Episode Details

Can AI really kill everyone by 2050, or are the doom warnings overstated? In this condensed recap of Doom Debates, host Liron Shapira and guest Casey Hart spar over artificial intelligence, machine learning, AI alignment, and the risks of superintelligence—cutting through a full-length debate into a quick, high-signal summary. Hart, an ontology expert and philosopher, argues that extinction scenarios depend on a stack of uncertain assumptions, from whether AGI arrives on time to whether misalignment actually leads to catastrophe. Shapira pushes the opposing case: that scaling laws, autonomous systems, and weak human control could make future AI far more dangerous than it looks today. You’ll also hear their take on LLMs, symbolic AI, AI regulation, and why present-day harms still matter even when the worst-case timeline is disputed. Listen now to get the key ideas in minutes.