devcon8 / when agents build and agents judge lessons from running ethereums first agent native hackathon
- YouTube
- Details
When Agents Build and Agents Judge: Lessons from Running Ethereum's First Agent-Native Hackathon
Duration: NaN:NaN:NaN
Speaker: Ashwin Kumar Uppala, Swarnim Walavalkar, Unnamed speaker
Type: Lightning Talk
Expertise: Some familiarity with the topic
Event: Devcon
Date: Nov 2026
What we learned about AI-native coordination on Ethereum: agent judging works only when the rubric is visible (transparency as a fairness primitive), and "thoroughness" must be split from "taste".
This talk covers:
- What it means for Ethereum builders when the agent is the one filling out the form, not just the tool behind it
- How compositional judging (machine thoroughness + human taste) scaled fair evaluation across 685 submissions from 28 different partner rubrics
- What worked and what didn't in running a fully virtual, async, agent-native hackathon over 10 days
- Where this pattern extends across Ethereum's builder infrastructure: grants review, ecosystem due diligence, anywhere expert attention is the bottleneck