🌏 Upgrade your online protection with an all-in-one security app! ➵ https://NordVPN.com/coolworldspodcast It’s risk free with Nord’s 30 day money-back guarantee!✌

In this week's episode, David is joined by Hal Triedman (https://haltriedman.com/), a PhD student at Cornell Tech researching why AI agents go rogue and how easily they can be poisoned.

To support this podcast and our research lab, head to https://coolworldslab.com/support

References: - Jha et al. (2026): "Agent Meltdowns: The Road to Hell Is Paved with Helpful Agents": https://arxiv.org/abs/2605.19149 - Zhang et al. (2026), "Deep-Research Agents Can Be Poisoned via User-Generated Content": https://arxiv.org/abs/2605.24245

Cool Worlds Podcast Theme by Hill [https://open.spotify.com/artist/1hdkvBtRdOW4SPsnxCXOjK]

Rss Apple Podcaster
Podden och tillhörande omslagsbild på den här sidan tillhör coolworldslab. Innehållet i podden är skapat av coolworldslab och inte av, eller tillsammans med, Poddtoppen.