Cisco's AI Bet: Smaller Models for Cybersecurity

Cisco is betting on smaller, specialized AI models for cybersecurity. The company believes general-purpose AI models aren't built for security tasks, leading to issues with customization, data privacy, and high costs. Yaron Singer, Vice President of AI and Security at Foundation AI, Cisco, explains that smaller models trained specifically on cybersecurity data offer better control, privacy, and cost efficiency. These models can be fine-tuned for precise tasks and deployed without sending sensitive information to outside providers. What's interesting is how these smaller models can dramatically improve analyst capabilities, turning month-long tasks into just 10 minutes of work. This approach is seen as groundbreaking for security operations. The bottom line is that tailored AI can significantly enhance cybersecurity efforts.