Powered by RND
PodcastsTechnologyUnsupervised Learning

Unsupervised Learning

Daniel Miessler
Unsupervised Learning
Latest episode

Available Episodes

5 of 510
  • A Conversation with Bar-El Tayouri from Mend.io
    ➡ Get full visibility, risk insights, red teaming, and governance for your AI models, AI agents, RAGs, and more—so you can securely deploy AI powered applications with ul.live/mend In this episode, I speak with Bar-El Tayouri, Head of AI Security at Mend.io, about the rapidly evolving landscape of application and AI security—especially as multi-agent systems and fuzzy interfaces redefine the attack surface. We talk about: • Modern AppSec Meets AI Agents How traditional AppSec falls short when it comes to AI-era components like agents, MCP servers, system prompts, and model artifacts—and why security now depends on mapping, monitoring, and understanding this entire stack. • Threat Discovery, Simulation, and Mitigation How Mend’s AI security suite identifies unknown AI usage across an org, simulates dynamic attacks (like prompt injection via PDFs), and provides developers with precise, in-code guidance to reduce risk without slowing innovation. • Why We’re Rethinking Identity, Risk, and GovernanceWhy securing AI systems isn’t just about new threats—it’s about re-implementing old lessons: identity access, separation of duties, and system modeling. And why every CISO needs to integrate security into the dev workflow instead of relying on blunt-force blocking. Subscribe to the newsletter at:https://danielmiessler.com/subscribe Join the UL community at:https://danielmiessler.com/upgrade Follow on X:https://x.com/danielmiessler Follow on LinkedIn:https://www.linkedin.com/in/danielmiessler Chapters: 00:00 - From Game Hacking to AI Security: Barel’s Tech Journey03:51 - Why Application Security Is Still the Most Exciting Challenge04:39 - The Real AppSec Bottleneck: Prioritization, Not Detection06:25 - Explosive Growth of AI Components Inside Applications12:48 - Why MCP Servers Are a Massive Blind Spot in AI Security15:02 - Guardrails Aren’t Keeping Up With Agent Power16:15 - Why AI Security Is Maturing Faster Than Previous Tech Waves20:59 - Traditional AppSec Tools Can’t Handle AI Risk Detection26:01 - How Mend Maps, Discovers, and Simulates AI Threats34:02 - What Ideal Customers Ask For When Securing AI38:01 - Beyond Guardrails: Mend’s Guide Rails for In-Code Mitigation41:49 - Multi-Agent Systems Are the Next Security Nightmare45:47 - Final Advice for CISOs: Enable, Don’t Disable DevelopersBecome a Member: https://danielmiessler.com/upgradeSee omnystudio.com/listener for privacy information.
    --------  
    45:53
  • The 4 AAAAs of the AI ECOSYSTEM: Assistants, APIs, Agents, and Augmented Reality
    In this episode, I break down what I believe is the emerging structure of the AI-powered world we're all building—consciously or not. I call it the “Four A’s”: Assistants, APIs, Agents, and Augmented Reality. This framework helps make sense of recent developments and where it’s all headed. I talk about: 1. Digital Assistants That Understand and Optimize Your LifeYour DA (like “Kai”) will know your goals, preferences, health, schedule, and context—and proactively optimize your day, from filtering messages to planning meals or surfacing relevant information in real time. 2. APIs and the Real Internet of ThingsEverything becomes an API—from businesses to people to physical objects. Your assistant interacts with these APIs to act on your behalf, turning the world into a navigable ecosystem of services, tools, and resources. 3. Agents and AR Bringing It All TogetherAgents act autonomously to complete multi-step goals, and AR glasses will display their outputs contextually as you move through the world. These systems will collaborate, search, and act—quietly transforming how we live, work, and perceive reality. Subscribe to the newsletter at:https://danielmiessler.com/subscribe Join the UL community at:https://danielmiessler.com/upgrade Follow on X:https://x.com/danielmiessler Follow on LinkedIn:https://www.linkedin.com/in/danielmiessler See you in the next one! Chapters: 00:00 - The AI Ecosystem We’re Building Without Realizing It01:33 - Assistant: Your Most Powerful Digital Companion03:08 - APIs: How DAs Interact with the World07:54 - Agents: The Step Beyond Automation11:00 - Augmented Reality: The Interface Layer of the AI Ecosystem14:20 - Combining APIs, Agents, and UI for Real-Time Situational Awareness17:17 - Summary: A Unified Ecosystem Driven by the Four A’s23:36 - Industry Trends: How Companies Like OpenAI, Apple, and Meta Fit In25:11 - Final Thoughts on Timelines, Winners, and Interpreting AI NewsBecome a Member: https://danielmiessler.com/upgradeSee omnystudio.com/listener for privacy information.
    --------  
    27:04
  • Using the Smartest AI to Rate Other AI
    In this episode, I walk through a Fabric Pattern that assesses how well a given model does on a task relative to humans. This system uses your smartest AI model to evaluate the performance of other AIs—by scoring them across a range of tasks and comparing them to human intelligence levels. I talk about: 1. Using One AI to Evaluate AnotherThe core idea is simple: use your most capable model (like Claude 3 Opus or GPT-4) to judge the outputs of another model (like GPT-3.5 or Haiku) against a task and input. This gives you a way to benchmark quality without manual review. 2. A Human-Centric Grading SystemModels are scored on a human scale—from “uneducated” and “high school” up to “PhD” and “world-class human.” Stronger models consistently rate higher, while weaker ones rank lower—just as expected. 3. Custom Prompts That Push for Deeper EvaluationThe rating prompt includes instructions to emulate a 16,000+ dimensional scoring system, using expert-level heuristics and attention to nuance. The system also asks the evaluator to describe what would have been required to score higher, making this a meta-feedback loop for improving future performance. Note: This episode was recorded a few months ago, so the AI models mentioned may not be the latest—but the framework and methodology still work perfectly with current models. Subscribe to the newsletter at:https://danielmiessler.com/subscribe Join the UL community at:https://danielmiessler.com/upgrade Follow on X:https://x.com/danielmiessler Follow on LinkedIn:https://www.linkedin.com/in/danielmiessler See you in the next one!Become a Member: https://danielmiessler.com/upgradeSee omnystudio.com/listener for privacy information.
    --------  
    9:35
  • A Conversation with Patrick Duffy from Material Security
    ➡ Secure what your business is made of with Martial Security: https://material.security/ In this episode, I speak with Patrick Duffy from Material Security about modern approaches to email and cloud workspace security—especially how to prevent and contain attacks across platforms like Google Workspace and Microsoft 365. We talk about: • Proactive Security for Email and Cloud PlatformsHow Material goes beyond traditional detection by locking down high-risk documents and inboxes preemptively—using signals like time, access patterns, content sensitivity, and anomalous user behavior. • Real-World Threats and Lateral MovementWhat the team is seeing in the wild—from phishing and brute-force attacks to internal data oversharing—and how attackers are increasingly moving laterally through cloud ecosystems using a single set of compromised credentials. • Customizable, Context-Aware Response WorkflowsHow Material helps teams right-size their responses based on risk appetite, enabling fine-grained actions like MFA prompts, access revocation, or full session shutdowns—triggered by dynamic, multi-signal rule sets. Subscribe to the newsletter at:https://danielmiessler.com/subscribe Join the UL community at:https://danielmiessler.com/upgrade Follow on X:https://x.com/danielmiessler Follow on LinkedIn:https://www.linkedin.com/in/danielmiessler Chapters: 00:00 - Welcome & High-Level Overview of Material Security02:04 - Common Threats: Phishing and Lateral Movement in Cloud Office05:30 - Access Control in Collaborative Workspaces (2FA, Just-in-Time, Aging Content)08:43 - Connecting Signals: From Login to Exfiltration via Rule Automation12:25 - Real-World Scenario: Suspicious Login and Automated Response15:08 - Rules, Templates, and Customer Customization at Onboarding18:46 - Accidental Risk: Sensitive Document Sharing and Exposure21:04 - Security Misconfigurations and Internal Abuse Cases23:43 - Full Control Points: IP, Behavior, Classification, Sharing Patterns27:50 - Integrations, Notifications, and Real-Time Security Team Coordination31:13 - Lateral Movement: How Attacks Spread Across the Workspace34:25 - Use Cases Involving Google Gemini and AI Exposure Risks36:36 - Upcoming Features: Deeper Remediation and Contextual Integration39:30 - Closing Thoughts and Where to Learn MoreBecome a Member: https://danielmiessler.com/upgradeSee omnystudio.com/listener for privacy information.
    --------  
    26:47
  • AICAD: Artificial Intelligence Capabilities For Attack & Defense
    AI is changing cybersecurity at a fundamental level—but how do we decide what to build, and when? In this episode, I outline a structured way to think about AI for security: from foundational ideas to a future-proof system that can scale with emerging threats. • Rethinking Human Workflows as Intelligence PipelinesBy mapping tasks into visual workflows, we can pinpoint exactly where human intelligence is still required—and where AI agents are most likely to replace or enhance us. • Using AI to Understand and Manage Organizational StateI introduce the concept of AI state management: building systems that track your current and desired security posture in real time, and using AI to bridge the gap—automating insights, decisions, and even actions across your environment. • Building a Cyber Defense Program Inspired by Attacker PlaybooksInstead of waiting for threats, I propose a new framework based on attacker capabilities—what they wish they could do now and in the near future—and how to proactively prepare by building a continuously adapting AI-powered defense system. Subscribe to the newsletter at:https://danielmiessler.com/subscribe Join the UL community at:https://danielmiessler.com/upgrade Follow on X:https://x.com/danielmiessler Follow on LinkedIn:https://www.linkedin.com/in/danielmiessler Chapters:00:00 - Framing the Future: Two Key Questions on AI and Cybersecurity01:28 - Intelligence Pipelines: Visualizing Human Work as Replaceable Workflow06:10 - Theory of Constraints: How Attackers Are Bottlenecked by Human Labor10:42 - Defining Agents: What Makes AI Different From Traditional Automation12:08 - AI State Management: The Universal Use Case for Automated Intelligence16:53 - Real-World Demo: Unified Context AI for Security Program Management26:30 - Advanced Uses: Reassigning Projects, Updating KPIs, and Security Reports34:58 - Automating Security Questionnaires With AI Context Awareness38:43 - ACAD Framework: Predicting and Preparing for Future Attacker Capabilities47:40 - Defender Response: Building AI-Driven Red Teams and Internal UCCs52:25 - Final Answers: How Software and Security Change With AI AgentsBecome a Member: https://danielmiessler.com/upgradeSee omnystudio.com/listener for privacy information.
    --------  
    42:52

More Technology podcasts

About Unsupervised Learning

Unsupervised Learning is about ideas and trends in Cybersecurity, National Security, AI, Technology, and Culture—and how best to upgrade ourselves to be ready for what's coming.
Podcast website

Listen to Unsupervised Learning, Super Data Science: ML & AI Podcast with Jon Krohn and many other podcasts from around the world with the radio.net app

Get the free radio.net app

  • Stations and podcasts to bookmark
  • Stream via Wi-Fi or Bluetooth
  • Supports Carplay & Android Auto
  • Many other app features
Social
v7.17.1 | © 2007-2025 radio.de GmbH
Generated: 5/8/2025 - 1:05:25 PM