The Evolution of Open-Source Intelligence: A Conversation with Nico Dekens

In this episode of Integrity Insights, we are joined by Nico Dekens, a recognized authority in the world of open-source intelligence (OSINT). With over 30 years of experience in the field, Nico shares his insights into the evolving OSINT landscape, the tools and techniques he uses, and the ethical considerations of this critical field. Nico spent over two decades with the Dutch police, working on intelligence and human operations, before transitioning to the private sector where he now leads engineering efforts at Shadow Dragon.

We dive into how OSINT has changed over the years, with new technologies such as AI revolutionizing the process, but also the complexities of using these tools responsibly. Nico shares practical advice on how to start an OSINT investigation, how AI and machine learning are reshaping the future of intelligence, and how to stay up to date in a fast-evolving field. He also gives us a glimpse into his work on high-profile cases, where he uses his investigative skills to piece together puzzles from seemingly limited data.

Key Takeaways:

• The Evolution of OSINT: Nico provides an overview of how OSINT has progressed over the past three decades, with a focus on the increased integration of AI and machine learning in the investigative process.
• Practical OSINT Techniques: Nico explains his approach to conducting investigations, from starting with a name or photo and using facial recognition tools to building actionable intelligence.
• The Ethics of OSINT: Nico discusses the ethical boundaries of OSINT, focusing on the use of publicly available data and the challenges of navigating the dark web and leaks without crossing legal or ethical lines.
• AI and OSINT: While AI tools provide huge advantages in terms of processing large data sets, Nico emphasizes that human judgment remains critical for interpreting and validating the data.

Time Stamps:

• 00:00 – Intro & Background: Nico’s professional journey and his transition from the Dutch police to Shadow Dragon.
• 04:20 – OSINT Techniques: How Nico begins an investigation and the tools he uses, including facial recognition software.
• 10:00 – The Ethics of OSINT: Discussing the ethical challenges and limitations of OSINT, including using publicly available data.
• 15:30 – The Role of AI in OSINT: How AI and machine learning are changing the field and the importance of human oversight.
• 21:45 – The Future of OSINT: The rise of commercial OSINT, network-building in the community, and keeping up with the fast pace of technological change.

Relevant Articles:

• The Slow Collapse of Critical Thinking in AI – Nico’s recent article exploring AI’s limitations in OSINT.
• OSINT is Still a Thinking Game – Discussing the balance between technology and human input in OSINT investigations.

Listen to the Episode:

• Available on Spotify
  , Apple Podcasts
  , and Amazon Music

Don't forget to follow Integrity Insights for more expert discussions on integrity, financial crime, and investigative techniques!

Connect with Us:

• LinkedIn: https://www.linkedin.com/showcase/integrity-insights/?viewAsMember=true
• Berlin Risk Linkedin: https://www.linkedin.com/company/berlinrisk/?viewAsMember=true
• Website: https://berlinrisk.com/

Transcript

 All right, Nico, welcome on the podcast. Thank you. Thanks for having me. It's an honor. It's, well, the honor is mine. I I, i, I don't want to get too much into detail regarding your biography because there's really like a plenty of material that everyone can find, uh, online, but, uh, we need to say, we need to start somewhere.

So could you maybe just give us like a very brief version of your, you know, professional trajectory? Yep. Yeah, of course. Uh, so I worked, um, just under 25 years for the Dutch government doing osint in intelligence and human operations. Um, mostly online when it comes to open source intelligence. So it always had an online component.

Then I moved over to Bellingcat, uh, for a very short period of time. And now currently I'm the SUP of Engineering and Chief Innovator at a company called Shadow Dragon. And, and at Chato Dragon, is it like a, because I know this is a US-based company, but I guess you have operations all over the world.

Like what do, is it a full-time position or what, what, what is your Yeah, it's, it's a full-time position. So basically I work with the, with the CTO to um. To steer the development team and, and enable all of our users, uh, all over the world in, uh, giving them a, what we call a sauce os ocean solution. So basically a scalable os Ocean solution platform.

Mm-hmm. And you work remotely from Yep. From your, yeah, so I work from the Netherlands. Mm-hmm. That's, that's my home base. But, uh, typically I'm somewhere around the world, um, every month. Mm-hmm. Depends on where they want a month to half. Me, I just came back from Washington and Washington, Japan earlier, so I, I travel a lot.

Oh wow. And, and I, and I saw on your LinkedIn that you have there after you left the police in 20 or since 20, uh, 19, you have this company, Dutch Osint Guy Intelligence Services. Uh, what does that company do exactly? Uh. So it started as a consulting, uh, one man consulting company. So, uh, what I do typically, I, um, go in in companies and floor walk and see how they can implement open source intelligence, uh, train them, lecture them, um, strategical planning, um, but always with, let's say the OSI slash human intelligence component.

That's, that's my consultant part. So you, but nowadays with my very busy day job, it's mostly giving talks, keynotes, and presentations on from my company, Dutch guy. So you basically like develop tools. You, you give lectures, you, you teach, you do also to yourself. Like what do you, is there, what do you, what do you enjoy the most?

Uh, I still enjoy the most doing an investigation that's just old habits. I just like to dig. I just like to, to. Uh, to piece together puzzles. So, um, yeah, I just sometimes just read the news or find something in my timeline like, Hey, this might be an interesting use case. And most of the times all you have is a first name and last name or a location, and that's your starting point.

And, uh, if all goes well, you find a lot, and then either I, I never. Disclose those use cases publicly. I always share them with the according authorities if there's something that's really sensitive or that they need to know, because the fact that you can find things publicly does, for me at least, doesn't mean that you should always then, uh, share that information or at least the analysis publicly.

Mm-hmm. This is actually a perfect segue. What, what you just said, that you, uh, started investigation with just a name, because I, my, this was exactly my next, next question I was wondering. You know, because I would really like to get more practical since, you know, I'm talking to one of the best known Osint investigators.

I was wondering, you know, practically, how do you start like an osint investigation? If you say, if you have just, let's say a name, uh, like how would you start? And is it, is that, is the, has the change radically over the last, let's say, I dunno, 5, 6, 7, 8 years. Yeah, good question. Yeah. So there's to, to answer the, the change.

Yes, definitely. There's definitely a lot of change, especially since all the ais and machine learnings and LLMs of this world started popping in. Um, it has its advantages, but it also has its downsides because I think when you now do a typical search engine search, so think of Google, Bing, Orex, they first give you a AI tailored results, which for me very often are.

The less reliable results because they are summarized. So I like to do it the old school way. So let's get back, let's just say recently I read, uh, a news article that in Germany they arrested two Russian nationals that were living in Berlin that were, uh, arrested, um, on suspicion of spying for the Russian government.

So. In the news articles, there was just one profile picture shared. So I had a face and I had a first and last name that was just mentioned publicly in the news. So that's my starting point. And then, but then my goal is, is to come up with a research question. What do I really want to know? Because of course I can start looking up everything, but my goal was, can.

Proved. First of all, can I find this exact same person based upon that profile picture? Um, so to build at least confidence for myself that I'm dealing with the right person. And then my next question is, can I find information that connects them to, to Russia or Russian intelligence or let's say spying orientated?

Um. Connections businesses. So in this case, there were a lot of businesses connected to them. And when you looked at those businesses, you could see other people that were working in the company were clearly Russian National. Some were even sanctioned on cha known sanction list. So again, you pivot off those and yeah, by the end of the day, I could at least confirm what the news article stated.

Yes, this is, uh uh, someone that has been living in and around Berlin for a long period of time have set up businesses. All these businesses are. Untransparent and have always connections to Russia. And when most of those connections, uh, I can with high confidence label as people that have connections to Russian intelligence, RS, or at least the Russian government.

And then the next step for me always is where it's really getting interesting. Can I find unknowns that were not? Uh, identified in the news. So can I find new individuals that may help me, uh, reach out to Russia, ba uh, uh, uh, Russian, sorry, uh, German baker, ah, or German police, and say, Hey, did you know this is what I was able to find with my techniques and tools?

Did you know about this person or did you know about that information? So that's, that's one of the recent use case examples, but it literally starts with a name and the hard part always is. A name can be spelled in so many uh, ways. So if you look at my name, Nico Daikins, it could be just Nico Dash Daikins.

Nico Do Dakin. Nico underscore Daikins, last name, first name, uh, translated into Chinese lic Arabic because I don't know how someone or someone else put their name online and then of course, start guessing. Most common email providers. So if these were German and Russian related people, you could think of all the major Russian in internet service providers as well as ger, uh, German internet providers, and then combine all those names with emails well, and in the end, after a couple of hours, she would at least have a lot of information and.

That's not osint for me. That's just finding information. Then the next step is, is making sense of all the information that I've now collected. What is noise? What is actionable? What is new information? What allows me to pivot and dig deeper? So that's. That's basically how I approach it. That's a very short, condensed version of how I approach it.

Mm-hmm. A lot of, a lot of things come into my mind as you speak. Uh, one interesting, uh, thing, what you said about, uh, names being written in different ways, uh, I remember you, you wrote recently an article on this about Amsterdam, right? About the, about the, the, the Dutch capital and how it can be spelled in so many different names and how this is a challenge for us and investigators.

But I was also wondering. Uh, when you said, uh, in this particular case you have a name and you found a, a photo, uh, how do you work with, uh, like this type of data with, with photo? Do you, for instance, rely on, uh, some fo facial recognition? Uh, tools. Yeah. So I, I use facial recognition tools that are commercially available.

Um, so in the ocean world, there are a common, uh, there are a few ones that are out there that are pretty well known. So you have pim, you have lenzo.ai, and those are, let's say, commercially available for an individual. And of course there are also facial recognition tools that. Are mostly used by government, but I sadly don't have access to them.

Um, but yes, and of course just traditional reverse image searching in Google, uh, Google being yandex, Dr. Go all the search engines. And nowadays, um, AI is getting better with recognizing faces, uh, or at least. Telling you where that same image or a similar looking image might exist somewhere on the internet, but it's still not where I want it to be.

So mostly commercially available facial recognition tools. Mm-hmm. And are there any tools, you know, generally speaking, not just facial recognition, but in general that you relied the most, you know, these days? Like most recently. So specifically for facial recognition, you mean, or No, I mean in general.

General. Like for your, for your research. Well, of course the company that I work for, shadow Dragon, so we, we, we, I, I basically don't need anything else mm-hmm. Because we cover almost all sources that are out there. Um, but other than that, it's, I use a lot of, uh, breach data information or steal the log information.

Um. We can have a debate on if that's open source intelligence. Um, but again, I, I'm very strict on that. I will only use breachs and leak data if it's really publicly available. So I like to call it low hanging fruit breaches. Someone else could be, should be able to find them by doing. Just a search engine search.

Now the hard part is always, if you download eight gigabytes of data, how do I process and parse that and to find what I need to find? But that, that helps me pivot a lot. So that's what I use very often. Um, I'm still pretty old school when it comes to that. I don't use that many, let's say commercial ocean tools because yeah.

They have a price tag. And I'm Dutch. And the Dutch are known to be cheap. That's true. I had, I had a lot of, or a couple of Dutch roommates throughout my life. And that's, that's actually accurate. Uh, it's, it's an interesting, yeah, it's an interesting thing. But you mentioned the, uh, the bridge data and how you only rely on those if, if they're in fact publicly available.

Anyone can, anyone, can anyone can access them. 'cause this is a, this is a very. Controversial topic. I have, uh, I have recently read a, a Guardian investigation. I dunno if you saw that one. They did an investigation into Asad, uh, like Bahar Asad and his, um, his life in, in Moscow. And they use data from. And I think Bellingcat, Bellingcat does that as, as well.

They use data that they, I, I suppose buy on, on Telegram, like this flight, you know, about flight, like flight records. And they, uh, uh, guardian used that to track basically the movement of the children of Bashar Alad. And so if I understood you correctly, you would not go to this length, you would not like buy No, I'm, I have, I've, my ethical standards are slightly different.

Um. I will never pay for stolen data, period. Um, that's just something that I will never do. So I will only use data that's easy to find for anyone that just has an internet connection. Um, I know the information is out there, and it does not mean that I am, you know, let's say someone hires me that I sometimes know that the data might be there.

And I tell them, Hey, there is data out there. You need to talk to your legal department. We could collect it and do something with it, but we need to make sure that legally, um, we have the right backup to do this. Because yes, I do acknowledge there are certain channels or groups or places on the internet in the deep dark web where you can buy data.

Um, but again, um. I'm not a huge fan of that. Mm-hmm. Mm-hmm. Yeah, I mean, that's, that's always a question like where, you know, where you draw the line between legitimate osint and, uh, creepy behavior or violation of, of privacy. And I think it also depends on the use case. You can only imagine if I need to research someone that's stole a bike, uh, versus someone that's planning a terrorist attack.

I'm like, that's, that's a different, different game. So it really depends on how serious or how. Urgent. A use case is where you sometimes just need to maybe cross certain ethical boundaries, but only again, with the legal backup. You, you, I'm, I'm, I just want to. Play this by the rules at all costs. That's, that's my biggest goal when I do investigations or I, or when people hire me.

Yeah. Uh, another question I wanted to ask you about the, regarding your approach, because this is something I've been thinking of recently a lot is, uh, where do you kind of, and. Your investigation, you know, do, do, do you ever feel comfortable, uh, saying that, you know, you have, you have, you know, you have, you have investigated everything and there's no, there's nothing, there's no information, uh, on this individual or this, or this entity.

Uh, you have, you, you know, you have seen it all. Do you ever feel like comfortable in saying that? Because I feel like there's still, there's so much data these days that it's, it's just difficult to know like where to actually conclude the, the investigation. Yeah, I like that question a lot. And there's something that I, um, yeah, I teach classes, two, six day classes.

That's what I hammer on with my students all the time. Where do you stop? When do you stop? And I think that all comes down to if you're able to define an answerable research question, you should know where to stop. Because if I need to know, um, and always make this analogy, uh, where were you last Wednesday?

That's a very constraint. Research question. So I could probably spend weeks looking into your entire life and find things from three years ago and whatever, or even planned for the upcoming weeks, but that's not my research question. My research question is, where were you on that specific date and time?

So that's narrowing down. And with that, you probably know your first step would be, let's see where these guy's online. Can I find it? If I can find it, then I can narrow it down to that timeframe. If I cannot find it. The not finding an answer is still an answer. I, I've had many cases where I've had very, let's say, um, large criminals at at large that just had very good operation security and they had little to no online footprint.

But that by itself is also an an a deliverable, an answerable like, Hey, you asked me to fight this individual because you wanna, you wanna arrest him. But this individual is not online, so that could potentially mean that they take great pride and effort in hiding themselves online. So. I hope that answers your question, but for me, it all comes down and that's, that's what I'm seeing currently and that's what we started off this discussion with the landscape has shifted for me the past five years where people just want to grab everything about anyone, and I'm like, then it's just information.

I always want to, the, the in, in osint means intelligence, which means that you need to. Define a research question. Come up with a research plan. Define your sources, collect information process, and analyze the information, and then turn that into an answerable deliverable. So someone always comes to you with a question.

Find me. Anything about everything. About anything is something that I will simply say, can do it. I cannot download the entire internet for you. That's impossible. Yeah. Yeah. No, it makes sense. Um, and I mean, you already mentioned this is a huge topic. We cannot ignore artificial intelligence. I, um, what I wanted to ask you, you know, I have, I'm a bit younger than you.

I have been, uh, you know, doing this for, I don't know, 8, 9, 10 years. And, uh, for me, or like in general, you know, prof being in my professional life and for me, this is so far the biggest game changer that I have experienced. In terms of, you know, everything changing so quickly, is it, is it the same for you or was, I don't know, like the internet, you know, even bigger, um, in your early days?

Um, yeah, the, there's definitely, this is definitely, definitely the, for me in my entire career. So I've been doing this for over 30 years now. Um, so yes, I'm that old. Um, so that. Especially the AI is the most significant change because everybody jumps on the hype wagon. For me, um, osint or intelligence gathering has always equal change for me.

Uh, in the good old days, for example, platform like Facebook was very easy to find information. Then they shut down certain capabilities and basically you're back to the drawing board. I think what now really? Has become more useful is that I, I learned programming. I learned how to program Python and other things.

Nowadays with ai, people don't need to know how to code anymore. They can just ask an ai, Hey, I have an idea for a tool. And as long as you can articulate your prompt well enough, the AI will bring you the code. Now, the hard part is can you trust the code because I also know that AI. That does a very bad job at coding something or may add in lines of code that communicates to a super secret server somewhere else that makes you highly uncomfortable.

But it does mean that where I, 10 years ago, if I had, let's say, 10 million pieces of data, I had to come up with how to process and analyze that data myself, nowadays, I can most likely throw most of that data in a local running ai. Um. On my local machine. So I don't prefer to use online ai. Um, I wanna be in control.

Um, so I download models, I have very special graphics cards and computers. I wanna be in control. So I, I can now easily say, Hey, here's 10 million lines. I only want to see this piece of information carve out that, create a dashboard for this. So that's something that has become definitely easier, but still.

It always requires the human in the loop to validate that information. So that's what I think is the biggest change. Mm-hmm. And maybe you notice as well, in certain cases it makes life easier, but in other cases I spent literally more time prompting than if I go back to in traditional way of doing it, I could have probably found, found my entry in half an hour instead of spending one and a half hour finding the right prompt to get the right answer.

Yeah. Yeah. That definitely happens. I mean, I, I would really invite everyone to, uh, read your. Articles on the, on the subject, on the Dutch OSI Guy website. You have written most recently an article titled, uh. The slow collapse of, no, this one is all a bit older, right? The slow collapse of critical thinking in ai.

In ai. Uh, yeah. The most recent one is Osint is still a thinking game. Uh, so anyone can read those, but, uh, could you maybe just kind of briefly summarize your position? So I, I think people are forgetting about the trade craft. Uh, it all comes down to, um, critical thinking. And for me it's always about the five W one H questions, which means the who, what, when, where, why, with what, with whom, how.

Um. And AI is not capable of reading between the lines and interpreting what you want to see. It's great at showing what it can see in the data, but it doesn't understand the context of the data and it doesn't understand the context in a research use case, even if you give it a prompt with guardrails and telling it, Hey, this is my goal, this is what I've collected.

Even then it's not good at, so this is why I'm a little bit worried. On the newer generation, but also my generation, let's say the old school ocean folks, um, that we rely a little bit too much on algorithms and AI where we forget that we have our own mind and that we should always challenge the outcome.

Of any tool. So it could be ai, but it could also be a commercial tool. You should always, let's say, try to validate that, try to challenge the outcome. How, how certain can we be, can we find alternative routes or way ways in, into this information to, to create more certainty in my report, for example. So that's, that's definitely what I see.

And I think it also, the risk is that. The younger generation for me feels sometimes like that. They are button operators. You give them a tool, you give them a button, and it's almost like they're looking for a tool that says, fight me the criminal or fight me the suspect. I. I would never want to use a tool that does that, an algorithm that is great that it can do, let's say the first triage, but there should always be a human in the loop in every step to validate any outcome out of any algorithm like my own CTO.

Jason always said AI and ML is just fancy. That's it. Nothing, nothing more. And, and math can break and math will make mistakes. So you need to make sure that you double check, triple check everything. Mm-hmm. Mm-hmm. But I mean, still omnipresent. And, uh, you know, I was wondering if the, how do you see the future of tools like Shadow Dragon?

Can you, you know, uh, foresee a, a world where they, those tools are absorbed in a. In, you know, some, uh, agentic AI systems? Yeah, to some extent. So currently we are literally on our roadmap. We have, um, more analytical layers, what I like to call 'em and assistance because we do see a new breed of investigators.

They, they don't have, they. Barely trained on open source intelligence, which also worries me. I think companies should train their people before they give them powerful tools in, let's say the fundamentals of intelligence scattering and open source intelligence. But we're trying, we're trying, we're going to build in tools that will assist the user in ensuring that they get the most out of the tool, um, that it points out things they forgot or should at least pay attention to and we need to deliver.

Um. Quicker insight because. People are also becoming more and more impatient. Uh, people don't wanna wait. Even waiting 30 seconds for a result nowadays, for a lot of people is too long, which means that we need to speed up things and then tell them basically in one picture it's, well just like we are doing a podcast now.

It, it was easy to do a free hour podcast three years ago. Nowadays, the attention span of most people is this short and the same counts for using software. So the software needs to be catered more to. Impatient, but also unexperienced users. And that's literally what, what we're building. Mm-hmm. We're trying to build more, what I like to call ease of use in there.

And I was also wondering in terms of, you know, using AI for the different, um, parts of the intelligence process like, uh, collection extraction analysis, do you, do you, do you feel more comfortable using ai uh, for. You know, uh, one aspect rather than another. Yeah. I think AI works best with large volumes of data.

So if I need to, um, oh, maybe a good, good use cases. Recently in the Netherlands we had, um, the. Uh, no, my mind. The NATO Summit, which basically meant all the world leaders came to the country, and you can only imagine that the Dutch government had a reason to understand if there were violent protests or other forms of violence being announced.

That means that you cannot look at individuals. Because that's from a legal perspective and no, no go. But you can look at messages from social media or the internet where people announce that they're going to riot or set things on fire or going to attack people because that means that maybe the riot police should be positioned in a certain location knowing that a large group of people will be coming there.

Um, and this is where I think, um, monitoring comes in. On top with analytical layers. So it should be able to extract locations, it should be able to extract, uh, threat indicators, so keywords. And that's hard because most people who announce violence will not literally say, Hey, I'm going to shoot you, or, Hey, I'm going to set this on fire.

No, they're probably going to say, we need to get her here and here. It's going to get spicy. Something like that, or let's make a fist together, or let's make sure that we let them know that we are serious. So being able to interpret those nuances in language. This is where I think, uh, well, I know because I've experimented this a lot, ai, but typically machine learning is pretty good.

In, in understanding a narrative of a large volume of group. So I'm not interested in finding that one person that's just angry with the world. There are too many of there. Mm-hmm. I just want to know how can we keep a certain area as safe as possible during a certain event? So there would not be just the collection, but also analysis.

Yes. So it's parsing the data. So first step is collect the data based on keywords and indicators. Um. And then analyze that information for meaning. And for me that's triage, uh, triage because the next step is the dashboard will present you, Hey, these are five announcements. These are people that reply to it.

This is the, the overall narrative. And then I want to humanly validate. The accuracy of it and then maybe pivot into them and say, Hey, is this a group that's well organized? Are they financed? Do they have backup? Um, well, we've seen that also a lot during the, the Hamas Palestine protests all over the world where we could literally see that some of these protests were clearly financed and influenced by others, where the pure goal was violence.

I'm all for democracy and I'm all for free speech, but when it comes to people. Deliberately joining protests with their, their only goal is to, to, to be violent in any way, shape or form. I think that's where the tipping point comes in. And this is where I think open source intelligence can, can, and is super useful if you use that with the right collection methods and then analytical layers on top of that.

Mm-hmm. And you also mentioned the human, the human element. Do you, do you have the impression that. In this day and age with again, the growing amount of content data that the human is becoming more important. Yes, certainly. Um, well we've seen most, uh, a lot of social media platforms closing down or locking down.

So the only way to really get access is to have accounts on those platforms. Uh, but also sometimes when you look at, for example, telegram groups where a lot of, uh, nation state actors are, are active or maybe hacking groups are active sometimes just asking a general question, Hey, what are the plans for the weekend?

Or, Hey, what's up next, guys? Um. That's a form of human or at least elicitation that sometimes could give you just that little lead or pivot point that can take you into the next step of your investigation. Is that Osint? Um, I'm not sure. Uh, but I do think that is the evolution of osint, especially in the phase that we are in the internet, is simply becoming more locked down and decentralized.

So for certain ETS use cases. You need to interact, or at least you need to ask questions, and that's something that you can do openly or coly, depending on the use case. Uh, I suppose you just need to have, make sure you have a really good, uh, operation security when you're engaged. Yes. And that's, that's, that's the hardest part because you do want to interact, but you do not want to let others know that you are you with a specific interest and, and, uh, operation security on the internet, there's no such thing as 100%, um, a waterproof opsec.

Um. Because your computer will have to communicate with the internet, so no matter what you're sending out information to the internet, but you can certainly try to lock down things as to your best of your ability. So obfuscate your IP address, obfuscate your computer, fingerprint your username, your everything, your time zones, your language settings.

There's so much that you can do when it comes to hiding yourself and your mission from others online. Mm-hmm. Big, big question. Where do you see the, this, this whole field, because it's still, I don't know if it's still, but it was always kind of relatively obscure field, kind of, uh, emanating from intelligent services to slowly creeping into the corporate world, but it's still relatively, I guess, niche, but it's becoming, I feel like it's becoming a bit bigger.

Like where do you see the future of, uh, OSINT and its future applications? Good question. I think since last year for me it really felt like ocean has become mainstream. Um, so even my neighbor now knows. And my mom knows what Osint is, or at least what publicly available information is and what that means to find information.

I also see that there are a lot of more companies popping up like Shadow Dragon. Um, that's something because 10 years ago there were basically free companies. Us and two others. Now there's dozens. Um, and you also see acquisitions of companies where large companies buy osint companies like Shadow Dragon, or recently there was, there were some others, uh, from Australia bought.

So there's definitely a market interest as well where Osint has become a. I think more like a commercial commodity, if that's the right way to say it. There's definitely market interest and just like you said, um, it used to be a government only thing, but the past three, four years, the finance world, um, due diligence, um, um, marketing in general, there are so many, it's no longer.

Limited to government. Everybody or most companies nowadays have seen the light when it comes to open source channel, which makes me very happy because I've been advocating for osint for quite some while now. So it really makes me happy that it's, that it's finally resonating with a broad audience. Yeah, no, I mean, you are the OG of the, the Osint, the osint world, which leads me to my next and, uh, let's make it the last question because you, you mentioned earlier that you.

Been doing this for, you know, what is 30 years, uh, which is a long time. And as we discussed this, this, this field is just an, an enormously quickly changing. I really wonder how do you, you know, make sure to stay up to date. Yeah, good question. So this is why I, um, network a lot. So, uh, on my LinkedIn, on my ex, or still like to call it Twitter, blue sky.

So basically I have accounts on all the socials, but I attend all, um, conferences, workshops, read blogs. So I spend a lot of time catching up and typically my, my Fri my Friday is what I like to call my. Nerd Day, which stands for Never Ending Research and Development. Um, so that's where I just try and keep on top of the game.

Uh, but I must admit it has become harder because just like we just said, 10 years ago, there were only, let's say, a small group of people, but now we have thousands of people talking about Osint. So for me now, it's more about filtering the actual valuable new information that I want to know about and learn about.

But for me, OSINT is. It's always, always equals learning. So I learn something new every day. It could be a GitHub repository, it could be a block by someone else, could be a podcast like yours. That's what I encourage people to do. Reach out, join Discord communities, join slack groups, join other OS and groups.

There are many out there just to build a network and get to know people. Hmm. That's a, that's a great, great advice. Well, thank you so much, Nico, for your time. I, this has been real, an actual privilege talking to you. I really appreciate it. You took the time and uh, yeah, I wish you all the best in orial in all your roles.

Thank you. Thank you very much and thank you for having me.