hackathon
Won THE MAVERICK at Podium Hackathon
Built AIMy, an AI shopping assistant for hardware stores. 300+ applicants, 100+ builders, 38 teams — we took the top prize.
Hackathon AI Winner
Hackathons, meetups, and what I'm building in public
Built AIMy, an AI shopping assistant for hardware stores. 300+ applicants, 100+ builders, 38 teams — we took the top prize.
Threw my name in the hat with 5 million other fans for the 2028 LA Olympics. Fingers crossed!
Key insight: Spec > Syntax. TDD is non-negotiable. Context engineering is critical for AI performance.
2026 is the year of AI orchestration. Excited to bring together powerful AI tools in seamless automated workflows.
8 Sumo oranges, roasted lamb, and time with friends. Reminded that friendship, knowledge, and food are the most valuable things.
Connects Azure DevOps with GPT-4o to trace bugs to their source commits via git blame. Zero manual effort traceability.
Captures 9 years of institutional knowledge. New bugs are semantically matched to historical context and routed to the right developer.
Super Bowl XLI champion shared his framework: Resilience 1.0 (survival) vs Resilience 2.0 (adaptation and growth).
Built Quorum: Your Virtual Board of Directors. 4 AI executives debate your business decisions via text message.
Turns napkin sketches into professional change orders. AI interprets rough drawings and generates Bills of Materials instantly.
Released the code for my AI system that automates morning planning. Aggregates weather, calendar, fitness, and priorities while you sleep.
Hot-swappable agents communicate through an Orchestrator. Failure in one API won't bring down others — scales from 3 to 30 agents.
Built a system to handle the 35,000 daily micro-decisions. Weather affects fitness, fitness affects calendar, calendar affects study time.
Verizon, X/Twitter, Yahoo/AOL, Microsoft 365 — all down this month. The cloud is just someone else's computer.
After cutting support staff 50%, Salesforce admits LLMs 'drift' after 8+ instructions. Now hiring humans again for high-touch roles.
Airlines have a dedicated counter for travelers who booked the wrong country. Same problem happens with LLMs: missing context = wrong destination.
AI doesn't need 'please.' Your coworkers do. Don't let prompting machines train you to speak with transactional coldness.
My GPU simulates protein dynamics for Alzheimer's, Cancer, and Parkinson's research. A small regular contribution to make the world better.
Learning to automate VM deployments on my home server. API-driven setup speeds up my project deployments significantly.
Playtested upcoming games and discussed how AI assists 3D modeling and development. Tools empowering creators, not replacing them.
Kaden Wilkinson on Model Context Protocol. Up to 98% more efficient LLM usage. Shift from models that 'know' to agents that 'do'.
Align with stakeholders, create high-value tests, master pipelines, build feedback loops. QA teaches you to ship quality code.
Matthew D. on getting data from any website and warehousing it. Learned about robust scraping tools and LLM integrations.