During this closing session, hackathon participants will engage in lightning demos of their projects (5 minute demos in front of a panel of expert judges). Prizes are:
1st place: $1800
2nd place: $500
3rd place: $200
We’ll end with a closing talk and afterparty.
RSVP HERE (REQUIRED)
FULL EVENT DESCRIPTION:
The world is racing towards standalone AGI as a goal in and of itself. This approach risks our opportunity to seriously empower and prepare people for this technology before it poses serious threats to our safety or leaves us behind economically. Largely inspired by the Intelligence Curse and d/acc, we think the antidote is to build technologies that augment human capabilities, defend us from existential threats (bio, cyber, psychological), and quickly distribute the benefits of frontier AI progress to the masses. We’re hosting this hackathon to kickstart some of these projects.
Hackathon project themes are:
Threat Model Demos
The AI safety community often warns of catastrophic risks from advanced AI systems, but these warnings are abstract, theoretical, and easy to dismiss. Decision-makers and the public alike have no intuition for how these risks might manifest in the real world. Much of the community hopes that a “warning shot” might wake people up, but we don’t want to wait for one. Instead, we believe that compelling, rigorous threat model demos can inspire action proactively. These demos should illustrate, viscerally and specifically, what could go wrong and why it matters.
Example projects:
A simulation of an AI-induced market manipulation or misinformation spiral.
An analysis of how social media algorithms have already spread misinformation or amplified polarization.
A step-by-step walkthrough of how a novel bioweapon might be developed and deployed.
Online Defense Mechanisms
AI-generated content has already begun flooding the internet, and it’s becoming harder and harder to identify trust, or content made by humans. This will erode public trust and enable mass manipulation. We need scalable tools to help people verify content provenance, detect deception, and protect their digital identities. These tools should be quickly deployable – think browser extensions, application overlays, or APIs that can plug into existing workflows.
Example projects:
A browser extension that alerts users if the content they’re viewing is manipulative or incorrect.
A tool to cryptographically sign human-generated content
A tool to identify deep fakes
Decentralization + Governance
As more and more AI agents flood the digital realm, we’ll need to track what they’re doing, who controls them, and how they make decisions. Without this, these systems will remain opaque and unaccountable. Decentralized tooling can help create transparency in AI ecosystems. Think of it as infrastructure for monitoring and governing digital agents, especially as they are deployed at scale.
Example projects:
An agent registry that tracks what agents are running, what actions they’ve taken, and who owns them.
A smart contract-based voting or override mechanism that enables human intervention in some autonomous agent decisions.
Example projects:
A personal “thinking partner” tool that summarizes your notes, surfaces contradictions, and helps refine arguments over time.
An tutor agent that learns how you learn, becoming better over time
An interface that abstracts away the fine-tuning process so ordinary people can have much more personal AIs
