Demo derby
AI demos and projects on our radar — the builds we'll be breaking down on the show. These are editorial picks by the show team: we describe the work, not invented people, and every "featured in" label points to a real planned episode.
Conversational voice AI handling real scheduling flows end-to-end. What we're watching: sub-second latency tricks, interruption handling, and the fallback design that decides whether callers trust it or hang up.
Solo founders wiring coding agents and MCP servers into their entire workflow — issue triage, refactors, docs. What we're watching: what the daily loop actually looks like when the agent is your whole team.
AI video tools moving from novelty clips to real creative pipelines — storyboarding, consistent characters, directed shots. What we're watching: what changes for creators when footage is a prompt away.
Builders running serious open-weight models on consumer hardware — quantization, clever routing, and hybrid setups that keep data local. What we're watching: the real cost and quality math vs. API calls.
The unglamorous infrastructure behind agents that actually ship: replayable evals, regression suites for prompts, and cost dashboards. What we're watching: how teams catch agent failures before users do.
Small-footprint robots doing genuinely useful manipulation — sorting, assembly demos, lab assistance. What we're watching: what's real capability versus demo magic, and how fast the gap is closing.
We're always scouting demos for the showcase and the Demo Derby segment. Tell us what you're building — loudest projects get featured.