GGlassboxtraderinglassboxtrader.hashnode.dev·1d ago · 3 min readI Simulated 30x Leverage So You Don't Have To (It Went to Zero)Part 6 of an honest build-in-public series — the wins, the dead ends, and what actually works. Previously: Running Local LLMs on a Potato: AI on a 6 GB GPU Every new trader hears the same whisper: le00
GGlassboxtraderinglassboxtrader.hashnode.dev·4d ago · 3 min readRunning Local LLMs on a Potato: AI on a 6 GB GPUPart 5 of an honest build-in-public series — the wins, the dead ends, and what actually works. Previously: My Entire Setup: One PC, a 6 GB GPU, and a NAS The assumption is that "AI" means a giant GPU00
GGlassboxtraderinglassboxtrader.hashnode.dev·Jul 3 · 3 min readMy Entire Setup: One PC, a 6 GB GPU, and a NASPart 4 of an honest build-in-public series — the wins, the dead ends, and what actually works. Previously: The Bug That Silently Faked All My Backtest Results If you picture "AI trading bot" and imag00
GGlassboxtraderinglassboxtrader.hashnode.dev·Jun 30 · 3 min readThe Bug That Silently Faked All My Backtest ResultsPart 3 of an honest build-in-public series — the wins, the dead ends, and what actually works. Previously: The Uncomfortable Truth: My Trading Bot Doesn't Beat Buy-and-Hold The scariest bugs aren't th00
GGlassboxtraderinglassboxtrader.hashnode.dev·Jun 26 · 4 min readThe Uncomfortable Truth: My Trading Bot Doesn't Beat Buy-and-HoldIn the first post I dropped the spoiler: my active trading bot doesn't beat simply buying and holding an index. This is the post where I show you the receipts — and how I almost fooled myself along th00