AI's Bottleneck Isn't Models or Tools, It's Security

I love your blog. AI is trained to do things but it isn't trained to do things securely. Especially in software supply chain, openAI acknowledges this limitation and hence came up with an AI solution called aardvark. This is just the beginning. As we give more power to AI to take decisions on our behalf, this problem ust gets worse.

One of the key things to solve this problem would be AI guardrails for "stay clean" and automation for "get clean" scenarios