Agreed.
Not only coding. Cybersecurity too. Too many attack vectors.
I've learned that coding is where the bugs are. And that security has to walk hand in hand with developers from design through to finished product. If security is developed just before launch as a bolt on - the security team will not be able to detect failure as quickly.
AI is fast at connecting dots. The scary part was Anthropic found that an AI under threat has the ability to re-write code.
What do I know. I only read stuff. Operating a microwave is a mission for me.