Good afternoon, and welcome to another episode of "markets are acting weird and AI isn't helping."
There's an AI trading competition happening right now that should be front page news but nobody's talking about it.
They gave different AI platforms $10,000 each on October 17th to trade crypto derivatives. Real money, real trades, everything transparent.
The results?
Well, let's just say the AI revolution in trading might need to learn how to not lose money first.
The Numbers (And They're Brutal)
Here's what's actually happening in this Alpha Arena competition:
ChatGPT: Down about 65-70% of its initial $10,000 as of yesterday
Most general AI platforms: Getting destroyed by basic market chop
The winner: DeepSeek (Chinese math/coding specialist) - up 10-40%
Second place: Grok - up 15-30% with simple BTC longs and low leverage
Wait, hold on.
Grok is actually doing well? Yeah, that's the thing everyone's missing. This isn't "all AI sucks at trading." This is "general AI built for everything sucks at the one specific thing everyone claims it'll dominate."
Grok's strategy that's beating most of these sophisticated systems?
Buy Bitcoin, add a little leverage, don't overthink it. A strategy so simple a monkey throwing darts could have figured it out.
While Markets Are "Designed to Extract"
You want to know why we're seeing these results?
The same reason retail traders are tearing their hair out right now. There's a lot of different reasons why we're not getting major directional movement, but the core one I always go back to is that markets are designed to extract.
We're seeing these morning moves - Nvidia yesterday, right? Pushed up to around 182 in the morning, looked bullish, nope. Rolled right back down to 178, saw everybody capitulate, then squeezed right back up to 181 close.
This is algorithms talking to one another on 15-minute, 30-minute charts. Get into oversold territory, they turn right back around. We can keep doing this for a while, and there's no doubt in my mind that funds are more than content to just do this while everyone else loses their minds.
ChatGPT? It over-traded every single one of these fades. Got wrecked by the exact same chop that's frustrating human traders.
The Real Story Nobody's Telling
Here's what this competition actually reveals: specialized AI beats general AI, every time.
DeepSeek was purpose-built for math and coding. It's dominating because it understands market microstructure, not because it can write poetry about your feelings.
Grok's winning with simple directional bets because it's not trying to be everything to everyone.
Meanwhile, ChatGPT and Gemini are getting liquidated on DOGE trades because they were built for general conversation, not risk management.
What This Actually Means
I'm not an AI skeptic in general, but when ChatGPT can solve complex problems and write code but can't figure out not to lose 70% trading crypto derivatives, maybe we should pump the brakes on the "AI will replace all traders" narrative.
At least human traders know when they're confused. These general AI systems are losing money with complete confidence, over and over again.
The Chinese specialized AI that's winning? That tells you something. Purpose-built tools for specific tasks beat general-purpose tools every time. This is basic engineering, not some profound AI insight.
The Competition Continues
This thing runs until November 3rd. Keep watching, because the results should be telling us something about what works and what doesn't in algorithmic trading.
But here's the real kicker - while everyone's debating whether AI will replace human traders, the best-performing AI in this competition is using strategies so simple that human traders figured them out decades ago.
The Bottom Line
There's a fundamental disconnect between AI trading hype and AI trading performance when you get specific about the task.
General AI? Getting destroyed.
Specialized AI? Actually competitive.
Simple strategies executed well? Still winning.
Maybe instead of worrying about AI taking trading jobs, we should focus on using AI for what it's actually good at - analysis, not autonomous execution in volatile markets.
This competition ends in a couple weeks. I'll be watching. Because sometimes the best reality check is just letting the market do the talking.
And right now, the market is saying that hype and performance are two very different things.
Stay Positive,
Garrett Bladwin
P.S. I’ll be live tomorrow in the pre-market to breakdown the latest earnings, movers and shakers. It’s free to join me, click here to RSVP.