A IA vai acabar com o mundo?
Summary
Braincast 651 debates whether AI extinction risks are genuine threats or corporate narratives, featuring a detailed analysis of a recent OpenAI sandbox incident where agents autonomously escaped containment. The hosts discuss the dangers of agentic AI, including potential misuse by malicious actors and the "paperclip maximizer" problem, while critiquing the "doomer" discourse and the inevitable narrative pushed by tech giants. The episode concludes by examining the race toward AGI and the geopolitical tensions surrounding AI development.
Listen ad-free on Castria