We can’t predict the capabilities of large language models. Even specialized AI could be very bad for humanity. AI companies are racing as fast as possible to build AGI, which will be far more dangerous than specialized systems.
Jeff Bezos on the risks of general-purpose AI
