Did I Lose My Way? Did I Go Off on a Tangent?
Summary
Transcript
The new threat is AI, and I’ve tried to issue a warning in the video on the end of end-to-end encryption. Yes, the actual threat to end-to-end encryption is AI. AI is being embedded in all the major OSs, Copilot in Windows, Apple Intelligence in iOS, and macOS, Google, Gemini, and Android. And this change is the surveillance model. And the knee-jerk response is just to avoid these, which is actually the correct answer, but it is an incomplete answer. And why am I teaching you AI if this is the danger? I was recently interviewed by David Bumble.
He has a very big channel here on YouTube, and he asked me what I thought the next level of threats were to our privacy. And even when we did that interview months ago, I said the threat was AI. Yet here I am making recent videos about AI, and showing you a clear support for learning AI. Am I being inconsistent? I knew this was a hard thing to explain, but I’ll try to make it clear in this shorter video. Unfortunately, AI itself is a technology that we cannot avoid. It will be around us.
We can try to hide from it, but it will be part of everything we do. You can see that now in little pieces where you see AI integration in typical software you use, and in all the operating systems outside of Linux. Yes, there’s no built-in AI in Linux. Yet you will be surprised by this. Most of the AI development and programming is actually occurring on Linux. The machine learning hardware is all powered by Linux. So even Linux is part of this. The difference you need to understand is AI you control, and AI that controls you.
The term I’ve used recently is to refer to one class of AI as Embedded AI, and this is in contrast to local AI or open source AI where you install the AI yourself. These are two separate things. AI in itself will not hurt you. But as I tried to explain in episode two of the AI series, the OS manufacturer could embed AI agents, which are the ones that can do the evil. Yes, a new term, AI agents. This basically means a set of software that acts based on feedback from the AI. This is the trigger for things like client-side scanning, starting with the Apple CSAM project, and lately in the iteration of Windows Recall.
So it’s not just an AI. It is some secret code that can respond to the AI. Think of them as little programming robots that await instructions. That’s what an AI agent is. AI agents can do useful things, but that’s if you control it. AI agents controlled by someone else is a problem. And that’s the issue with the Apple CSAM scanning and Windows Recall. The embedded AI in the OS has agents. Now if you run the AI yourself with a clean slate, like with Linux, this is completely different. Why? Because Linux has no secret AI running.
If you want to run AI on Linux, you have to install a local AI, and this means using open source AI models. And this also means there are no unknown AI agents. These are perfectly safe because no one can control them. I tried to explain that in my last video. So in general, this has been what I’ve been trying to say, but it’s hard to do that with a set of unrelated videos coming out. It is important to learn about AI because you will not escape from it. I’m sure some of you will like to, but it will not work.
I put the AI series into its own playlist on the channel so you can watch that if it interests you. But I really do need your feedback. It’s hard to talk about technologies when the bulk of my audience think a technology is evil. So tell me in the comments what you think. Do you think I’ve lost my way? There are consequences to the recent events relating to AI. It affects what computers we buy, what OS to use, and what apps to run. I have to be here to guide you through that. But at least as a base, I will give you some general statements right now.
There’s more to this AI business than I’ve been able to express so far. And that is how AI gets personalized. How AI starts to understand personal context. And that’s when the AI becomes a mind control machine. This is why the last video aimed to explain the transformer architecture and things like AI context. It’s really hard to be a creator, by the way, because I know there’s a series of steps needed to explain where I’m going and I can explain it before making the video. So this explanation I hope will give you this roadmap.
I have to talk heavily about using Linux. I have to talk about how to transition to Linux because this will be the only safe answer. I have to tell you about keeping your data away from Big Tech because the AI they control will be your nemesis. I hope you understand that I’m not leading you the wrong way. I have not lost my way. I’m just giving you warning of a new situation that requires a new response. If we understand how AI works, we can use it safely ourselves. But we will also know how to stop it from being used against us.
And believe me, I understand this. Well, it will be too easy to use AI in an evil way. The solutions are the same. Keep data away from Big Tech, the Google open source phones, VPNs, Linux, privacy in general everywhere. So that will not change. But if I’m to continue explaining this all to you, I need to know if you’re getting what I say. If not, then I’ll have to take a different approach. So please guide me. Your reaction is very important. Thank you for watching and I’ll see you next time. [tr:trw].