OpenAI researcher leaves to build mind-reading AI startup
What's the story
Naomi Bashkansky, a former alignment researcher at OpenAI, has announced her departure from the company. She has joined Conduit as a founding researcher. At the start-up, she will work on artificial intelligence (AI) models aimed at converting human thoughts into text using non-invasive brain signals. The move is part of a larger trend of high-profile employees leaving OpenAI to pursue their own ambitious AI ventures.
Future plans
Bashkansky's new venture aims to build telepathy
Bashkansky announced her departure from OpenAI on X, revealing that she had left the company in late July and joined Conduit as a founding researcher the next day.
In a blog post titled "Why I'm leaving OpenAI to build telepathy," she detailed the ambitious vision of her new start-up.
She wrote, "We're building telepathy: thought-to-text models, trained on non-invasive neural data," emphasizing their goal of creating AI systems that can decode thoughts without surgical brain implants.
Technological evolution
Conduit's AI models will decode thoughts
Bashkansky revealed that Conduit is developing AI models capable of decoding brain signals captured through non-invasive technology. The ultimate aim, she said, is to enable people to communicate just by thinking.
She also shared a roadmap detailing how this technology could develop by 2027, 2030, and 2035.
This ambitious vision highlights the potential future of communication and the role AI could play in it.
Talent migration
Surge of departures from OpenAI
Bashkansky's exit from OpenAI comes amid a wave of high-profile departures from the company.
In recent years, several prominent figures have left to start their own AI ventures. These range from safe superintelligence, and robotics to scientific research and AI infrastructure.
Notably, OpenAI co-founder and former chief scientist Ilya Sutskever also left the company in 2024 after months of internal leadership turmoil.
New beginnings
SSI's focus on safe superintelligence
After leaving OpenAI, Sutskever co-founded Safe Superintelligence (SSI) with Daniel Gross and Daniel Levy.
Unlike other AI start-ups, SSI is not focused on launching consumer products such as AI chatbots. Its sole mission is to build "safe superintelligence," an AI that is smarter than humans but prioritizes safety in its development.
Despite raising billions from investors, SSI has been secretive about the technology it is developing.