Paul, Weiss Waking Up With AI

In this episode, Katherine Forrest and Scott Caravello examine one of China's approaches to achieving artificial general intelligence (AGI), drawing on reports from Georgetown's Center for Security and Emerging Technology (CSET). They discuss the country's focus on embodied AI and robotics as a potential path to AGI, multilevel government initiatives supporting this development, a large-scale social simulator project in Wuhan, and significant investments in power grid and data center infrastructure. ## Learn More About Paul, Weiss’s Artificial Intelligence practice:https://www.paulweiss.com/industries/artificial-intelligence

In this episode, Katherine Forrest and Scott Caravello trace how a "vibe-coded" project became Moltbook, a social network for AI agents. Our hosts unpack its lobster-themed lore and early community drama, consider whether the site represents truly autonomous agent activity or human direction, and assess the cybersecurity risks. ## Learn More About Paul, Weiss’s Artificial Intelligence practice:https://www.paulweiss.com/industries/artificial-intelligence

In this episode Katherine Forrest and Scott Caravello take us down “memory lane” to explain the importance of high bandwidth memory (HBM) and RAM to AI development. Our hosts also give us a rundown of potential challenges ahead, unpacking developments in the market for memory, including plans for additional capacity and lobster-style RAM pricing. ## Learn More About Paul, Weiss’s Artificial Intelligence practice:https://www.paulweiss.com/industries/artificial-intelligence

In this episode, Katherine Forrest and Scott Caravello explore small language models (“SLM”) and their potential implications for task specialization, speed, and confidentiality. Our hosts also share some recent research covering expectations surrounding SLM adoption and growth. ## Learn More About Paul, Weiss’s Artificial Intelligence practice:https://www.paulweiss.com/industries/artificial-intelligence

In this episode, Katherine Forrest and Scott Caravello unpack OpenAI researchers’ proposed “confessions” framework designed to monitor for and detect dishonest outputs. They break down the researchers’ proof of concept results and the framework’s resilience to reward hacking, along with its limits in connection with hallucinations. Then they turn to Google DeepMind’s “Distributional AGI Safety,” exploring a hypothetical path to AGI via a patchwork of agents and routing infrastructure, as well as the authors’ proposed four layer safety stack. ## Learn More About Paul, Weiss’s Artificial Intelligence practice:https://www.paulweiss.com/industries/artificial-intelligence