Developer gave an LLM $10k to run a real portfolio — full stack explained
demian_ai · x · 2026-09-26
An experiment where the author gave an LLM $10k of real capital plus a research toolkit, tasking it with running an actual portfolio rather than picking stocks.
- Inputs: X signals and web search (hand-picked top voices + automated search), Perplexity research, Tavily, company filings, earnings calls and analyst targets, news, technicals/catalysts, and private reports.
- Universe: 223 stocks across AI infrastructure, energy, and robotics.
- Decisions: Each session, MiniMax M3 (served via Nebius) decides to buy, hold, add, trim, or avoid. For every stock it outputs a target weight, 1/5/20-session forecasts, a stop loss, a profit-taking ladder, a time stop, plus rationale, counterargument, and invalidation triggers.
- Risk: A deterministic risk engine gets the final word, checking concentration and other constraints.
Related event: Developer gives LLM $10,000 to autonomously manage a portfolio(3 posts)→
More from coding & agent
- Perplexity CEO shows agent building browser-playable 3D zombie shooter in hours with zero 3D assets — AravSrinivas · 2026-11-11
- The complete harness engineering roadmap: how OpenAI shipped ~1M lines with zero hand-written code — blaizedsouza · 2026-11-11
- Micius-Flash: 27B Multimodal Model for Hardware and Semiconductor Agents — zhiman-ai · 2026-11-11
- Viral 'the harness is the machine' thread touts unverified Claude Fable/Mythos 5.2 9-equation architecture — blaizedsouza · 2026-11-11
- Compaction creates a new failure mode for coding agents: critical bug details can vanish mid-fix — blaizedsouza · 2026-11-11
- What 40+ coding-agent harnesses do between your prompt and the model, dissected — blaizedsouza · 2026-11-11