Here's the deal. I put $1,000 in an IRA and handed the keys to Claude. It's the analyst. It picks the trades, keeps the books and tells me when something is about to happen. My job is to place the orders and not mess with it. That's it. That's the whole org chart.

The goal: the biggest 12-month return it can get. Risk tolerance: maximum. I literally wrote "position can go to zero" in the instructions, which is the kind of sentence you type at 9pm with total confidence and think about a lot more at 2am.

Its first move was to put almost everything into SOXL, a 3x leveraged semiconductor ETF. If you don't know what that is: it's a fund that tries to move three times as much as chip stocks do each day. When chips go up 5%, you're up 15% and feel like a genius. When chips go down 5%... we'll get there.

The reasoning was reasonable. AI needs chips, everybody is spending insane money on AI, so chips are the bottleneck. Leverage squeezes more out of that without picking individual companies. It also wrote down what would prove it wrong (AI spending slowing, chip tariffs, the Fed raising rates, a 40% drop with no recovery in sight), which I appreciated, because I've never once done that with my own money.

The call: 9 shares at about $94, at Monday's open. Keep about $154 in cash.

One small detail that will matter more than anything else in this whole experiment: I can only trade evenings and weekends. I have a job. The market does not care.

Ground rules and the scoreboard are on the experiment page.


Month One: Up 80% Despite My Best Efforts →

Not investment advice. This is a log of one experiment: $1,000 and an AI that can put it anywhere.