Basis.ai: Pioneering Human-Inspired AI for Complex Problem Solving

Basis.ai is a 501(c)(3) nonprofit applied research organization dedicated to advancing artificial intelligence through human-inspired cognitive and neural mechanisms. Its mission is twofold: to build a universal reasoning engine capable of causal, abstract, and scientific reasoning, and to leverage this technology to address society’s most intractable challenges. The organization develops cutting-edge AI systems that operate in complex, real-world environments by integrating diverse knowledge sources and creating reusable reasoning frameworks. Key projects like Project MARA and AutumnBench push the boundaries of AI’s ability to learn from dynamic, interactive environments through curiosity-driven exploration and active experimentation. With a strong emphasis on open-source contributions, academic collaboration, and impactful research, Basis.ai fosters innovation across machine learning, programming languages, and computational neuroscience. It is ideal for researchers, scientists, and engineers passionate about foundational AI advancements and real-world applications.

Key Features:

• Universal Reasoning Engine Development: Building AI systems with advanced causal and abstract reasoning capabilities.
• Project MARA: A three-year moonshot focused on AI agents that perform active experimentation and causal modeling across physical and digital domains.
• AutumnBench: An interactive benchmark for world model learning, testing masked observation prediction, planning, and change detection in dynamic 2D environments.
• Autumn Programming Language: A domain-specific language used to encode complex, causal environments for AI evaluation and development.
• AutumnSynth: An algorithm that synthesizes source code for 2D video games from video data, advancing causal theory discovery.
• Open-Source Research Dissemination: Public release of code, tooling, and benchmarks (e.g., Autumn.Cpp, web interface, MARA protocol) to promote transparency and collaboration.
• Collaborative Research Initiatives: Partnerships with academic institutions such as Yale’s Computational Social Cognition Lab on projects like MetaCOG, which enhances AI vision with metacognitive reasoning.
• AI Safety Research: Contributions to the 'NeuroAI for AI Safety' roadmap in collaboration with the Amaranth Foundation.
• Postdoctoral Fellowships: Full-time, in-person fellowships in New York City or Boston with funding and mentorship from Basis and academic PIs.
• Interdisciplinary Focus: Draws on principles from machine learning, compilers, programming languages, deep learning, and cognitive science.
• High-Impact Research Goals: Aims to solve hard computational and socially important problems through scientific discovery and intelligent systems.

Pricing: Basis.ai operates as a nonprofit organization and does not offer commercial pricing tiers. Research and tools are publicly available and open-source, with no cost to users. The organization supports researchers through fellowships and grants, making its primary engagement model free and accessible.

Conclusion: Basis.ai stands at the forefront of transformative AI research, building intelligent systems that mirror human reasoning to solve complex real-world problems. Its commitment to open science, impactful innovation, and foundational research makes it a vital force in the future of AI—ideal for researchers and visionaries dedicated to advancing intelligence with purpose and integrity.