New Episode of Security Now: Security Now 1029

In this episode of Security Now, Steve Gibson and Leo Laporte tackle several crucial topics related to cybersecurity and emerging technologies. The episode begins with a discussion on an Apple research paper that explores whether new large reasoning models truly think. Steve Gibson shares his opinion, stating that these models do not actually think but are better at language processing. One of the highlights of the episode is the discussion on how Meta and Yandex use native applications to listen on local ports to track users across different websites. This tracking method bypasses all existing privacy protections, including browser settings and consent policies. Steve Gibson explains in detail how this works and why it is so difficult to detect and prevent. The practical implications of this revelation are immense, as it shows how determined big tech companies are to track users, even at the cost of violating their privacy. The episode also covers the new initiative by the European Union to launch its own DNS service. Although this service is slow for users outside the EU, it is designed to offer a secure and privacy-respecting alternative to American and other DNS services. Steve Gibson discusses the different flavors of the EU's DNS service and their usefulness for governments, telecommunications operators, and home users. Another important topic is the discovery that Russia might be intercepting Telegram messages. Steve Gibson explains how the Russian FSB has managed to access communications from users with certain Ukrainian Telegram channels, which could expose these users to treason charges. This revelation raises questions about the security of Telegram and how users can protect their communications. The episode concludes with a discussion on an Apple research paper that examines the capabilities and limitations of large reasoning models. Steve Gibson explains that while these models can solve complex problems, they fail to develop problem-solving capabilities that are generalizable beyond certain complexity thresholds. This discovery has important implications for the future of AI and how we understand and use these technologies. In summary, this episode of Security Now provides an in-depth look at the latest trends and developments in cybersecurity and technology. It highlights the invasive tracking methods used by big tech companies, the EU's initiatives to offer secure DNS services, and the limitations of large reasoning models. To learn more, you can listen to the full episode at https://twit.tv/posts/transcripts/security-now-1029-transcript