AI & Enterprise
Palantir says 5 things matter more than models in AI security
Palantir’s product security team shared its experience adopting agentic AI for security work on a Medium blog. The team has used AI agents for about a year in tasks such as code review, vulnerability discovery and penetration testing, and is offering the know-how to customers under the name Security Forge. Palantir summarised five key points, including the need for a harness and organisational context, and said fixing vulnerabilities is now the bigger bottleneck.
