Summary

  • Open source large language models (LLMs) are being used by startups and cybersecurity leaders to develop adaptive, cost-effective defences against cyberattacks, which are evolving faster than humans can respond to them.
  • At last week’s RSAC 2025 conference, Cisco, Meta, and ProjectDiscovery announced new open source LLMs and community-driven attack surface innovation, which define the future of open source in cybersecurity.
  • One key takeaway from the conference was the shift towards extending and strengthening infrastructure at scale.
  • Cisco’s Foundation AI group focuses on delivering domain-specific AI infrastructure tailored to cybersecurity applications, most of the work it does should be open to encourage innovation.
  • Cisco’s foundation-sec-8B LLM offers a new era of open source AI and was purpose-built to tackle complex cybersecurity tasks.
  • Meta expanded its AI Defenders suite to boost security across generative AI infrastructure, including a multimodal classifier and a real-time security framework for AI agent decision-making processes.
  • ProjectDiscovery’s Nuclei is a customizable open-source vulnerability scanner driven by a global community that rapidly identifies vulnerabilities.

By Louis Columbus

Original Article