Virtual machine

So I didn't see in the announcement any real AI infrastructure play, right? They're not running VMware Kubernetes on bare metal.

They're running it inside of VMs.

So I guess I'm a little...

I understand the marketing value for saying, hey, we have an AI factory or we're building an AI factory, but for... enterprises using your, you know, if you're a large enterprise, you're probably building AI clusters as AI clusters.

Go ahead.

Please.

Excuse my ignorance about this, but does anybody in the market have anything even resembling a a way in which the customer can manage their VMs and the GPUs.

I mean, are you seeing it coming out of Red Hat? Is there anything that's doing much of...

And he's got the config there, pseudo echo block in quick from, uh, that IP address to any, uh, and he's appending that to etcpf.conf. Now for the virtualization of it all.

FreeBSD has its own native VM hypervisor called Beehive, and QEMU is also available if needed.

I'll be sticking with Beehive for now.

Experimentation beyond that is a stretch, stretch goal.

I am, you know, making sure everything runs, the, the main stuff runs.

Unfortunately, yes, it does mean that the test load is, you know, goes up every time I add a platform.

Like, I even toyed with the idea of adding Illumos and OmniOS support, but then I was like, "Eh, unless there's actually somebody that's gonna use that, let's not create more of a headache." But I have VMs for everything that I'm actually loading it up in and running, and I have, you know, a, a battery of tests that I'm, I'm checking on each system to make sure that I didn't break the obvious.

But like with any complex system, there's always edge cases, and that's when I just need people who are using it to, like, if you find something that's broken, don't just grumble silently.

Another story that I thought was pretty interesting is that the new cyber models can escape VMware machines.

So you can't contain them with a virtual machine, which is what I've been using, unless you use a special virtual machine called Firecracker that is stripped down to present a smaller attack service.

And Firecracker, even GPT 5.6 cyber, cannot escape.

vulnerabilities

and finding zero

days in the VMware.

Because remember, VMware specifically is not doing the same security patches that it used to.

It's not VMware anymore.

Another story that I thought was pretty interesting is that the new cyber models can escape VMware machines.

So you can't contain them with a virtual machine, which is what I've been using, unless you use a special virtual machine called Firecracker that is stripped down to present a smaller attack service.

And Firecracker, even GPT 5.6 cyber, cannot escape.

They're changing vulnerabilities and

finding

zero days in the VMware.

Because remember, VMware specifically is not doing the same security patches that it used to.

It's not VMware anymore.

And, you know, like now we also have the fact that you can actually just have an agent do the maintenance for you.

So, you know, if you're going to be non-technical and run a relay, just have the agent sound the VM and do the maintenance work for you.

Mm-hmm.

And that's just going to get automated.

And that's kind of rough for a lot of things at the beginning.

Yeah.

But I think models like the data vending machine, like the DVMs, or rather context VM, which is what, you know, Jesus worked on, that protocol spec would help a lot in terms of funding,

right?

Last few companies like Rockwell have been selling their industrial data center, which was a virtual platform to run a lot of stuff on, but we still find, but it's not throughout the entire plant.

They might have VMs.

If you got 500 HMIs in your plant, you might have a third, two thirds of them might be VMs.

You still got standalone machines out there, right? At the same time, we're bringing in more applications specifically around cybersecurity, AI being another one.

It's like, okay, the plants don't necessarily have that infrastructure per se.

We're actually in the IO path.

So we are the storage that your application sees.

Whatever that hardware happens to be underneath of it, it doesn't make any difference, right? Software defined 101, right? So you've got the data on the storage, and then you've got the hypervisor, the compute layers that utilize that data, right? And today is in the form of VMs and or containers, right? Kubernetes, Docker, Podman, all of that support within the platform.

But before we do, Allison, what's the latest going on with the stuff that you're clotting?

Well, I am working on a big update to AVM, Accessible Virtual Machine.

I've more or less got USB pass-through working.

So soon you'll be able to use your USB devices like microphones and – And some drives and all the various USB things that we plug in every day on your accessible virtual machine.

I also have something that I've released called Access Term.

Access T-E-R-M.