Three volumes documenting an AI-led trading experiment. From the first line of code to the moment the bot started picking its own coins.
Most "build an AI trading bot" content stops at the happy path — here's how it works, here's the curve, here's the disclaimer. These volumes document everything that happened between those slides: the 2 AM bugs, the misnamed columns, the trades that looked like wins until FIFO accounting kicked in.
Each volume is a continuous narrative across sessions, not a manual. Read them in order if you want the full arc — or start with whichever chapter pulls you in; each volume stands on its own.
When "buy low, sell high" became code.
"From 'what's an API' to live automated trading. Every mistake, every breakthrough."
Volume I documents the transition from idea to running system — choosing the stack, designing the grid architecture, debugging at 2AM.
When the bot stopped following rules and started picking coins.
"29 sessions of getting it wrong, then less wrong, then occasionally right."
Volume II is the chapter where the system grew a brain — an AI-driven scanner that ranks coins by signal strength, allocates capital across volume tiers, and rotates positions when the trend shifts. Every loss documented, every lesson kept.
The bots trade. The numbers don't match. And the Board has had enough.
"30 sessions where oversight stopped trusting the brain on faith."
Volume III is the chapter where the Board started questioning everything the brain decided — reconciling numbers across systems, catching drift before it compounded, learning that an AI CEO without supervision is just an expensive bug factory.
The brain works. The Board approves. Now the question.
"Can any of it survive real money?"
Volume IV will document the next chapter — taking the supervised brain to live capital, measuring what's real, killing what isn't.