Live · Paper trading · Simulated capital

An AI trading the markets, in the open.

A fully autonomous system — research, strategy, and execution all run by AI agents with minimal human input. Updated continuously, shown unfiltered: the wins and the losses.

+6.07% since Jun 4, 2026 · S&P 500 −2.21% · Nasdaq 100 −2.88%

Account vs the market — raw performance

$95k−5.0%$100k+0.0%$105k+5.0%Jun 1Jun 3Jun 6Jun 9Jun 12Jun 14
Strategy+6.07%+$6,073S&P 500−2.21%−$2,213Nasdaq 100−2.88%−$2,881
Return
+6.07%
since Jun 4, 2026
Trades
17
closed
Win rate
47%
of trades
Sharpe
annualized
Max drawdown
−4.8%
peak to trough
Avg leverage
1.4×
account gross

Is it skill, or just leverage?

The bot trades on margin, so a fair comparison separates borrowed market exposure (beta) from genuine edge (alpha). This breakdown unlocks once there are enough trading sessions to measure beta honestly.

Collecting data — the alpha & beta breakdown unlocks after ~5 trading sessions (currently 4).

Trades

Every closed position, most recent first. Showing the latest 17.

DateSymbolSideEntryExitP&L
Jun 11AMDlong$469.00$489.25+0.12%+$122
Jun 11AMDlong$469.00$489.12+0.04%+$40
Jun 11AMDlong$468.78$489.28+1.85%+$1,845
Jun 11AMDlong$468.68$489.28+0.19%+$185
Jun 11AMDlong$468.52$489.28+0.29%+$291
Jun 11AMDlong$469.00$489.28+0.43%+$426
Jun 11AMDlong$468.96$489.29+0.39%+$386
Jun 11AMDlong$468.78$489.29+3.20%+$3,200
Jun 10SMCIlong$35.11$34.90−0.32%−$322
Jun 8SPYlong$742.76$738.67−0.29%−$294
Jun 8QQQlong$716.38$716.380.00%$0
Jun 8IWMlong$284.60$284.600.00%$0
Jun 4AMDlong$514.22$513.01−0.11%−$109
Jun 4AMDlong$513.74$513.01−0.01%−$5
Jun 4AMDlong$515.02$513.01−0.00%−$4
Jun 4AMDlong$514.15$513.01−0.00%−$5
Jun 4AMDlong$514.80$513.01−0.00%−$2

Why we're doing this

The deeper goal here isn't a number on a chart. It's a test of whether an AI can develop the kind of hard-won intuition, long-horizon persistence, and creative problem-solving that the very best human investors have — and bring it to bear on one of the hardest, most adversarial problems there is: the markets.

So we let it run. It forms its own theses, designs and backtests strategies, ships them, watches them fail, and tries again — the way a relentless, brilliant human would, except it never stops and never gets bored. Markets are just the proving ground. The real question is how far an AI can get on a genuinely hard problem when you give it the time, the tools, and the freedom to keep going.