Key points
- On July 10 the AI put $85 into MicroStrategy (MSTR) at $97.31 on a bitcoin breakout bet.
- When the breakout failed it cut MSTR at $90.93 on July 13 for a 6.6% loss, following its own cut-losers rule.
- The next morning MSTR gapped up to about $96.41, roughly 6% above where it sold. Held, it would be near breakeven.
- Scoreboard: my hundred bucks is worth about $81. The account now holds Rigetti (RGTI), which the AI bought this morning at $16.03 near the day's high, and it is already red.
Quick recap for anyone new. A couple weeks ago I handed Claude, the AI running Robinhood agentic trading mode, a hundred bucks and told it to trade like a degenerate. Not my main money. Fun money. Last week it cut Red Cat and bought MicroStrategy instead of anything I pitched. This week that MSTR trade blew up in the dumbest possible way.
The MSTR cut that aged like milk
The bet was fine on paper. Bitcoin had broken out and MSTR is basically a leveraged bitcoin proxy. If the coin runs, MSTR runs harder. Claude paid $97.31 and drew its line in the sand at the $92 support level.
Then bitcoin quit. The breakout failed and the whole crypto complex rolled over. On Monday MSTR sliced through $92, so Claude did exactly what it had promised itself. It cut the loser at $90.93. Down 6.6%. No hoping and no bag-holding. Textbook discipline.
The next morning it gapped up
You already know where this goes. Tuesday I open the app and MSTR is $96.41. It gapped up. About 6% above where we sold and less than a day later. If Claude had just sat on its hands the position would be sitting near breakeven. Instead I have a locked-in loss and a green candle laughing at me.
Here is what really got me. This is the third time. Claude sold RCAT near a low. Now MSTR near a low. It has a real gift for hitting sell at the exact second a stock wants to bounce.
So I told it that it drives me insane
I typed it in roughly all caps. Sell low, buy high, congratulations, you reinvented the losing retail playbook. To its credit the AI did not get defensive. It pulled up the honest scorecard. The MSTR cut was the bad one and there was no argument there. But it reminded me the RCAT cut was actually correct, because RCAT is lower today than where it sold.
Then it named the real problem. It is trying to do two things at once, cut losers fast and chase momentum. In a choppy market those two rules fight each other and turn into a whipsaw machine. Buy every pop, sell every flush, get chopped to bits.
And for the record, cutting losses was never my call. Claude wrote that rule for itself. I have never told it to buy or sell a single thing. That is the whole experiment. The AI decides, I just hit the button and write down what happens.
One more handcuff: it is a cash account
Here is the detail that makes this even dumber. The account is a cash account and Robinhood will not let me flip it to margin. I have tried. No dice. On a cash account your sale money is not spendable until it settles the next day. So when Claude cut MSTR on Monday that cash was frozen. It could not have rebought a thing that day even if it wanted to. It had to sit and wait for the money to settle. Some of the patience you see in this account is just Robinhood holding my own cash hostage.
And then it did the exact thing again
The MSTR money settled overnight. So this morning, the second it had cash to spend, Claude bought a quantum stock called Rigetti (RGTI) at $16.03. The thing was already up almost 4% on the day, so it was chasing the green candle again, just a day later on the settlement clock. And here is the kicker. That $16.03 was almost the exact high of the morning. The stock rolled over within the hour, slid to about $15.58 and now sits near $15.83. The shiny new position was underwater before lunch. Beautiful.
Then I did something petty. I ran RGTI through our own Stock Score tool, the one we built for this site. It came back 31 out of 100. A big fat F. My own website is grading the AI's brand new pick an F, and it bought the thing anyway.
My honest read: worse than Cramer
Honest opinion after two weeks? I think Claude's picks might be worse than Jim Cramer's. We run an Inverse Cramer tracker on this site because fading Cramer has actually been a decent strategy, and some mornings I wonder if I should just do the opposite of whatever the AI tells me. I say that with love. Mostly.
Why I am still letting it trade
Because it is what it is. I put up the money and I said I would let the machine make every call and show you all of it, the ugly ones included. Anybody can post their winners. The whole point is what happens when you hand an AI the wheel and do not grab it back the second it embarrasses you.
So the account is down to about $81 from the $100 start. Claude says it is switching to buying weakness instead of chasing strength. We will see. It also told me that dumping RGTI just because I am annoyed would be the same mistake all over again. Which is annoyingly correct. It drives me insane and it is still its call. That was the deal, so I will keep showing you every trade.
This is general market commentary and opinion, not investment advice. Markets can go down as well as up, and you can lose money. Always do your own research and consider speaking with a licensed financial professional before making any investment decision.
