Грозит ли нам ИИ-зависимость?
Summary
This episode features insights from the exclusive Cybercamp conference, with guests Sergey Gordechi, Vyacheslav Kassimov, Amar Ganiyev, and Alexey Yudin. They discuss how AI is reshaping cybersecurity, including the rise of autonomous SOCs, LLM firewalls, and blockchain security. The panelists also explore the shift toward "AI-native" businesses, the obsolescence of junior roles, and the critical need for new security paradigms to manage AI-driven risks.
Topics discussed
Introduction: Cybercamp conference and LLM hacking experiments
The paradigm of power, global AI investments, and control
Russia's position in the AI game and corporate automation
Cybersecurity budgets, pen-testing vs APT, and developer adaptation
AI as a mechanism for capital growth and future governance
Presentation on AI in SOC and controlling intelligence
Internal LLMs, security tools, and the firewall problem
LLM usage in text analysis and the need for local models
Amar Ganiyv: AI in banking, fraud detection, and white hat behavior
Model dependence, Mythos/Fable, and the future of Red Teams
AI Native companies and the speed of AI implementation
Security awareness, agent risks, and limited rights
New attack vectors, local models, and data protection
Ensemble models for detecting APT attacks and graph interfaces
Model training speed, fine-tuning, and reliance on foreign tech
Red teaming GitLab: privilege escalation and secret discovery
Hardening results, business requirements, and infrastructure stability
Conclusion: Cyber-sovereignty, telemetry, and final thoughts
Listen ad-free on Castria