Decisions
A lightweight log of significant decisions — tech stack choices, investment stances, business direction, and operational calls. Reference for context and consistency.
All decisions
Astro for burnsideco.com dashboard
Dashboard was static HTML. Needed a structured, maintainable rebuild with data-driven content for stocks, subscriptions, and learning paths.
Use Astro with static output. No framework adapters. Data files in TypeScript. No remote fonts or CDNs.
Live at burnsideco.com via GitHub → Cloudflare Pages. Build currently generates 21 routes.
Copilot Studio learning vehicle: Morning Agent
Needed a real project to build Copilot Studio skills against, not tutorials. Goal: portfolio piece + interview-ready demo.
Build a personal version of an executive intelligence agent — stock prices, news briefs, thesis recall, portfolio summary — delivered into Teams.
Defined 6-phase learning plan. PL-900 cert in Phase 5.
Tideform positioning
AI reshaping BA and business analyst roles. Need a differentiated offering before the market commoditises basic prompt consulting.
Position Tideform around AI system integration for SMBs — not generic prompt work. Copilot Studio + OpenClaw + agentic workflows.
McIntosh as first client case study in progress.
IREN position sizing post-Microsoft contract
Microsoft $9.7B / 200MW GB300 contract confirmed. IREN pivot from BTC mining to AI cloud is now the primary thesis driver.
Accumulate selectively below $55. Treat like early RKLB-style long hold. Track Horizon deployments as proof points.
RKLB remaining shares are hold-through-volatility core
Patrick sold 89 RKLB shares at $148 on May 28, received $13,167 USD and transferred proceeds out to pay the majority of credit-card debt.
Hold the remaining 70.5175 RKLB shares through ups and downs as free-carry/core. Old Tranche A/B sell-all plan and $64 stop-alert framework are no longer active.
Portfolio, KB, dashboard data and position-alert cron updated.
MU post-mortem: overbought means size down, not skip
MU thesis at ~$776 was strong but RL/Opus treated overbought chart as an entry blocker. MU later closed at $928 and tagged the first StockWaves target zone.
For high-conviction names with material upside targets, overbought becomes a sizing/risk-control signal rather than automatic no-entry.
L-035 filed in LEARNINGS.md; MU KB and dashboard thesis updated.
StockWaves subscription review
StockWaves (EWI) is $79/month. Usage has been irregular but recent RKLB/MU levels were important.
Keep on review. Value is justified only if levels are actively used in trade decisions and not ignored due to over-conservative reads.
To add a decision, update src/data/decisions.ts and rebuild. Keep entries factual: context, the decision made, and outcomes as they develop.