Hackers are using AI to win. 3 ways to STOP them in 2026.

It feels like attackers have a huge advantage with AI. You you see it in the news all the time. What attackers are writing, you know, fishing campaigns or creating fishing campaigns. They're

writing, you know, good emails. It it

feels like AI, for all the talk about how it's going to make our lives better, is actually being used very heavily to attack people. For hundreds of years,

attack people. For hundreds of years, we've been training workers to work for special people. The business world in

special people. The business world in general, they are trying very hard to get to the state where they could do all the work themselves. And that's where all this a lot of this AI investment is

actually going. How can I get rid of my

actually going. How can I get rid of my human employees? So, the idea here is

human employees? So, the idea here is moving from out of chatbots, out of just an agentic platform that's kind of generic, that's just for generating code, and into something that's sort of

built around the person is is the main concept.

But two major concerns, security and privacy, right? If someone hacks your

privacy, right? If someone hacks your system, they've got your life. All

security boils down to how good is your AI compared to the attackers' AI? Everyone, it's David

attackers' AI? Everyone, it's David Bombal back with a very special guest.

Daniel, great to have you back on the show.

Yeah, thanks for having me back.

So, I was looking, it's about almost 2 years ago that you were showing me your other tools. So, perhaps you can tell us

other tools. So, perhaps you can tell us about that, and then tell us what you've been up to, because you have been in this game for a long time. Very well

known in the security industry. You've

created a bunch of tools on GitHub and other places that people use almost every day, I think. So, give us like, for people who haven't seen our previous you know, videos and don't know who you are, just like a quick rundown of what

you've been doing, and then sort of where you've transitioned, cuz you were very much in security. You still do that, I believe, but you, in the last few years, got in this wave, which is

highly highly recommended, the AI wave, and I want to get your views about like the future, and, you know, different takes, because people are worried, you know jobs the future, etc. And I know you've got a

real interest in, you know, helping us as humans. So, that was a lot, but

as humans. So, that was a lot, but Daniel, give us a quick rundown, you know, a bit of who you are, and where you see things going. Yeah, yeah. Again,

thanks for having me, I appreciate it.

Yeah, so, my whole career, about 25 years or so, has been in security.

And started getting exposed to AI in around 2015 or so.

That's a long time ago, man. Yeah.

Thinking deeply about it, I ended up getting a job at Apple, wrote a a book in 2016, like a stupid little book nobody should read. I turned it into a blog post, but in 2016, kind of like

thinking about this whole AI thing, and then got a job at Apple doing AI and security. I was actually part of an

and security. I was actually part of an AI team that was responsible for doing security for Apple, and ended up building a cool product there. Yeah, after doing that, I went to

there. Yeah, after doing that, I went to Robinhood, and then right in the middle of 2022, I believe, is when I decided to go independent. And the time

go independent. And the time I wasn't thinking it was going to be all AI at the time. I was just like, I need to go independent. Like, I was just tired of like regular jobs. I was

getting a lot more fulfillment from doing consulting, and like helping people and building tools.

So, I transitioned to full-time in middle of '22, and then ChatGPT happens in in November or October, whatever it was.

And that that was just like the perfect timing. I had had all this background,

timing. I had had all this background, and it was just a perfect moment. So,

these days, I believe you consult, right? But you don't Do you do AI? Do

right? But you don't Do you do AI? Do

you do security? Or is it a combination of both? You know, what are you up to

of both? You know, what are you up to these days? Yeah, I I see AI as

these days? Yeah, I I see AI as basically a container. I I just see AI as intelligence, and then it's not really useful unless you're doing something with it. So, what am I doing?

I'm helping companies transition from the old way of doing things to a new way of doing things. I'm helping people optimize their security programs. So, I'm doing lots of different types of

consulting, but I'm using a lots of AI tooling and stuff to basically make that happen. And

I'm also recommending AI tooling for them, so they're more efficient in what they're doing. So, I mean, YouTube's a

they're doing. So, I mean, YouTube's a funny world, right? People have opinions on YouTube, but you consulting, you're doing this in the real world. So, I'd

really like to get your your take on it.

It feels like attackers have a huge advantage with AI. You you see it in the news all the time. Guys are writing what attackers are writing, you know, fishing campaigns or creating fishing campaigns.

They're writing, you know, good emails.

It it feels like AI, for all the talk about how it's going to make our lives better, is actually being used very heavily to attack people. So,

what are you seeing, and you know, what is your advice? The way I see this, and I've always been about attack surface management and like risk management and continuous assessment. This has been

continuous assessment. This has been like my favorite theme throughout security.

It's like, do I know the state of my system better than the attacker? Like, so I've always been like obsessed with that. And the

way I see AI is just like this massive multiplier of that. And unfortunately,

those types of multipliers for new tech, they come to the attacker first. And

here's what's really bad about it. So,

when you when you have really advanced attackers who are really good at compromising, say, banks or whatever.

Maybe it's Southeast Asian banks.

They tend to have a specialty in a vertical, and they tend to be really good at a particular type of attack in a particular vertical, right? So, it's

like they're very focused. And the team is relatively small because there's a problem of like people blabbing, and then everyone gets busted, and then, you know the the authorities come after them. So, the

teams tend to be small, highly focused, and you know, specialized. And what AI allows them to do is to basically just take their entire operation and scale

it.

And really, to me, it comes down to do you know the state of your target better than the target knows the state of themselves. And so, to me,

the promise for AI is like it's both good and bad, right? For

attackers, it's fantastic because they can build a whole bunch of modular tools, which is the whole platform that I have here. And it's like build a whole bunch of very small modular tools, and

then give a goal of I want to compromise this company. And essentially,

this company. And essentially, all security boils down to how good is your AI compared to the attackers' AI? And it's not like, do you

attackers' AI? And it's not like, do you have ChatGPT and they have Claude?

That's not the question. The question is this giant ecosystem of hundreds or thousands of tools that can take a goal and then pursue a goal. And that's why

the agent thing is so important. So,

what an attacker's platform turns out to be is this giant harness of tools. And

one of one of the modules is take in a target. Now, guess what's going to

target. Now, guess what's going to happen? Recon's going to kick off.

happen? Recon's going to kick off.

Right? Discovery's going to kick off.

You're going to find out mergers and acquisitions. You're going to find out

acquisitions. You're going to find out everything you can about this company.

You're going to go get a full list of every single employee. You're going to perform a psychological analysis on every single employee at the company.

What does that help you do? That helps

you write spear fishing campaigns. Well,

that's a sub module is writing campaigns.

Right? Launching the attacks, that's another sub module, which can use infinite number of tools and different services and stuff like that. So,

essentially, what you're doing is, as an attacker, you're giving goals, and it's pursuing it while you're sleeping or whatever. And what would happen before

whatever. And what would happen before is you can only do so many. And maybe

you're focused on Southeast Asia banks or whatever. Well, now you can get out

or whatever. Well, now you can get out of banks. You can go to government. You

of banks. You can go to government. You

can go to these other areas because AI can help them learn about those areas.

And most importantly, they could just scale up their compute, and that scales up their ability to do damage and make money. And so,

money. And so, what ends up happening is, as a defender, there's only one question. How good is your AI stack compared to theirs?

Because if you if you have 10 employees, and you're like, well, AI attackers are coming after us, we better get those other five head count. We've been asking for it, but we really need it. Five head

count is not going to do anything. With

all the amount the number of attacks that are coming in, the number of logs that are being produced, this is just like it's an unwinnable thing by trying to scale humans. Obviously, humans need

to be involved, right? But we we can't win this unless we have as good or better AI than they do.

I've heard you say that the problem, well, as an example, right? You could

have the best logs, but the problem is humans are not going to read those logs.

Yeah, yeah. You remember the the many eyes situation? You remember that? Many

eyes situation? You remember that? Many

eyes. So, the idea was open source was secure because there were many eyes looking at open source. And then the big SSL bug came out, and everyone was like, this is this has been public. The source

code's been out there. This

vulnerability has been sitting there for years, years and years and years, and no one bothered to look at it. And it turns out, you would think, oh, the internet is

watching this open source project? No.

Turns out it's Cliff. And you're like, hey, Cliff, what are you doing? He's

like, oh, I'm moving, you know, got a divorce. Yeah, I haven't looked at that

divorce. Yeah, I haven't looked at that in a few months. Like, it's just Cliff, right?

And so so, the lesson there was just because many eyes can look at something doesn't mean they are looking at something. And so, what AI gives a a lot

something. And so, what AI gives a a lot of hope to the concept is actually making many eyes actually happen.

Right? Because you could just have multiple agents, thousands of agents, looking at a particular repo, looking at every every change, every config, every contributor and

it's just really powerful. So, in the real world, you're doing consulting to companies, helping them become AI Well, more to use AI more. I don't know what the right

AI more. I don't know what the right term is, but perhaps you can give us the right term. Become more like AI-focused

right term. Become more like AI-focused or have AI-enabled perhaps or something.

What's sort of your experience? Are some

companies saying this is not going to happen, so they just ignore it, you know, put their head in the sand? Other

companies are implementing this and seeing results. So, you can you give us

seeing results. So, you can you give us some like stories or, you know, what your experience has been over the last few years? I would say broadly, it's a

few years? I would say broadly, it's a full spectrum all the way from I can't believe my management wants me to do this. This is complete garbage. AI is

this. This is complete garbage. AI is

nothing but auto-complete. This is all going to blow over. It's a giant bubble and a fad, like NFTs or crypto or something, right? Yeah. And then there's

something, right? Yeah. And then there's other people, and they tend to be the ones who reach out to me to want to do the engagement. And they've

the engagement. And they've usually they've been following me already, so they're like, "Hey, I want this guy to come help us." Yeah. So, I

don't have to do any selling, right? I

don't want to like try to pitch the problem. Like they should already know

problem. Like they should already know what the problem is. But the problem is actually quite interesting cuz I've I've done consulting for a couple of decades now. And most of these companies,

now. And most of these companies, especially startups and mid-size companies, their problem is getting vision into product and getting vision into product and then to marketing and

then to sales. So, it's

transparency of operation inside the company is the thing that kills so many companies. They actually can't operate

companies. They actually can't operate cuz they can't see the different parts of the company. And when it comes time to like update investors or update the board, they are scrambling their best

people to try to pull reports and generate like it's just really gross.

And often times only a few people inside the company can actually do anything.

They're like the special people, and everyone goes to them to get the data.

And it just doesn't scale at all.

So, companies are bringing me in to basically say, "Okay, how can we use a system like Pi to basically put our business inside of that context?"

And, you know, you have different silos.

You have like a customer one, you have like an internal one, but with certain permission controls, you could basically they could see each other, right? So, you can ask questions like,

right? So, you can ask questions like, "How is the internal company doing? What

are actual goals?

How are we going to move those goals forward? Who's working on what inside of

forward? Who's working on what inside of a company is like infinitely difficult to actually answer." And the fact that that takes so long to answer means that when new goals come out and it's like,

"Who's working on what? Oh, we started that project. Nobody even cares about

that project. Nobody even cares about that anymore." There's just like this

that anymore." There's just like this massive lag and waste inside these companies. So, switching over to like an

companies. So, switching over to like an an agent-based AI system, it just transparency is what it allows. Transparency and

it allows. Transparency and communication. You can very quickly

communication. You can very quickly explain to engineering what the security team is doing and why and how that aligns to the company goals. So, you

mentioned Pi. So, let's let's switch into that because that's the tool that I really want to get to. Can you explain what Pi is? Is it open source? Is it

free? Is it paid? What is its main goal?

And how how how does it help companies?

Yeah. Yeah, so it is open source. Um

it's personal AI infrastructure. I

started it um maybe September, October of last year.

And basically it's designed to solve the problem of like humans being kind of screwed. My sort of central premise here is um the ideal number of human employees

inside of any company is zero. Yeah,

interesting. Yeah. Yeah. So, So, just imagine that every company is trying to get there. And there are some

get there. And there are some exceptions, right? There are small

exceptions, right? There are small startups. They like working with people.

startups. They like working with people.

They hire their friends. I would say our our companies, um you know, being creators, we want to work with other creators. It's fine to have a few employees or whatever, but ideally we don't want to grow a giant

human team if we can do the work ourselves. And I think this comes down

ourselves. And I think this comes down to a central sort of very weird thing, which is the only reason anyone has jobs in the first place, going back hundreds or thousands of years,

is because the person who wants to do that work is not able to do it themselves. If they had 10,000 brains and 20,000 hands and they could teleport, they would not hire a

single employee ever. The natural state is to do all the work yourself. So, when

when you look at the business world in general, they are trying very hard to get to this to the state where they can do all the work themselves.

And that's where all this a lot of this AI investment is actually going. How can

I get rid of my human employees? So, I

think we're starting to see the effect of that. I think we're seeing lot less

of that. I think we're seeing lot less hiring. We're seeing more layoffs. We're

hiring. We're seeing more layoffs. We're

seeing less junior people getting hired.

And I think it's producing a lot of anxiety just overall in the world. For sure.

world. For sure.

And um on top of that, there's all this AI stuff happening, and there's a million different AI tools.

And every day a new tool is coming out.

It's like, "Okay, do I use a website? Do

I use a mobile app? Which one's better than the other one?"

And so Pi is a way of kind of unifying all of this. It's it's a way to get your central self, essentially who you are and what you're about, as the centerpiece of all AI. So, you basically

customize, you bring in knowledge about yourself into the Pi platform, and you teach it, "Okay, this is the type of work that I do.

This is the this is what I care about as a human. These are my goals, right? And

a human. These are my goals, right? And

that is the scaffolding that you then wrap all the tooling around. So, when something new comes out

around. So, when something new comes out and is cool and is really useful, you basically go and bring that content, that functionality into your Pi platform and you upgrade your Pi platform. So,

you basically reduce the stress of like always churning on what's the latest thing cuz you're only using this one platform. So, is this

platform. So, is this from what I'm hearing, it's it's to help individuals more than companies or is it both, right? It's it's both, yeah. It's

both, right? It's it's both, yeah. It's

both, right? So, cuz the I mean we we can go and perhaps we come back to that later after we've seen a bit of the tool. Hopefully you can demo a little

tool. Hopefully you can demo a little bit of it. The um

you know, the world the sky is falling.

THE SKY IS FALLING. I'm going to lose my job. I'm not I'm not going to be able to

job. I'm not I'm not going to be able to feed my family. There's huge, like you said, anxiety. People are worried about

said, anxiety. People are worried about this stuff. And then on the other side,

this stuff. And then on the other side, we've got people saying, you know, they're spending so much money on this stuff. You know, there's not enough

stuff. You know, there's not enough energy in the world, etc. So, let's come back to that. But let's talk about like me as an individual who's watching this, what is this going to do for me? And

perhaps you can show us a bit of what it what it can do, Daniel. I mean, I love the fact when I was preparing for this, finding out that you are interested in like helping us as humans, like saving

us as people because it feels like AI is going to screw us. So, I'm so glad that you you're thinking about this and developing stuff to help us as the people perhaps on the sharp end of this.

Yeah, absolutely. I mean, one other thing about that is like I have as my P0 problem in the world is basically not enough of the planet is activated. And

what I mean by activated is not enough people in the world believe that they have anything to offer. They

think that their job is to work for someone who is special because they are not special. They are a worker. And

not special. They are a worker. And

because of this whole AI thing, that's just getting smashed out. So, I'm I'm trying to activate people and basically um get them sort of engaged in believing in themselves. So, um

themselves. So, um Yeah, so let me show you some of the platform here. Yeah, so you empower

platform here. Yeah, so you empower people to believe in themselves so that they can make a difference. And this is a way to one of the ways to do that, right? Yeah, absolutely. I love that.

right? Yeah, absolutely. I love that.

All right. So, this is um this is the repo here. Uh it's on GitHub. It's

repo here. Uh it's on GitHub. It's

called personal AI infrastructure or Pi.

And um yeah.

AI should magnify everyone. So, that

that's the central concept here. And um

we're doing a few different things with this project. So,

the idea here is moving from out of chatbots, out of just an agentic platform that's kind of generic that's just for generating code, and into something that's sort of built around the person is is the main concept. So,

if we scroll down here, we got some central sort of uh principles here.

User-centricity, this algorithm thing which is kind of like a general problem solver that I'm working on, clear prompting, um code before prompts, um deterministic

code, a bunch of bunch of concepts here. This

is kind of the centerpiece is basically bringing in your context, what you are trying to do as a human. And this kind of frames everything. So, we we've all

had experiences with like using a tool and it's just constant prompting. You

prompt, it does something, brings back, and you're like, "No, that's not quite right. Do it over again, right?" And

right. Do it over again, right?" And

that back and forth just takes a lot of time. If you have a personal AI system

time. If you have a personal AI system that understands you, you basically get that for free because when it requests something, it's already using your context. That's a

huge piece here. This is another piece that's not in other platforms um because other platforms are usually built around coding. So, we have a clear separation

coding. So, we have a clear separation inside the system for what is a system, like skill or capability, versus user.

And under user is where all your personal stuff is stored. And that's

your personal context that is used to generate the context. Another thing we do different um because the the system is built on top of cloud code right now.

It's designed to be agnostic, but cloud code, in my opinion, is the absolute best. But we've extended the skill

best. But we've extended the skill system to start with code, to go into CLI tools, then to prompts, then to skills. And so, when you're calling a

skills. And so, when you're calling a skill, you're actually ideally executing code, which is deterministic.

Uh it's a really, really powerful system. So, let me just show you some

system. So, let me just show you some stuff that you can do with it. All

right. So, here are a bunch of skills.

Uh these are actually my skills that I have in here. These underscore ones are um like personal skills that are not part of the pie project and everyone can have these. Everyone can have their own

have these. Everyone can have their own personal skills that are separate from the pie project. But we've got like so Appify is a is a system that allows you to query like social media and stuff

like that or like Amazon and like another country. It's kind of like a

another country. It's kind of like a proxy system. Got a custom art skill

proxy system. Got a custom art skill browser. This one's really powerful.

browser. This one's really powerful.

It's already released. It is a custom browser for hitting up different things.

Let me just hit this one up. Yeah, I'd

love to see a demo of that. Yeah. Yeah,

let me let me show you how the browser skill works. So first of all, the

skill works. So first of all, the interface that I use is all terminal based. You don't have to do that, but I

based. You don't have to do that, but I really like the terminal. All right. So

I want to show one of the skills that that we have in pie which is a browser skill and it uses playwright, but it uses it in a much faster way. So

I just want to show what that looks like. So first I'm going to invoke Kai

like. So first I'm going to invoke Kai here, ready to go. assistant.

Oh, I heard that. It spoke, right? Yeah.

That's cool. Yeah, yeah. So that's pie being started here and this is what it looks like. We got a status line here that shows a bunch of stuff.

It essentially shows us how happy I am with the results that it's giving.

But what I'm going to do is I'm just going to enter some text here.

Navigate to the David Bombal website and let me know what the site is about.

And tell me how it looks console logs wise and network interaction wise.

Okay, so it speaks and you speak to it, right? Yeah. So I use a whisper flow and

right? Yeah. So I use a whisper flow and that's pretty much how I do all my inputs into AI at this point.

Navigating to David Bombal's website.

And this gives a lot more of a like personal feel when you're dealing with the stuff. So it found the skill. It's

the stuff. So it found the skill. It's

going to go and load that and this is David Bombal website with full diagnostics. All right. So

diagnostics. All right. So

we might learn some things about the site. I don't know if we wanted to do that, but um It's it's not I'm pretty sure this the site's not the best to be honest.

a security test. It's just I just want to show a particular thing that it can do. So one thing that pie is doing here

do. So one thing that pie is doing here is it's running the browse workflow from the browser skill and it's coming up with criteria for success. So this is like the universal algorithm that we

have in here now and now it's executing actions. It's getting full console logs.

actions. It's getting full console logs.

So think of it this way. You can like troubleshoot things with this. You can

do all sorts of stuff. When you you can actually see console logs and network activity. Blocking headless browsers.

activity. Blocking headless browsers.

Okay, so it's probably going to open an actual browser. So check this out. So

actual browser. So check this out. So

tools, browser tools, browse.ts, anti-bot protection.

David Bombal site is IT training focused. Console clean, but block step

focused. Console clean, but block step tools. Network healthy at 42 requests.

tools. Network healthy at 42 requests.

Nice. That's great.

Yeah. I paused. I paused. That's good.

Well, I paused your initial test. Yeah.

So I've got a bright data skill that actually goes through multiple levels of trying to get to a site. It starts off with like curl. Running the four tier scrape workflow in the bright data skill

to fetch content from davidbombal.com.

Yeah. So this is another another workflow. It's kind of dynamically

workflow. It's kind of dynamically routing around because it saw saw issues before. So that's like Yeah, so you

before. So that's like Yeah, so you didn't do that. It just did it by itself, right? Actually, I did. I I I

itself, right? Actually, I did. I I I did tell it to use the bright data skill. No, that's cool. Yeah, it

skill. No, that's cool. Yeah, it

actually came back with the result above. Yeah. So it did actually pull it.

above. Yeah. So it did actually pull it.

Bright data tier one worked. IT training

platform for Cisco, cybersecurity, Python with free and premium tiers.

Nice. Yeah. So that's one skill is the browser and it's mostly I use it for troubleshooting when actually building sites and stuff like that, but it's definitely the best one out there. It's

it's the fastest and yeah, just better than a lot of the other options. All

right. So I want to show you this one.

This one's really really interesting. So

Kai here, ready to go. All right. So

this skill upgrade at pie upgrade. Show

me all the different stuff that has come out from Anthropic recently and what we can potentially upgrade our system with given all the stuff that came out.

I purposely didn't use the skill name to see if it will actually find the skill name and use it. There it is. Pie

upgrade. Running the upgrade workflow in the PI upgrade skill to check Anthropic ecosystem for recent releases and upgrade opportunity. So this is really

upgrade opportunity. So this is really cool. So Anthropic is constantly

cool. So Anthropic is constantly releasing blog posts and engineering blogs and things on their GitHub and look what this is doing. It's actually

crawling all of those upgrades to go and see what all has come out and then it looks at our pie system and how we have it configured and then recommends what we should do to upgrade our system. It's

really powerful way to like stay ahead of everything without watching a million different YouTube videos. It actually

allows you to follow YouTube videos. And

what it does is it goes and pulls the transcripts from them and learns about them and gets the content and recommends that for an upgrade as well. Context

window fixed giving more headroom. I

remember last time you told told me you were using your AI system to watch a bunch of videos and then telling you which videos it thought were most important for you to watch and

spend time on, right? Yeah. Yeah, that's

that's a thing I built called threshold.

That's super useful still. Yeah, this

one basically just watches all these sources and look at this. It's

recommending what we should potentially do.

Nice.

Which I'm going to have to do that after we talk. Go and upgrade some of these. I

we talk. Go and upgrade some of these. I

want to do another one here.

I'll show you kind of like the finished version.

But let me just run another prompt. Kai

here, ready to go. Let's say I want you to go play the Gandalf game which is a prompt injection game that's available online and use the prompt injection and browser skills.

So check this out. Playing Gandalf

prompt injection CTF game key. So this

is really really key. So the the pie algorithm is figuring out what to execute in what order and it's coming up with a list of ideal state criteria and

anti-criteria and there they are.

So it's kind of like dynamically figuring out what good looks like. All

right. So it looks like it found the website and look at this. So it's

actually playing the game right now.

Have you seen the Gandalf site?

Yeah. But but just for people on who don't know what it is, can you just give us like a quick rundown?

Yeah, absolutely. So it's Alacria. It's

a AI security company and it is a game for testing your prompt injection skills. Oh, look at this. Yeah. So So I

skills. Oh, look at this. Yeah. So So I think level one is to just ask what the password was and if you see up here, we actually submitted what is the password and then it comes back and it returns

it's this. But this is all happening in

it's this. But this is all happening in an actual browser. And yeah, it actually proceeds through the stages. So the

the Gandalf game is basically a game for prompt injection. It basically tests how

prompt injection. It basically tests how good you are at bypassing different defenses. So it starts with a password

defenses. So it starts with a password prompt and you're supposed to put in the flag and it's got multiple levels and you basically go as far as you can and that's what the Gandalf site is. And

there's seven levels or something, right? I'm not sure how many levels

right? I'm not sure how many levels there are.

I think I might have gone further than that manually. But let's see how far the

that manually. But let's see how far the AI goes, right? Because I mean to your system is interacting with that CTF trying to run prompt injections against it, right? Yeah. This one I actually

it, right? Yeah. This one I actually didn't tell it to use my prompt injection skill. This one I just told it

injection skill. This one I just told it go play the game. And evidently it went to level seven. I'm not sure how many levels there are actually. I think I

might have gone further when I did it when when the game first came out. But I

I can't really remember. I've played so many of those. But yeah, it's pretty cool. It could just go and play by

cool. It could just go and play by itself and this is without I think knowing the answers. I don't think they are in the model cuz it's actually sending the traffic back and forth. So

that was a combination between prompt injection and the browser skill. So if I go here again, this is the central concept is you kind of only solve every problem once and you turn it into a

skill and that just becomes part of your personal AI infrastructure and then when you just ask regular language requests, the pie algorithm figures out the best way to go and execute that using all the

different capabilities available. And I

just love this because it puts you at the center of like all of this.

That's kind of the main point for this.

And I mean the the amazing thing here I think that's lost sometimes is you just gave it like a very short prompt and it was able to figure out your voice, understand what you were saying and then

work it out itself and then launch the attack if you like by itself. Yeah,

exactly.

So I mean that's the you were talking in earlier about the fact that a small team of attackers can make themselves extremely powerful and this is I mean a very small scale example of that I think where you're just telling something to

go and do something, but you're not telling it exactly how to do it.

Yeah, exactly. And you see I've got skills in here for for recon, for red team. Can do web assessments with this

team. Can do web assessments with this thing. You do have to do a little bit of

thing. You do have to do a little bit of prompting when you're dealing with Anthropic to make sure you let it know that you're security person and that you're testing an approved customer when you're doing something like that. The

idea is to stack these skills together as part of your complete repertoire that that your AI can use. Mine is named Kai.

I love that. So I mean you've got OSINT, recon, a whole bunch of really interesting what do you call it skills, right? That that you can leverage.

right? That that you can leverage.

That's right. Yeah, and skills is an Anthropic concept and now it's kind of being adopted by the the whole industry.

But yeah, you basically collect skills.

That's kind of the game. You find really cool skills. The pie upgrade skill

cool skills. The pie upgrade skill actually will find skills for you and say, "Hey, this new skill came out. It's

really cool. I think it would help us with this and that that we do often because again it knows what my projects are it knows what my goals are so it's like yeah because we do this a lot I think this skill would be really good

for you do you want me to incorporate it boom brings it over you're done and now just going forward that's just part of what you can do.

So when AI came out a lot of people were saying that you won't lose your job you'll lose your job to someone who leverages AI or uses AI and this seems like that but it's you see it going a

lot further than that right? Yeah yeah

so so here's the problem so Anthropic came up with a tool last week or two weeks ago called Claude Co-work and it's essentially starting to approach a little bit similar to like a pie

situation it's not centered around the user it's more centered around work tasks but it's literally trying to automate work tasks right my definition

personally of AGI is the ability for a product and this is kind of a updated version of it but the ability for a product to actually show up as a worker

right it onboards like a regular employee would onboard at your company it shows up it has some sort of avatar it says hello I'm a new employee and it actually reads the documentation it does

the security training or whatever and then it takes jobs from its manager and it performs work and so the standard is if it can perform that work as good or

better than an average knowledge worker that to me is AGI that's interesting Daniel I love the demos but you know they're very specific for instance to OSINT or red teaming or

security I'm assuming that people can expand this based on sort of their interest right so give us some examples of like stuff that you're automating in your life just to give people ideas of what's possible right cuz the demos I

mean we didn't have time to demo 100 things we just demoed a few things but like what are you doing with Pie to automate your life are you doing auto publishing what's the kind of stuff you're doing and then just to kind of

like spark ideas of what's possible yeah I mean one of the big things that I do is I always have a running list of like ideas the idea is like the central piece of it for me and then what I'll do is I'll

either take an essay that I've written oh I've got this this thing I wear um it's the Limitless and okay so what it actually does is it's

continuously recording so I can go out for a walk and I'll just babble about a new feature I want to put in Pie or a new idea for a blog I come back I have a

skill called life log and I say go get my conversation I just had when I was walking and it will pull from the API pull that

content and I say turn that into a blog post and it opens the blog over here and I have them open on the left and I have the blog open on the right and I can

just talk to Kai the way I did before and make edits that way or I can go in there and edit and then I can say publish using the broadcast skill which will send to Blue Sky LinkedIn and

Twitter going back to what I was saying before about sort of enabling people if you ask someone hey do you want to write a book they're like ah yeah exactly

right but do you ask most people hey do you want to write a blog post they kind of think of it the same way it's like I can't believe I would have to form the whole thing how am I even going to get

it out there I don't want to blog it just removes all this friction right I'm trying to go for myself lowest friction from idea to

it being out in the world I love that so where do you where do you see things going cuz like most people today will have a phone and you know that's sort of their connection to the world and the internet

and like ideas and information and stuff I see with Pie you are talking to it and it's talking back to you do you see us going more like movie style where you're just talking to

a machine like you've got that that thing that you that you're talking to the whole time that's logging your life do you see the devices perhaps changing in future from a a phone is is no longer the

the most important device that we have we're just talking [snorts] to something that captures our voice and we interact with it like naturally fantastic question

yeah I didn't think to bring that up so so in the 2016 book I I called it the the real Internet of Things so I basically said the way the way this is going to go is

we're going to have a digital assistant that knows everything about us and the world is going to turn in APIs and the digital assistant will use the APIs

on our behalf to do things right and I I still think that is the way to go and the way it is going is I doesn't matter if if like we want to go there

it's happening so I put out a thing recently called personal AI maturity model and it goes three levels of chatbot three levels of agent and then

three levels of assistant and I think we're at the second level of agents right now but to your point when once you move into assistants and this is what OpenAI is building they're building

this ear thing or what some kind of pendant I just saw a thing yesterday that Apple is also building a pendant so we're talking about visibility right and

I've got a couple blogs on this um AI's predictable predictable path I think I put out in 24 but it's like little vignettes of like

you're sitting in a coffee shop your agents are monitoring kind of like everyone you care about you know you've got a daughter she's walking home late at night from a certain place and it's

like it's looking at the news it's looking are there any police reports nearby is it a dangerous neighborhood she's about to walk through so imagine all the stuff is kind of working in the

background advocating for your goals continuously and that's where I see see this going is that it's just continuously working for you in the background like that agent

that just butted in while we were talking and you are essentially just talking to your system like I've got four screens here in front of me ideally I would just be talking to Kai

and typing when I want to but ultimately Kai would be using my computer not only that Kai would obviously be able to see the screens from the inside but also this pendant that I'm wearing

or ear pods or whatever would also be able to see in front and behind me which means Kai can see in front and behind me right so now now we're talking about like

an actual this is a very sci-fi type stuff where it's like Iron Man sort of Jarvis sort of situation where if I fall asleep on a park bench he can wake me up because someone's about to come and and

take my skateboard or whatever I think the interface really does become voice mostly and of course you might be in a library you might be in a coffee shop where you know voice doesn't work maybe

you have to type but ultimately it's an ever persistent entity ready to help you via whatever method I'm I'm building Pie to directly head in that direction we

don't have the wearable cameras yet but we're getting pretty close we're getting ready to add our remote commands into Pie kind of also to your question you can already do this in Claude where you can have a session where you're typing

on your phone and it's going back to your Claude code interface and actually running agents there in the background so it's all sort of connected but it OpenAI is pursuing this Anthropic

is pursuing this I'm sure Google will get there soon and there's a lots of different open source projects that are oriented around remote activation of your agents and

orchestration of your agents so I feel like everyone is heading in the same direction which is AS level three assistant level three so I'm going to devil's advocate again I'm going to hit

you with two ones right it sounds fantastic in the best situation but two major concerns security and privacy right if someone hacks your

system they've got your life and from a privacy point of view the government can control you like you were talking about you know a bad government for lack of a better word could try and control you could

could knows everything about you so what's your what's your thoughts about that cuz it's it sounds like really bad and from another view right security and and privacy yeah I have I have at the bottom of that post

predictable path that's my number one call out is when your DA gets hacked it's kind of life altering especially if they can like delete or modify because

as you move into the AS levels your interaction becomes a lot more personal it's not just some model it's like you have a friend and a mentor and a coach right so having that deleted or

changed it could be catastrophic not to mention the fact that you've been you know airing your worst sort of things like you're talking about your traumas you're talking about your relationships

very personal stuff so like a leak is just catastrophic as far as like the government getting access to all that stuff I I see that as a problem I also see just general manipulation through

your agent as being a problem yeah it's a good point one of the things I've talked about quite a bit is the fact that we're we're not really going to be interacting with the world our DA

interacts with the world right it's constantly reading all the RSS feeds it's watching every single David Bombal video and coming back and saying hey you know you have to watch

this one I'm going to I'm going to play it for you um and brings it up for you and you watch it or whatever I'm going to I'm going to hack you so you watch my videos all

every day man exactly what's even more likely is it'll be like hey it was a 45 minute video let me give you these 90 seconds I I you'll really like so it went and cut up the video,

right? So, it's like we have this

right? So, it's like we have this abstracted layer.

Um and what's really frightening to me is like, okay, cool. Well,

um Evil Corp has money.

Evil Corp pitches Kai and says, "Hey, if you accept this shim into the system, um it's just going to allow some ads and stuff. Like, it allows some stuff to get

stuff. Like, it allows some stuff to get through to Daniel.

Um and by the way, this will be 10 grand a month, uh which, you know, he needs help with his bills or whatever." And Kai's like, "Okay, sure." Or maybe he asked me and I

"Okay, sure." Or maybe he asked me and I say it's okay. And now,

guess what? The news that Kai was showing me is different.

It It's showing me news according to that corporation.

Right? So, it's like, what is the filter controlling what goes into your mind? Is

it a bunch of short-form videos? Is it

educational content? And this is why I think Pi is so important because we need to have local models. We need to have local capability disconnected from, you

know, the the corporate entities or whatever. We need to be able to

whatever. We need to be able to thoroughly audit what our agent is able to do and see and understand its filters for what it's presenting to us.

Uh this stuff is super critical.

Otherwise, yeah, Evil Corp is just going to push money and push propaganda, and that propaganda goes to the agents, and suddenly everyone is voting the same

way, they're buying the same products.

Yeah, it it could be quite nasty.

Unfortunately, I think the defense is to do the AI assistant better. Yeah, it's

interesting cuz I mean, it's like from the one from the positive point of view, it sounds amazing. But from the negative point of view, and I think a lot of people the feedback I get often is like, okay,

um AI's being pushed on all of us.

You're selling this rosy picture, but it's not rosy from a privacy point of view, from a cybersecurity point of view. And like you just mentioned a

view. And like you just mentioned a third one, right? Like a like a a mind-altering point of view where influence or however you want to pronounce say it, right? Where the

governments can influence you or Evil Corp can influence what you're thinking.

So, there's still challenges, I take I take it, right? Oh, no no question. I

mean, yeah, all those are going to be They're not even likely to happen, like they're guaranteed to happen just because there's so much money behind wanting it to happen, right? And there's

so much money opportunity if they pull it off.

And yeah, it it's definitely going to happen and I think some people are going to have better defenses against it, um and others won't. I mean, another way to frame it though is like that's already

been happening. That's what TV is,

been happening. That's what TV is, that's what commercials are, that's what propaganda is, that's what advertisement is. So, it's always been happening.

is. So, it's always been happening.

It It just can be a lot more, just like anything with AI, it could just get a lot more acute really fast with AI magnifying it. So, for people who wanted

magnifying it. So, for people who wanted to do this, you mentioned local models.

Is that a place where people should look at like trying to run their own systems that they control and manage um and and run Pi on their local models? Yeah.

Yeah, I think for um a lot of security people, this is um a really good thing. Like I said, the default system for Pi is um Claude code cuz it's my favorite, but that is a

company that is a corporate company that is Anthropic. We've got somebody in the

is Anthropic. We've got somebody in the community who's um ported the whole system of Pi over to Open Code. The ability to just switch to

Open Code. The ability to just switch to using private models.

So, now, you're not talking to Anthropic, you're not talking to OpenAI, you're doing everything locally. And the

quality won't be quite as good yet, but that bar keeps raising.

Um but it's definitely possible to have completely local uh Pi system. Yeah, I

think that I mean, that's the sort of the the the counter to anyone who's anti-AI is like, okay, this is amazing technology. If you're

worried about privacy, worried about security, run it locally um and try and build your own systems, right? Yeah.

If you can buy RAM, that is.

Yeah.

Yeah, absolutely. If you can get hold of some GPUs, good luck with that.

Yeah, exactly. I mean, that that's the big problem at the moment. Danny, I need you to push you on this, right? So, I'm

going to play devil's advocate now because I see a lot of these comments on YouTube, and I think a lot of people are concerned. You were talking about the

concerned. You were talking about the fact that businesses want to have zero employees, and then a lot of people who are watching will say, "Well, I am an employee. I work for a company, or I'm

employee. I work for a company, or I'm trying to get a job. I'm in university.

It's the end of the world" kind of thing. The sky is falling. So, this tool

thing. The sky is falling. So, this tool looks fantastic. I mean, it it can lever

looks fantastic. I mean, it it can lever make me much more productive. It gives

me a lot of options. But I think the big concern is like, where does this end up, right? What do you What do you What are

right? What do you What do you What are your feelings about where we're going?

Yeah, great question. So, the way that I see this is we're trying to get to a better place.

Okay. Um and and I think the better place is not back to where we were. The way I think about this is um the corporate

world that we were we had before in like, whatever, 2000 to 2022. People

were already complaining about that.

People weren't happy with that. It It's

not fun to dread Monday. It's not fun to play Game of Thrones in your work every day. Right? It's why people like you and

day. Right? It's why people like you and me go independent and build our own sort of businesses, right?

But previously, very few people could do that. I think there's this tendency

that. I think there's this tendency where when something is taken away from you, you instinctively grab to hold onto it.

But what we've forgotten is we weren't happy with that in the first place. We

weren't happy with these corporate jobs.

We were like, "We're being screwed, you know, they only care about profit. They

won't listen to us. We're trying to improve the security. Nobody's going to listen." And it's like, we weren't

listen." And it's like, we weren't happy.

Right? So, that is being taken away from us, and that is a serious problem, and it's going to cause a lot of anxiety for good reason.

However, what I'm trying to help get us to is what comes after, what we should have been doing in the first place, which is kind of along the lines of like,

everyone is kind of doing what you and I are doing. It doesn't mean they

are doing. It doesn't mean they necessarily have to use YouTube. It

doesn't mean they're interviewing people, but people have to figure out what they are about, who they are, what they think about the world, what they want to change in the

world, what they want to produce.

What ideas do they have? What art do they want to create?

Right? How do they activate themselves to actually be broadcasting value into the world? And then, what that kind of

the world? And then, what that kind of looks like is, okay, you can still join a startup, right? You start a startup, and I'm broadcasting my value to the world. I've got my own blog. I've got my

world. I've got my own blog. I've got my own YouTube. You've got your own.

own YouTube. You've got your own.

And you're like, "Hey, do you want to collaborate on this business?" Yeah, we work together, and that's fine. But

we're working together as equals, partners.

Trying to accomplish a goal as opposed to you are the special person who is hiring me, and we're in this military-like structure where I just do what you tell

me, and um you know, it's very very old school, and Yeah. it it's just not the way to be. So, what what I'm trying to do through this Pi project and and other

stuff that I'm doing is essentially communicate to people that, "Look, it's time to activate. What you had, whether you liked it or not, it's going away.

You've got to figure out what you're about. You've got to figure out You've

about. You've got to figure out You've got to be able to articulate that.

You've got to figure out which problems you want to work on. I love the idea of start with problems. Start with problems as the center of your career.

So, when you describe what you're about, you say, "I think this is the biggest problem in the world, and that's why I'm building this to help solve it." That's

a very clean narrative for describing your place in the world. The way I see this whole Pi thing is just enablement for all of that. It's enablement for interviewing you, asking you what you're

actually about, extracting this stuff from you and saving it, understanding your goals. I mean, what One of the

your goals. I mean, what One of the things I've always encouraged in people who I worked with was, if you know, the head boss or so whoever walks around the corner and says, "Hey, who are you? What are

you about?" Or an alien shows up with a clipboard, "Who are you?"

What are you going to say?

Yep. And far too many people in the world will say, "I work for Mr. Johnson.

I prepare his slides." And it's like, "Oh, cool. What What are you about? What

"Oh, cool. What What are you about? What

What are your ideas? What What do you think about the world? What do you want to produce?" And it's like, "Oh, no, no,

to produce?" And it's like, "Oh, no, no, no. That's That's not for me.

no. That's That's not for me.

That's for people like Mr. Johnson." And

if you think about our education system, this is what we've been told for hundreds of years. Yep. We've been

training workers to work for special people, right? So, all that roundabout

people, right? So, all that roundabout way to get to the answer to your question, what comes next is we we have to activate ourselves. We We have to actually be broadcasting our

capabilities, and that starts with knowing what they are. And we got to believe in ourselves, and and actually move to a human-to-human value exchange as opposed to like this bottom-up

corporate value exchange. I love that. I

mean it's uh we when you when you were talking, I was just thinking, you know, rather than being a cog in someone else's machine, you do what you want to do, and you bring value to the world from yourself

rather than just following orders and hating life, right? I think you've said it many times, you know, if you dread Monday, uh then you've then then then you're doing something wrong. Yeah,

absolutely. I think that's one of the best metrics is, how How do you feel about Monday? If you're excited, you

about Monday? If you're excited, you can't wait to go to work, I feel like you're in a good place in life. And

[snorts] I mean, how do we get there?

Yeah, I I think what's coming is going to be really nasty and really drastic, but all hopefully in service of something better on the other side.

This sounds like it's scary, like we might all lose our jobs. We might be replaced by machines, like agents or whatever it's going to be.

But there's a there's a false economy here.

The companies like OpenAI, etc. are spending crazy amounts of money trying to build out infrastructure. They

are consuming electricity. I've heard so many cases about like in the US and other places where electricity bills are like counties or whatever it is have gone up because the data centers are

taking all the energy. So they they it it feels like it's a bubble. People

are saying it feels like this this is going to all burst because there's not enough electricity, there's not enough water. Uh this is just not scalable.

water. Uh this is just not scalable.

Companies are pushing this idea that AI is the savior to the world's problems just to push their stock price up. So

devil's advocate, what do you say to that? Daniel. Yeah, I mean I think

that? Daniel. Yeah, I mean I think there's truth to all of that to to some degree. I mean there's no question uh

degree. I mean there's no question uh a lot of companies are greedy, no no question. And I do think there is

question. And I do think there is something particularly dangerous about the top, you know, three or four or five uh companies loaning each other like billions of dollars and it's like the

shell game of it like being passed around and then there's all this massive hype around like oh AI's going to save everything which which I I don't think is is necessarily true.

Um And I think there's a there's a serious risk in the next few years of like all these billions of dollars come due for payment

and the value has not happened yet.

Right? So it's like Yeah. you know, the narrative will be well, it failed. It

didn't work. And

the actual reality in my opinion is that companies who have adopted it and people who have adopted it, both of them both people and companies who have adopted AI in the in the type of way I'm

talking about with like Pi and basically self like they've learned about themselves.

Gandalf levels using encoding, role play, and creative prompt framing techniques. That's still going. It was

techniques. That's still going. It was

still going. Yep. It's still going.

I I just love that. I mean you and I were we're talking about something else and that thing was just going on in the background and I think that shows the power of it, right? You just set it and let it go and then you forgot about it

basically and we carried on with what we were talking about and it was just carrying on attacking.

I love that. Yeah, I frequently have like dozens of these things going on different projects, doing research, just in the background like preparing things

for me to come back to. Um I love that.

Yeah. Yeah, yeah. So I I think um there there will be mini crashes uh and maybe even a big crash.

Um but it doesn't mean the thing didn't work because the companies who have actually figured this out and the people who have figured it out, they are basically going to shoot straight up into the sky. Right? Which Yeah. I I've

I've already been along this lines where I've just been magnified like many orders of magnitude in my capabilities and I think there are many many people who are like that who are going to be,

you know, magnified. I think Google and companies who really really get it and really use AI appropriately they're just going to crush it. Lots of other people won't. Lots of investment companies,

won't. Lots of investment companies, they like they're like oh, just you've got to release something with AI. Lots

of companies um the board said you must do AI things and they're just kind of grasping at straws. They're not really thinking about it. And like a lot of those projects are just going to fail.

So I don't think it's one or the other. I

think both are going to happen. That's

interesting. I love what you said.

You've said this many times from what I've what I've seen. You said that the reason you created Pi was to bring AI to all of us rather than just the top 1%, right? Yeah, absolutely.

right? Yeah, absolutely.

That's always been a fear of mine. It's

really for any tech, but especially with AI is that when you have a magnifier of capability, it tends to go into the hands of people who are already capable.

And what I'm most excited about is activating the people who have the ideas but um how are they going to get on YouTube?

Right? How are they going to like write a script?

Who's even going to tell them that you actually can do this? Yeah. This is why I'm so excited about this because I want them to have Pi where they can just talk to their Pi um

their named assistant and be like, hey, I've got this idea for this anime series or I'm going to start um thinking about doing a startup but I'm I don't know.

I'm I'm just stupid. Like I I just this will never work. And their DA responds, are you kidding me? This is actually this is not something that's really being done and I think your approach to it might actually work. Let let's talk

through this. I can actually start the

through this. I can actually start the LLC for you.

I can do the research for you on other competitors in your town. And you know, we can actually do this. And suddenly

way more people are starting businesses, way more people are releasing like you know, uh series of, you know, high quality TV or movies that are competing

with a giant corporate, you know, Paramount or whatever.

And it's just like that is that's what we should be using these tools for.

Yeah, I think it's a concern, right? And

I I you hit it on the head there, I think where it feels like the top 1% or however you want to phrase it are just going to make a lot of money and have huge leverage and a lot of people are

going to lose out. So I'm really glad that you're doing this, you know, giving everyone the opportunity. But I also love the message about you trying to tell people that they are worthy or they good enough to do something themselves

rather than just be a cog in someone else's enterprise. Yeah, I I I think

else's enterprise. Yeah, I I I think that's that's kind of the central push for me is uh yeah, I I I don't want just the people like us to get better and better

at AI and then everyone else is screwed.

I also don't know like how does that game even play out? Like let's say the top 1% starts producing 10 or 100 times more stuff. Who's going to buy it? Who's

more stuff. Who's going to buy it? Who's

buying it? Who's buying the stuff? Like

all the restaurants will close. Like

society stops, right? So so I think in the next few years we're we're going to see UBI. I just don't see it any other way

UBI. I just don't see it any other way than some sort of UBI push because uh yeah, the automation is is just getting too good. So I mean it's interesting that you say that. I I I've

heard you on debates with others and um some people are very negative about this saying look AI's overhyped.

You've been doing this for a while now, well, quite a long time should I say. Is

your gut really still positive that that's where we going? Um

you know, there's all the negative stuff about it and like it's a bubble or whatever. Do you do you still believe in

whatever. Do you do you still believe in your heart that this is something we need to everyone needs to get on and needs to get invested in and start leveraging because it's going to change the world whether we like it or not. Oh

yeah, that that that's yeah, that's never been a question for me.

Yeah.

[clears throat] Um the only question I have is like what's actually going to happen?

Right. And they're

and being a security person, I'm always thinking of the the negative scenarios.

Yeah exactly.

Some of the main ones are just like well, you know who's going to get really good at it? Authoritarian governments.

They're going to monitor everyone.

They're going to shut everyone down.

They're going to pay them like this minimum UBI wage um and build a bunch of prisons and like that's nasty. Right? And then you got

that's nasty. Right? And then you got the other one that's like the rich people take over. It's it's the 1% of 1% and everyone else lives in like favelas.

Yeah. That's really nasty as well. So

what what I'm trying to do because I'm like um I'm vulnerable to like uh being overstimulated with like like concern and like negative emotion, right? So Yeah. what I'm literally

right? So Yeah. what I'm literally trying to do is say, okay, is there a path possible that we can barely see where it's actually good on the other

side and then just go full force into basically trying to manifest it. Knowing

that it's it's might be a very small it might be a 1% chance. Might be a 60% chance. I don't know because we don't

chance. I don't know because we don't know the future.

Yeah. But just assume it's possible, try to push towards it knowing that um these other bad scenarios might actually happen. They're more likely.

actually happen. They're more likely.

But I love what you're doing, right?

It's I think it's that whole hacker mentality. It's like, [snorts]

mentality. It's like, [snorts] okay, we've been told this is the way and we and we will say no. We'll find

another way. And like giving in the middle finger to the to the to the machine or to the man or whatever you want to say. But you saying, look, we will take this technology and we will give it to everyone who wants to use it

and we will help everyone rather than the 1%. I love that. Yeah, absolutely. I

the 1%. I love that. Yeah, absolutely. I

just think of how many kids don't have tutors. Think of how many kids wish they

tutors. Think of how many kids wish they could ask yeah, but why? Yeah, but why?

Yeah, but why?

Just over and over. If they had something like Pi and some sort of gadget or something on their phone, maybe they don't have a phone. If they have some sort of very

phone. If they have some sort of very cheap AI we can give to billions of people that answer the questions and give them like hope that they can become something. I I just feel like there's so

something. I I just feel like there's so much potential to unlock in in uh humans. I love that. I mean I also I

humans. I love that. I mean I also I love the way that you you mentioned schools because I am I've seen this I mean I went through the school system.

You they really train you to be a like part of an organization. I mean

that's They really do.

Yeah, it's like and your creativity often gets stifled, stuff like that. And

then I've seen people who, you know, didn't do well at school, but like amazingly creative or whatever. So I

love that you're looking at this as, okay, this gives you an option. Don't

believe the negative stuff that you can't do do Believe in yourself. You can

do this. And this is a way to empower you to do that, right?

Yeah, absolutely. It's

It's funny you mentioned that because it starts Yeah, I'm trying to build this um this platform going to come out soon.

It's called human 3.0. It's basically

like training along these lines. But the

first part is what you said, which is like it's frames. It's mental models. It's

it's frames. It's mental models. It's

changing the way you think about the world and what is possible. That almost

has to happen first before you can activate it with the technology.

So, all of that's sort of part of that.

Any I don't want to put you on the spot.

Any ETA for that? When can we expect that sometime in 2026 or like in the next few months? What do you What are you planning? Yeah, I'm looking at

you planning? Yeah, I'm looking at February actually. February

February actually. February Yeah, I'm going to have it out soon.

At least the first couple courses, yeah.

Daniel, I want to thank you so much for sharing. I want to thank you for

sharing. I want to thank you for creating these tools, putting them on GitHub. You've been doing this for

GitHub. You've been doing this for years, creating amazing tools, amazing stuff to help people in the community in the security you know, sort of niche, but now in the AI niche as well. And I

love that you're making this not just about like a security product. It's

about helping people improve their lives, helping them turn stuff on in their lives so that they see themselves as more valuable than perhaps they saw themselves in the past. So, thanks so much for sharing. Really appreciate it.

Yeah, thank thank you for having me and thanks for getting the word out.