Jed McCaleb, the co-founder of major cryptocurrency projects Ripple and Stellar, is spearheading an ambitious effort to create Artificial General Intelligence (AGI). McCaleb is committing substantial capital to develop AI systems modeled directly on the operational principles of the human brain.

The Astera Institute's Neuroscience Approach to AI

McCaleb is funding this initiative through the Astera Institute, a nonprofit organization that has recently pivoted its focus toward brain-inspired AI architectures. The goal is to achieve AGI—the point where AI can match human-level task performance.

Researchers at the institute are currently planning experiments involving mice, monkeys, and eventually humans. These studies aim to record neural activity patterns as subjects perform tasks, such as navigating mazes.

Mapping Brain States to Actions

The immediate objective is to create a comprehensive library mapping specific brain states to corresponding perceptions and actions. This data will then be translated into code, forming the basis for a novel AI system.

This research could establish a powerful feedback loop: brain experiments will inform new AI architectures, which in turn will suggest new hypotheses for neurological testing.

Science Fiction Ambitions: Brain-Computer Interfaces

The project carries a futuristic ambition involving brain-computer interfaces (BCIs). Researchers envision using BCIs not only for reading neural data but also for writing information directly into the brain.

This includes concepts like 'uploading' knowledge or inserting specific visual memories, such as the image of an apple, into a person's thoughts.

McCaleb's Vision and Funding Commitment

Forbes estimates Jed McCaleb’s net worth at $3.9 billion. He is channeling significant personal wealth into this long-term scientific endeavor. This $1 billion commitment for AGI research is in addition to $600 million already pledged to neuroscience research via the Astera Institute.

McCaleb, described as future-obsessed and capital-rich, spends several days a week at Vast, his space company based in Southern California, which aims to replace the International Space Station.

"Crypto was, in some sense, a big detour," McCaleb stated. "I’d been wanting to work on AI the whole time, but only really got the chance once I stepped back ...I think AI is going to be the most transformative thing that humans ever create. So it’s the most compelling thing to work on."

Founding and Leadership

McCaleb established the Astera Institute in 2020 alongside his spouse, Seemay Chou, a former biology professor at the University of California, San Francisco. Last year, the couple committed to donating the majority of their wealth through the Giving Pledge.

Dileep George, a former executive at DeepMind, has been hired as a key leader for the AGI effort. George previously co-founded Vicarious AI, which was acquired by Alphabet, and Numenta, an early neuroscience-focused AI firm.

Divergence from Current AI Trends

George intends to expand the Astera lab to approximately 30 researchers this year, prioritizing mission-driven scientists over those seeking massive compensation packages from large tech firms.

Astera plans to publish its research openly, reminiscent of early OpenAI before competitive pressures led to secrecy. George noted that a philanthropy-supported model avoids the distraction of constant fundraising demos required by startups.

Skepticism Towards Transformer Models

McCaleb and George share a fundamental skepticism that simply scaling transformer models—the architecture behind systems like ChatGPT—will lead to AGI. They believe the field needs to return to foundational research.

McCaleb previously suggested that AGI development will require "a lot of medium-sized breakthroughs" rather than one single major discovery. He argues that transformers primarily handle prediction, lacking crucial elements like planning, decision-making, and motivation.

The Path to Safer, Understandable AI

The Astera Institute views brain-inspired AI as a pathway toward safer artificial intelligence. Current frontier models are often opaque, making it difficult to understand their internal reasoning.

McCaleb believes that if AI functions more closely to the human brain, researchers will have a better chance of understanding its processes. This contrasts with current systems, which he describes as "this kind of abstract mathematical thing that ends up being very alien."