Summary
The All About AI channel delivers a 30-day performance review of three autonomous AI trading bots running live on Polymarket, Kalshi, and Hyperliquid, all deployed on VPS servers. The experiment covers strategies built with AI and machine learning, ranging from a simple arbitrage lottery to a quantitative weather-event system and an ML model trained to predict 5-minute Bitcoin price direction.
Results diverge sharply by strategy. A 1-cent Bitcoin lottery bot returned 3.5% net over three weeks but experienced a $63 maximum drawdown — profitable but noisy. A quantitative weather-trading system was the standout, returning 28.9% ($121 net on roughly a $400 account) with a 58% win rate and only a $10 maximum drawdown over 25 market days. A custom ML model targeting BTC up/down markets posted a 19% return and 72% win rate, though at very low frequency — 32 trades in a month, roughly one per day — with a single strategy adjustment mid-run after a notable drawdown cluster.
The creator also previews a live deployment of the Kalshi weather strategy, showing autonomous order placement in real time. Looking ahead, the channel plans to expand these approaches into a micro-scale “AI agent hedge fund” that would allow agents to invest in fractional amounts as small as one-millionth of a dollar, targeting a new class of agent-native financial participation unavailable through traditional funds.
📺 Source: All About AI · Published August 24, 2026
🏷️ Format: Workflow Case Study
