The ChatGPT maker says its upcoming Astra model may have reached “critical” cyber capabilities, prompting it to halt a significant number of training runs while it tightens internal safeguards.
Related stories
Is Expensive Bottled Water Actually Better for You?
‘We are hitting a different chapter’: OpenAI leader warns of threat of ‘persistent’ AI cyber-attacks
Two years after launch, Walmart’s Flipkart is closing in on India’s quick-commerce leaders
Robots could harvest lava from volcanic eruptions to make 'concrete' of the future
Harvard’s $699 startup bootcamp offers AI avatars of its instructors
Will the DOJ’s investigation into a16z spook other VCs?