The Human-Powered Breakthrough That Will Finally Make AI Smart
Artificial General Intelligence won’t be built on scale alone. It demands something harder: better data, fairer ecosystems, and sharper evaluations.
That’s the vision of the 2077AI Open Source Foundation, a global initiative advancing AGI through open-source benchmarks, sustainable infrastructure, and a thriving AI innovation ecosystem. Their work includes benchmarks for document parsing, logical rigor, graduate-level reasoning, and real-world agent performance — setting the standards that define AI’s next frontier.
Now, PublicAI is joining the mission alongside visionary institutions including Harvard, Stanford, MIT — and through the launch of Project EVA, the world’s boldest attempt yet to test AI’s ultimate limits.
Project EVA: A $10.24M Global Challenge
The $10.24M Project EVA (Evaluate AI) competition challenges the world’s brightest minds — “Reality Architects” — to push today’s most advanced AI models to their breaking point. Instead of solving problems, participants will design novel, adversarial challenges that expose the failure modes, blind spots, and reasoning limits of AI.
Because the truth is clear: traditional benchmarks are no longer enough. If we want AI capable of serving humanity, we must test it in ways as complex, creative, and unpredictable as humanity itself.
Strategic Partners
- RemoExperts provides the official registration and talent hub, building the new home for elite AI innovators.
- Abaka AI joins as co-organizer, supplying technical infrastructure and expertise.
- PublicAI powers the Human Layer, mobilizing real people to source the culturally nuanced, human-verified data.
We’ve spent years building the Human Layer of AI: decentralized data infrastructure powered by 3M+ contributors across 200+ countries. By compensating experts to provide authentic, human-verified data, PublicAI delivers the one input AGI can’t fake, and that’s the human experience.
Together, this collaboration means:
- High-quality training inputs. PublicAI’s global contributor base complements 2077AI’s datasets like Matrix (bilingual pre-training) and PIN (multimodal learning), grounding them in cultural nuance and lived expertise.
- Stronger evaluation. PublicAI will help source “hard questions” from real users to sharpen benchmarks like KOR-Bench and SuperGPQA, exposing AI’s blind spots where it matters most.
- Sustainable ecosystems. By aligning incentives and rewarding contributors, PublicAI extends 2077AI’s vision of a sustainable, human-centered innovation cycle.
Raising the Ceiling of AI
This partnership isn’t about incremental gains. It’s about building the foundations for AI that doesn’t just compute faster, but understands deeper.
PublicAI brings the human layer. Abaka AI brings the engine. 2077AI brings the vision.
Together, through Project EVA, we move closer to AI that isn’t just more powerful, it’s more human.
Learn more about Project EVA: https://www.2077ai.com/Project-eva
Register Now: https://www.remoexperts.com/login?inviteCode=NjhiZmFjOWIyYzRjM2ZiMzgwNDAyOWNl
About PublicAI: PublicAI is building the Human Layer of AI — a decentralized data network of 3M+ contributors providing authentic, human-verified data to accelerate the development of safe and human-centered AGI.
About 2077AI: 2077AI Open Source Foundation is a global initiative advancing AGI through open-source benchmarks, sustainable infrastructure, and a thriving innovation ecosystem. Its inaugural competition, Project EVA, sets a new standard for AI evaluation worldwide.
About Abaka AI: Abaka AI provides high-quality data solutions and technical infrastructure to power next-generation AI systems. As a co-organizer of Project EVA, Abaka AI supports building a sustainable and human-centered AI ecosystem.