
30 - AI Security with Jeffrey Ladish
AXRP - the AI X-risk Research Podcast
Security Concerns: Model Weights versus Source Code
Model weights are an attractive target for theft, but the source code is even more critical as it holds the key to creating more powerful models that may not align with safety measures. While securing model weights is challenging, protecting the source code, which contains less information, is much more difficult. Time is essential for aligning systems to ensure safety, and if source code is stolen, there is a risk of creating dangerously powerful models. Ultimately, the focus should be on safeguarding the source code to prevent misuse and potential catastrophic outcomes.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.