AI won't be taking over infrastructure and security anytime soon. 🔒
Not because it isn't capable of generating plausible-looking output - it absolutely is.
But plausible-looking is exactly the problem.
In a standard dev cycle, a bad output costs you a failed PR. In infrastructure or security, it can cost you an outage, a breach, or worse. The blast radius of getting it wrong is in a completely different league. 💥
AI supervision is already a skill. In these domains, it's a specialisation.
The engineers who understand that distinction - who know where to let the agent run and where to keep a very tight leash - are going to be the most valuable people in the room. 🎯
Not because they're afraid of AI. Because they respect what's at stake.
Part of an ongoing series on what AI actually looks like on the ground. Read the full article: We Know. We’re Just Not Saying It.
More notes on engineering, AI and the systems around the work.
Back to all writing