Show one failure in your AI demo — it builds more trust than a polished screenshot
brandon_galang · x · 2026-11-08
A counterintuitive take: the part you leave out of an AI demo — where it broke — is what earns trust.
Instead of a finished screenshot, explain what the model got wrong, how you noticed, what you checked, and why you changed approach. Walking through one failure-and-fix cycle gives reviewers something concrete to evaluate, especially when your resume doesn't yet vouch for you. Put that explanation next to the demo.
More from coding & agent
- Llama App adds official agent integrations: run Pi and Codex with local models, fully offline — mervenoyann · 2026-11-08
- RustingStrata hits 160 tok/s on RTX 3060 running 35B MoE, ~2x faster than llama.cpp — Serious_Scarcity7818 · 2026-11-08
- 10-minute Sherlock Holmes anime made with ComfyUI and Claude Code, files included — shiwano · 2026-11-08
- RustingStrata hits 160 tok/s on RTX 3060 running 35B MoE, ~2x faster than llama.cpp — Serious_Scarcity7818 · 2026-11-08
- Karpathy built a curated video platform in a weekend, filtering 43M HN comments with Claude — The Decoder · 2026-11-08
- Running Multiple LLM Slots on One 7900 XTX Boosts Total Tokens/Sec — c2btw · 2026-11-08