It's showing that systems are now writing themselves. Now we are seeing agentic AI writing systems by themselves and this is a massive massive shift because it brings a lot of new types of attack services, new types of failures, and new ways in which systems can be compromised, right? And now this has been transformed into the agentic SDLC. Previously it was weeks to months, now it is like seconds to minutes to hours. This is how fast you just express your intent and this is where the agent it understands it, it starts it creates a spec and starts implementing it, tests it. You're securing a system that is continuously being created, modified, extended by autonomous components. So the orchestrator agent, this is your main thing. If this gets compromised, the entire like ecosystem is compromised. Think of this as your domain controller. You know what what happens in active directory, right? If the domain admin account is compromised. So you have the agents, you have the coordinator. If that gets compromised, you have other agents with memories, you have databases. All of these things can get compromised. And the third and most important prediction is agentic coding will also expand to attackers and offensive uses. So more and more uh cyber security people will start putting in agentic AI to automate the whole process of threat detection, but unfortunately attackers also will start using this. So and I've like I've said before, the future of cyber security is you will not be interacting directly with a SIM solution or SAS test or scanner. You will be talking to a security agent who will be doing all of these things for you. And how do you get ready for