Liam Fedus, post-training head at OpenAI. departs to launch new startup | ChatGPT app | OpenAI Chat | OpenAI API | Turtles AI

Liam Fedus, OpenAI’s vice president of post-grad research, has announced his resignation to launch a startup focused on the application of AI in materials science. With a degree in physics, Fedus expresses his enthusiasm for combining his academic background with AI technologies. OpenAI recognizes the strategic importance of AI in science and plans to invest and collaborate with Fedus’ new venture

Key Points:

• Liam Fedus, OpenAI’s vice president of post-grad research, is leaving the company to launch an AI startup in materials science.
• OpenAI plans to support and collaborate with Fedus’ new venture.
• The startup will enter an industry that competes with the likes of Google DeepMind and Microsoft.
• Some experts remain skeptical of AI’s current ability to make breakthrough scientific discoveries.

Fedus’ decision comes amid growing interest in the application of AI in the discovery and design of new materials. Microsoft, for example, recently presented MatterGen, a generative AI model for the design of inorganic materials across the periodic table, capable of generating materials with specific properties. In parallel, Google DeepMind has developed GNoME (Graph Networks for Materials Exploration), a system that uses deep learning to predict the stability of new materials, significantly accelerating the discovery process. Despite these advances, some experts remain cautious about the ability of AI to make innovative scientific discoveries, highlighting the need for experimental validation and integration with traditional methods. Fedus’ new venture will therefore have to address these challenges, seeking to demonstrate the potential of AI in materials science.

Materials science represents a field in which AI can make significant contributions, optimizing processes and accelerating discoveries. However, the effectiveness of these technologies will depend on their ability to integrate with existing methodologies and overcome current limitations.

The scientific community is watching the evolution of these initiatives with interest, aware of the potential and challenges that AI brings to the field of materials research.