OpenAI Teams Up with Longevity Startup to Train GPT-4b for Extending Human Life By Elets News Network - 18 January 2025

OpenAI

OpenAI has stepped into the world of biological research, teaming up with longevity startup Retro Biosciences. Together, they’ve trained an AI model named GPT-4b, designed to improve stem cell production efficiency. According to MIT Technology Review, this collaboration signals a shift in how AI might reshape scientific discovery.

AI’s earlier scientific spotlight was dominated by Google DeepMind’s AlphaFold, a programme that predicts protein structures and won a Nobel Prize last year. Now, OpenAI has joined the race with a protein-engineering model that claims to visualise proteins capable of transforming regular cells into stem cells. Unlike AlphaFold, which forecasts protein shapes, OpenAI’s model focuses on biological research applications and has even outperformed humans in certain tasks.


This project is OpenAI’s first venture into biological data and raises the question: can AI fuel breakthroughs in scientific research? If successful, this could bring us closer to the realm of Artificial General Intelligence (AGI). Sam Altman, OpenAI’s CEO, recently expressed confidence in the company’s ability to build AGIs, describing them as tools that could revolutionise scientific discovery and innovation.

GPT-4b was trained on a dataset of protein sequences from multiple species, along with interaction data. While this dataset is significant, it’s smaller compared to what OpenAI uses for its flagship chatbots. GPT-4b is essentially a small language model built for specialised tasks with a targeted dataset.

Also Read :- Microsoft’s New MatterGen AI Model Can Help Create New Materials

The collaboration began a year ago when Retro Biosciences approached OpenAI. Altman, already a key investor in Retro with a personal funding of $180 million, facilitated the partnership. Retro aims to extend the average human lifespan by a decade and sees potential in using AI to push its mission forward.

Using suggestions from GPT-4b, researchers made modifications to two Yamanaka factors, boosting their effectiveness more than fiftyfold. These factors, four special genes, are known for their ability to transform adult cells into a stem cell-like state, where they can grow into almost any type of cell. Think of it as hitting the reset button on a cell, returning it to its original state. The project’s technical lead included OpenAI’s John Hallman and Retro’s Rico Meini.

Related News