Paul, Weiss Waking Up With AI

In this episode, Katherine Forrest and Scott Caravello break down three generative AI architectures—transformers, JEPA, and diffusion models—exploring what sets each apart and how they overlap. They also discuss Manifold-Constrained Hyper-Connections, a recent innovation aimed at improving how transformer layers communicate during training.

For the sources referenced in this episode, please see the links below:

DeepSeek AI: mHC: Ma...

In this episode, Katherine Forrest and Scott Caravello examine OpenAI's latest paper suggesting proactive policy measures to help society navigate the economic and social changes that advanced AI may bring. They unpack the paper's key recommendations and consider broader industry perspectives on managing the transition ahead.

For the sources referenced in this episode, please see the links below:

OpenAI: Industrial Policy for the I...

In this episode, Katherine Forrest and Scott Caravello examine world models and their growing relevance to enterprise AI, drawing on a recent IBM blog post that highlights use cases ranging from infrastructure management to atmospheric prediction. They also review the details that have emerged from a recent leak about Anthropic's upcoming Mythos model, including its reported advances in cyber, coding, and reasoning benchmarks.

For ...

In this episode, Katherine Forrest and Scott Caravello reflect on recent developments in AI regulation from the past month, walking through the White House's newly released policy framework and Senator Marsha Blackburn's proposed AI bill. They consider the nuances of each, including comparing where the two approaches align—and also diverge—on issues like preemption, copyright, and child protection.

For the sources referenced in thi...

In this episode, Katherine Forrest and Scott Caravello dive into new research exploring what happens when helpful AI agents end up helping the wrong people. Our hosts break down two recent papers and discuss what each may mean for the future: "Agents of Chaos," which examines security vulnerabilities and unexpected behaviors in AI agents under social pressure, and "H-Neurons," which presents groundbreaking findings on specific neur...

In this episode, Katherine Forrest and Scott Caravello examine two recent federal court decisions on whether AI-generated materials are protected by the attorney-client privilege and the work product doctrine. They break down those decisions, United States v. Heppner and Warner v. Gilbarco, explaining how and why the outcomes diverged, the different factual footings, and what these decisions may (or may not) mean for future dispute...

In this episode, Katherine Forrest and Scott Caravello continue their conversation on Moltbook—this time with a special guest. John Carlin, Chair of the firm's Cybersecurity & Data Protection and National Security & CFIUS practice groups, joins for a closer look at the cybersecurity risks of the agentic social network. In their wide-ranging discussion, the trio covers a host of concerns, from exposed credentials to hypothet...

In this episode, Katherine Forrest and Scott Caravello go inside the material world of digital minds. Our hosts explain how companies operate the massive data centers serving as the physical foundation for AI, break down the staggering energy demands behind them, and consider what powering the future might mean—by way of gigawatts and governance.

##

Learn More About Paul, Weiss’s Artificial Intelligence practice: https://www.paul...

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 ...

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/industrie...

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.paul...

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...

In this episode, Katherine Forrest and Scott Caravello kick off 2026 with predictions on humanoid robots, agentic AI and the legal flashpoints to watch. They break down the AI infrastructure boom and unpack the line between acqui-hire and acquisition. Our hosts also spotlight OpenAI’s Disney licensing for Sora and the rights issues behind AI-generated “behind-the-scenes” clips, foreshadowing a deeper conversation to come.

##

Lear...

In this episode, Katherine Forrest and Scott Caravello unpack how regulators are thinking about AI in the financial sector. Joined by Paul, Weiss colleagues Roberto Gonzalez and Sam Kleiner of the firm’s Economic Sanctions and Anti-Money Laundering (“AML”) practice group, they explore the Financial Stability Oversight Council’s initiatives on AI, the Office of the Comptroller of the Currency’s model risk management guidance, FinCEN...

In their first episode of the New Year, Katherine Forrest and Scott Caravello unpack OpenAI’s release of GPT-5.2, covering its performance on benchmarks like Humanity’s Last Exam and on OpenAI’s own GDPval—an evaluation designed to test the model's ability to match or surpass professionals' performance on real world tasks. Our hosts also examine the model’s sharp drop in hallucinations and break down OpenAI’s discussion of the mode...

In this episode, Katherine Forrest and Scott Caravello examine how deepfakes have evolved from static clips to adaptive, real-time impersonation driven by identity engines and behavioral replication. They explain why sensor-consistent forensics are being spoofed, the implications of the zero-trust evidence era for courts and companies, and how regulators and insurers are responding—plus concrete steps on provenance controls and inc...