P-1 AI, a newly emerged company focused on developing artificial general intelligence (AGI) for engineering, has announced a $23 million seed financing round led by Radical Ventures. Founded by Paul Eremenko, a former CTO at Airbus and United Technologies, and Aleksa Gordić, an ex-DeepMind researcher, P-1 AI aims to create an AI agent named Archie that can automate cognitive tasks currently handled by human engineers in the design of physical systems. Core Focus: Cognitive Automation for Engineering Tasks Archie's primary functions include distilling key design drivers from initial requirements, developing product concepts and derivatives, performing first-order design trade-offs, and selecting appropriate engineering tools for detailed design. Unlike conventional AI models, Archie is not intended to replace or compete with existing engineering tools; instead, it is designed to complement them by handling repetitive and time-consuming tasks, thereby enhancing team productivity. Addressing Data Scarcity One of the main challenges in engineering AGI is the scarcity of training data. Unique design sets are few, and the ones that exist are often proprietary and lack comprehensive multi-physics models. To overcome this, P-1 AI generates large, synthetic design datasets that are physics-based and informed by supply chain data. These datasets are used to train an AI architecture that implicitly learns the underlying physics of the product domain, enabling it to perform quantitative and spatial reasoning tasks effectively. Initial Deployment and Future Plans Archie is scheduled to be deployed later this year to assist in engineering data center cooling systems. The company plans a rapid expansion into other areas, including industrial systems, building systems, automotive and heavy machinery, and aerospace and defense. The ultimate goal is for every major industrial engineering team to have an Archie as a team member, starting with routine tasks and gradually moving towards more complex, innovative projects. Leadership and Vision Paul Eremenko, P-1 AI's co-founder and CEO, envisions Archie initially tackling the "dull and repetitive" aspects of engineering work. He sees this as a way to enhance team productivity and, through continuous learning from real-world data and feedback, evolve into a more sophisticated assistant capable of designing systems unknown to current human capabilities. Aleksa Gordić, co-founder and head of AI, emphasizes the need for fundamental breakthroughs in both data and models to achieve their ambitious goals. Gordić is a developer of llm.c, a significant contribution to the field of language modeling, and has extensive experience in generative AI platforms and community building. Investor Confidence Molly Welch, a partner at Radical Ventures, expressed strong confidence in P-1 AI's mission. She highlighted the team's deep understanding of customer needs and their world-class expertise in AI and physics-based modeling. Joining Radical Ventures in this seed round are investors like Village Global, Schematic Ventures, and Lerer Hippeau, along with notable AI experts such as Jeff Dean, Peter Welinder from OpenAI, and Bob van Luijt from Weaviate. Company Background and Team P-1 AI is headquartered in San Francisco's Mission district and is actively recruiting applied AI researchers and engineers. The company's leadership includes: Paul Eremenko (co-founder & CEO): Known for his contributions to advanced design and manufacturing tools at Airbus and United Technologies, and a key figure in launching significant initiatives at Google and DARPA. Aleksa Gordić (co-founder & head of AI): A prominent AI researcher with key contributions to language models and generative AI platforms, and a founder of the AI Epiphany community. Adam Nagel (co-founder & head of engineering): An experienced director of engineering at Airbus-Silicon Valley and associate director at the Model-Based Digital Thread Center, with a history of building versatile model-based toolchains across multiple industries. Industry Insights Industry insiders view P-1 AI's approach as groundbreaking, addressing a critical gap in the application of AI to physical systems. The ability to generate synthetic datasets and train AI models on them could revolutionize the way engineering teams operate, making them more efficient and capable of tackling previously insurmountable challenges. The investment from prominent firms and individuals underscores the potential and the high stakes involved in developing this AI architecture. P-1 AI's mission to develop engineering AGI has significant implications for the future of design and manufacturing. By automating routine tasks and providing cognitive support, Archie could free up engineers to focus on higher-value, more creative work, ultimately driving innovation and efficiency in industrial engineering.