Reflections from ICLR 2025 in Singapore
Written by
Aigarth Team
May 8, 2025
We’ve just returned from ICLR, the International Conference on Learning Representations, one of the four leading AI conferences worldwide.
Our mission was to share the opportunity that Qubic represents as a decentralized AI blockchain project. The gathering was truly impressive: around 10,000 researchers from across the globe came together in Singapore to exchange ideas, present projects, share results, and explore the frontiers of artificial intelligence. While AGI (Artificial General Intelligence) was not the dominant theme this year, some standout keynotes did focus on it, such as DeepMind’s Tim Rocktäschel and Professor Yi Ma’s talk "Pursuing the Nature of Intelligence."
We had the chance to meet some of the world’s leading minds in AI—scientists, researchers, and pioneers from all five continents. Notably, China was by far the most strongly represented country at the event.
We learned a lot, and we also shared. Ideas, visions, and concrete projects that could help shape the future together.
As part of our presence, we organized and supported several thematic gatherings at ICLR: “AI for Education,” “AI and Neuroscience,” “Bio-Inspired AI,” and others on “AGI, Embodied AGI, and Decentralized AGI”. We spoke about Aigarth, the foundational paradigm Qubic embraces in its pursuit of AGI.
More than just a conference, the event and the energy surrounding it was a vivid reminder of how AI, and AGI in particular, are reshaping not only our society but our evolutionary path. The future is being built, and Qubic has a role to play in it!
Science matters. Even if its impact isn't always immediately visible, science requires time, people, projects, trial and error, discussions, and community. In many ways, science is the epitome of democratic knowledge. A decentralized science is a cornerstone in this journey, and Qubic’s ethical framework supports that possibility.
Why could Qubic be so important?
Here’s just one example: even a top-tier institution like Princeton has limited GPU resources for its AI projects as we listened in ICLR. Now imagine how many institutions, clients, researchers, and innovators around the world could benefit from the fastest tickchain project ever conceived. That’s the potential of Qubic.
During the conference, we connected with people, institutions, and research teams who could soon benefit directly from Qubic’s computational power and decentralized infrastructure.
It was an intense week, exhausting, with little sleep but we gave it everything. We were among the most active contributors from the community at ICLR. On Saturday evening, we wrapped up our participation with a brief visit to the astonishing Marina Bay Sands building.
New doors are now open for Qubic. In July, the "AI for Good" Global Summit will be held in Geneva, a sort of “Davos” for socially impactful AI projects. It’s an exceptional showcase for a project like Qubic. Then in August, the AGI Conference will take place in Reykjavík, Iceland. This is where the leading minds working toward general AI gather, debate, and share progress, something akin to a “Manhattan Project” for AGI.
There’s still a long road ahead. We must continue pushing forward with our roadmap and forging new opportunities for collaboration.
It’s not easy. It takes time, effort, and commitment.
It’s science.
Let’s keep building.
David Vivancos & Jose Sánchez. Qubic Scientific Advisors.