But uh yeah, we can just wait this out. But basically, I think that's going to just sum it up. So, what is pretty cool about this is that the
[14:46] the agent or Claude Cod with Opus here, I think we're running Opus, yeah, basically has no issues doing this. So, I wouldn't say this is probably the last time I'm going to explore this uh since I just thought of this today. I
[15:00] explore this. But I think it's going to be interesting to see what we can do with this agent here on this predictions market. It could be something we can find here that is a bit more interesting in that what we have been doing so far.
[15:15] Uh yeah, this landed. So, we kind of we got some money back here. And yeah, like but this was just some money I had on Polymarket just for testing. But overall, I've been very impressed how Claude code navigates this and kind
[15:29] of how it performed on actually executing the trades. There was some small issues, but maybe, like I said, with a better strategy, you could find with a better strategy, you could find something here. So, if you want me to
[15:43] kind of share this skill, just give this video a like and I might try to set that up. I also want to to mention if we go to skillsstore.md now, you can see I made an update here.
[15:56] So, a lot of you asked about the browser agent I'm using to control this. So, this you can find more information here on skills.md.store. here on skills.md.store. Also, don't forget to check out AMD. You
[16:08] for that if you want to run some open-source models on very good hardware. So, yeah. Thank you for tuning in for this experiment and I'll see you in for this experiment and I'll see you again probably very soon.