So if we, as you said, if we think of MCP just another tool that gives access to resources, to anything that can be behind such an MCP server, so is it not an MCP vulnerability, in air quotes, but something where we see that it's just the focus point, the crystallization point for security problems that... come from mcp or do not so what is is mcp a new kind of security problem or is it just a new focal point where they happen

well again matthias you are hitting kind of the nail on the head and you are scratching a huge itch i had we discussed it earlier as well like People love inventing new labels for old things because labels sell.

Of course, if you can say, hey, we have just launched an MCP security product.

You have to buy to protect yourself from all these fascinating new problems.

Sure, people would fall for it.

But if you are following... coping equals year-long kind of advice and say, stop looking at labels, look at functionalities and capabilities behind.

And yes, MCP is just a new standard, a new protocol, if you will.

But what would be something that a developer, an application owner, a system provider, a system owner would have to do? How does that look in practice?

Does that work?

Oh, this is absolutely a great idea, and this is absolutely the right kind of thinking, but you have to temper your expectations very strongly.

First of all, just kind of to give you, uh, some, uh, background, I was born in the country which doesn't exist anymore, the Soviet Union, so I experienced it firsthand as a child.

What is it, what does it mean to be technologically sovereign to an extreme? But we do have other countries having kind of the same struggle in the more recent years.

To put it plainly, it, it is simply impossible for multiple reasons.

First of all, it takes huge amounts of investments and decades of research, throwing a lot of resources at just getting as good as your competitors, and as soon as you stop, you are lagging behind instantly.

So you have to continue that investment cycle.

Um, but what should an organization that is listening right now to you and me, what, what do they need to do n- next Monday morning to actually embark on a journey towards sovereignty and resilience? Because, uh, the reasons to do that, they are still good ones, right?

Is there anything new around that, and is it as dramatic as the news outlets, the news made it? Starting with you, Alexei, what's your opinion on that?

Well, Matthias, I think you all kind of, uh, hit it straight where I had my doubts a- all the time since I've heard the story.

So first of all, what we definitely should avoid is to kind of presenting it as a Skynet scenario, malicious AI, Doomsday attack.

Uh, are we all doomed as the entire society? No, it's definitely not that kind of issue.

But second and more important, it was not even a single incident, right? It was not that OpenAI hacked Hugging Face or a model hacked Hugging Face.

Those were two separate incidents with very different failures, if you will, with very different blast radio, and they really should have been studied separately from the very beginning.

The OpenAI failed to contain a, quote-unquote, "dangerous AI model." Basically, they had a pet tiger, and the tiger escaped.

Alexey, I think you, you looked at the defender side as well, right?

And yeah,

this is true because, again, kind of identity on its own is, of course, critically important.

But it's not the identity on its own that decides what you can do or what you cannot do at this specific moment, at this specific context, and under a combination of very specific conditions.

This is where real-time access occurs.

decision happen and this is exactly where all these quote-unquote zero-touch platforms are focusing their developments now because again this is a pretty important thing so continuous authorization is probably like the basic foundation of it like if you only do authorization once at login time This is not zero trust.

Avoid those tools like a plague.

Continuous registration is the baseline and I can report that 75% of vendors we analyzed do that already.

as analysts, right? The problem, again, is kind of