Weāve heard the drumbeat growing louder and louder by the dayā¦
Every boss, every CEO, every talking head on CNBC is shouting the same thing:
āAI is coming for your job!ā
Maybe thatās true for someā¦
But when it comes to trading, the most recent data proves (once again) that AI-based trading algorithms are nothing more than blueprints for a blown account.
Case in point, a recent Bloomberg article posted by Justina Lee titled āAI Bots Auditioning for Wall Street Trading Are Mostly Losingā concluded the followingā¦
AI isnāt ready to replace your fund manager ā and the public experiments testing it are showing why.
Across a series of new trading contests between the worldās leading AI models, the verdict so far is unļ¬attering. Most of the systems lose money.
They trade too much.
They make wildly diļ¬erent decisions when given identical instructions.
And no one yet knows if these shortcomings will fade with more powerful iterations ā or if they reveal something fundamental about the gap between large language models and how markets actually work.
The basis for the most recent findings put eight major frontier AI systems ā including Anthropicās Claude, Googleās Gemini, OpenAIās ChatGPT and Elon Muskās Grok ā up against each other in four separate competitions.
Each was staked with $10,000, and then they were turned loose on US tech stocks for two weeks. The challenge involved trading on a variety of signals, acting defensively, reacting to the competition, and using leverage.
The results were not pretty.
As a whole, the portfolios lost about a third of their capital across all 32 sets. Only six models finished with any sort of profit.
82% of the AI trading models lost money.
To me, this is the most public demonstration of what happens when AI systems try to take on some of the most lucrative and skillful work on Wall Street, and why the motto called āhuman in the loopā remains vital to success when it comes to trading real money.
Now, I admitā¦
AI is great at doing research, but itās still ineffective at identifying key variables that move stocksā¦ analyst ratings, chart patterns, insider transactions, sentiment shifts, etc.
And AI still mistimes trades, incorrectly positions sizes, and trades too frequently.
Sometimes, the various AI systems canāt even agree on a market direction.
In the most recent study, Claude mostly wanted to go long, Gemini had no problem being short, and Qwen was comfortable taking risks with big leverage.
Not surprisingly, they all mostly lost money.
Alexander Izydorczyk, former head of data science at Coatue Management and now at NX1 Capital, recently wrote that no AI trading bot he tracks has yet shown a lasting edge.
YOUR ACTION PLAN
My takeaway is simple. Leave the wonders of AI to research (and other data-mining tasks). But leave trading to those who have been fully immersed in it for their entire livesā¦ because guys like Karim and I know more about what moves markets than any AI robot will ever understand.
Maybe thatās a bold statement, but the most recent data continues to support this point. Itās also what makes The War Room one of the strongest trading communities in existence.
Thatās why weāre happy to have you in the room with us. And together, weāll continue to beat the bots.