CI/CD

It's also possible.

I've seen people doing that a lot on CICD, for instance, because in the end of the day, most of those coding agents, they're just JavaScript programs that could be quickly installed in the GitHub Action Runner, for instance.

And then if you have your private roles, you can consume through the role, then you can do a code review.

We actually have a blog post that I will... link to the show notes like one of the consultants in Firefox self-wrote it where how you can define the role and that will be like so basically you can set a cup uh through the i am on uh bedrock spending which was like not the obvious thing to do like it's not very obvious how to do it but it's actually possible and then you can say like well start declining requests on the i am level when we like spend three thousand

Every change comes with a tax.

The debt has reached its terminal stage when it touches architecture, tests, CICD, docs, and processes.

One change triggers a dozen others.

So, the signals are clear.

And I think we just we can't be as loose as we have been.

But we're going to build everything, which is a great answer for a CICD from a CICD perspective to say, every time there's a change, we're going to build for 20 products.

We're going to build for five different chips.

Is that necessary? You know, that's not as much fun, but can you do that a little bit smarter where maybe you build for one representative tool chain and that covers half of what you need.

I wonder, I guess especially when you're pushing code into production on a network level.

To give you guys a bit of background, I was just on a call earlier and we were talking about Megaport acquired a bare metal company, right? And one of the use cases that they're seeing a lot of is they have AI agents that are continuously testing for CICD, right? They'll go, they will make code improvements, they will deploy their own compute instance on our bare metal farm by themselves, make sure it works, and then push the change into prod.

So if we're looking at that from an infrastructure perspective or from a network perspective, can you take something like a forward networks where it is, what's the term for what they do? It's like- Digital twin? Yeah, a digital twin, right? Where you have a digital twin of your environment You have an AI agent or an instance where you can go and take a snippet of code from AI, deploy it in your digital twin to make sure nothing breaks, and the change is actually doing what you want it to before you actually apply it to production.

And that's where I think, yes, I think there's some...

And so we build for the most users.

If you wanna get from point A to point B, whether that's, you know, CICD systems or, you know, mailing systems, whatever it might be, we'll build the platform that allows you to kind of decouple your thinking.

You can just think about the product, the customer.

You don't have that cognitive overhead of maintaining everything else down to the nuts and bolts.

I wonder, I guess especially when you're pushing code into production on a network level.

To give you guys a bit of background, I was just on a call earlier and we were talking about Megaport acquired a bare metal company, right? And one of the use cases that they're seeing a lot of is they have AI agents that are continuously testing for CICD, right? They'll go, they will make code improvements, they will deploy their own compute instance on our bare metal farm by themselves, make sure it works, and then push the change into prod.

So if we're looking at that from an infrastructure perspective or from a network perspective, can you take something like a forward networks where it is, what's the term for what they do? It's like- Digital twin? Yeah, a digital twin, right? Where you have a digital twin of your environment You have an AI agent or an instance where you can go and take a snippet of code from AI, deploy it in your digital twin to make sure nothing breaks, and the change is actually doing what you want it to before you actually apply it to production.

And that's where I think, yes, I think there's some...

I wonder, I guess especially when you're pushing code into production on a network level.

To give you guys a bit of background, I was just on a call earlier and we were talking about Megaport acquired a bare metal company, right? And one of the use cases that they're seeing a lot of is they have AI agents that are continuously testing for CICD, right? They'll go, they will make code improvements, they will deploy their own compute instance on our bare metal farm by themselves, make sure it works, and then push the change into prod.

So if we're looking at that from an infrastructure perspective or from a network perspective, can you take something like a forward networks where it is, what's the term for what they do? It's like- Digital twin? Yeah, a digital twin, right? Where you have a digital twin of your environment You have an AI agent or an instance where you can go and take a snippet of code from AI, deploy it in your digital twin to make sure nothing breaks, and the change is actually doing what you want it to before you actually apply it to production.

And that's where I think, yes, I think there's some...