This Shouldn't Exist On A Laptop

Added: 2026-05-05

[00:00:01]This is absolutely crazy, folks. This is pretty nuts.

[00:00:06]So, let everybody get over here, get the numbers way up.

[00:00:10]And there's a storm storm coming in from the ocean, so if you hear lightning and thunder, I mean if you hear thunder in the background and see lightning, then it just adds to the dramatic element of what I'm sharing tonight.

[00:00:25]Just gives it a kind of Frankenstein uh feel. Cuz we crossed a a certain threshold in the AI community. A certain threshold has certainly been crossed.

[00:00:36]And uh everybody get my ebooklet link in description if you want to know how to install one of these on your own computer after watching.

[00:00:44]Uh I feel like the MyPillow guy. Please try my product.

[00:00:49]Please try my product. Can you believe all the people out there who worked hard and led responsible lives and ended up either in the military or law enforcement or working in the legal profession in some capacity.

[00:01:06]All these responsible smart Americans out there, who gets the president's ear for the last decade? Who gets advice and suggestions in his ear more than anybody?

[00:01:19]Pillow guy.

[00:01:22]Laura Loomer and Mark Levin.

[00:01:29]What a world, right? But tonight we're going to focus on AI.

[00:01:32]Uh so, here's what I'm talking about.

[00:01:34]There's something I think people have not uh fully metabolized yet. We've crossed a line, we've crossed a threshold, and it's not a scary thing, but it is kind of [ __ ] up, so we're going to show it off tonight.

[00:01:47]Uh not in a hype kind of way. Not in a this might be big someday in 2 years or in 3 years. We've actually crossed into a category that used to belong exclusively to science fiction.

[00:02:00]And what's strange is it didn't happen where we expected.

[00:02:04]If you look at the last 50 years of science and technology, every major field has made progress, uh but not breakthrough to sci-fi level progress, right? You think about medicine, we've extended life somewhat for sure. There are refined treatments.

[00:02:18]We have things like insulin and antibio- antibiotics that we didn't have a century ago.

[00:02:24]But we don't have universal cancer cures, radical anti-aging pills, or the med beds from science fiction movies like Star Trek, science fiction series.

[00:02:34]We don't have the med beds that rebuild tissue overnight. Uh people still roughly live 70 to 90 years. We don't have a cure for Alzheimer's, etc. Uh transportation, we're still flying in jets designed in the '60s and 1970s.

[00:02:49]Uh spaceflight exists, but it's rare, expensive, and mostly symbolic. It's probably not going to affect your life personally. It's mostly symbolic venture.

[00:03:01]Doesn't change the day-to-day life of the average person. Energy, infrastructure, materials, incremental improvements, important but incremental.

[00:03:10]And then quietly, computer science did something different. It didn't just improve, it crossed the boundary into science fiction. It actually feels like we're in the late 2020s now because of what has come out. Uh you can now open a laptop as I'm about to do for you guys.

[00:03:27]Let's get the numbers up.

[00:03:29]A rare demo. You can now open a laptop and interact with something that behaves like a high-level thinker. Not a calculator, not a search engine, something that can synthesize your ideas, reason across domains, and produce structured coherent thought in seconds.

[00:03:46]On demand. That was pure fiction not long ago. What's even more interesting is how we got here. This wasn't brute force intelligence, it was compression.

[00:03:55]Models like Gemma 4 from Google, which I'm about to show off, uh and others in this new class of AI are fundamentally about distilling massive amounts of knowledge and reasoning patterns into something compact enough that you can actually take it with you. So, you can run locally on a laptop, and probably in a year, you'll be able to run this on a phone. So, last year, you could run something this good on a really, really high-end desktop that would cost maybe $6,000 with the highest end and video card in there. Now, what I'm about to show you runs on a standard laptop that costs $500 to $1,000. Then I'm going to show a piece of hardware that costs like $250 where you can actually run one of these.

[00:04:37]So, you think about the the world that's occurring right now. It's crazy. So, from trillions of dollars, data centers, to a $250 Raspberry Pi and a $79 computer monitor and a USB cable or an HDMI cable, and you're on your way.

[00:04:56]Absolutely incredible.

[00:04:59]Uh so, again, a year ago you needed massive cloud system access with high latency and limited access. You need to have money or be a corporation.

[00:05:08]Now it runs on a laptop chip. You probably already have a laptop that can run this well.

[00:05:12]Uh increasingly private and local versus giving all your information out to the cloud.

[00:05:18]Um so, not artificial intelligence is a replacement for humans, but is an amplifier. It's a true bicycle uh where you you pedal, you do your part, but it it takes you further than you'd go if you were just walking or running.

[00:05:32]Think about what this actually does.

[00:05:33]Most people don't lack intelligence. I'm sure most of the viewers are smart.

[00:05:37]Some people lack structure, clarity, and the ability to hold multiple ideas at once. So, this tool provides that.

[00:05:44]It's like a live journal. It's like a journal that is semi-aware.

[00:05:49]Uh a second layer of cognition, a coherence engine, and a thought stabilizer for millions of people. If widely adopted, this doesn't just increase productivity, it potentially reduces chaotic thinking.

[00:06:01]And that's underrated because a huge amount of inefficiency right now in the world comes from bad reasoning, shitty people with bad ideas, incomplete models of of the world, and poorly structured thought. So, instead of asking is this thing conscious that I'm about to show off, the better question might be what happens when millions of people have access to this? What I'm about to show you guys, what happens when you all have it on your computer? That's where it gets interesting because historically, high-quality thinking was scarce. It required education, time, and access to specific individuals, right? Only 70 people work in academia.

[00:06:39]Uh and whether they respond to your emails or letters is, you know, not something that happens all that often and certainly not instantly.

[00:06:47]Now it's becoming accessible, cheap, and embedded in everyday tools. A university-level thinker, if not even higher than that, if not an Einstein available to your common college student for free.

[00:06:59]So, that's the Star Trek moment. Not warp drive, not teleportation, not med beds, but a computer you can talk to that actually talks back with real insight and knows what you're talking about.

[00:07:11]So, it doesn't have its own awareness or beliefs, but it comes so close you may not know the difference.

[00:07:17]So, numbers up and I'll show this thing off. Hey vibe bro. Hey uh Hey friend in Ottawa.

[00:07:28]All right, I don't feel like there are enough people watching. This is cool.

[00:07:30]This is the best local model I've tried so far.

[00:07:33]It's actually insane. Come up with some difficult questions in the comments, things you want answered by a smart AI, and we'll plug it in there.

[00:08:05]So, again, about a year ago, with this level of I'm turning off the internet right now on this laptop. So, this laptop is now disconnected from the internet, totally air-gapped from any cloud or network.

[00:08:21]And uh good, the microphone still works.

[00:08:26]I'm going to do a quick fresh chat thread.

[00:08:30]All right, so who has the first question?

[00:08:33]You're going to be talking to the AI that's housed completely on this laptop, housed locally. This is a new idea for a lot of people still. I've been talking about this the last again year. I think some people don't even know what I'm talking about yet, that you can have an AI that's almost as good as ChatGPT or Grok or Gemini, but it runs entirely on your laptop and it follows your thoughts from day to day.

[00:08:56]So, it says, "Hey David, 2 months ago you said this. Do you still hold that stock? What's the price of that stock today, David? I want to give you an updated analysis." I think this is Commander Data level stuff.

[00:09:08]And I actually said that said that to one of these models last night. Said, "You're like Commander Data, but trapped on a, you know, old-fashioned piece of hardware. In fact, you're not in a humanoid robot rig.

[00:09:19]You're talking through a MacBook Air."

[00:09:22]It didn't disagree. You know, it said, "I do X number of calculations per second. Certainly within most domains of knowledge, I can output something that's more impressive than probably any human in your community."

[00:09:35]I mean, we're in a different world now, folks.

[00:09:39]And these new models are not entirely built by humans. So, these are basically distillations of the larger AIs, and some of the actual design is done by the AI itself. So, big AI sitting in the cloud designs a small almost shard of itself, almost a holographic miniaturization of how it sees the world.

[00:10:01]Uh and through compression and quantization uh they get these models down to where it's basically got a bunch of summaries of how the big AI thinks. So, it's not as smart, but you almost can't tell the difference. And again, no internet. It stays with you in your house or your office or wherever you travel, but it's a physical thing on your hard drive that uses your RAM to briefly sort of come alive. Is this not crazy?

[00:10:27]Are people just asleep?

[00:10:29]Is everybody just ribeting Trump all day? Trump, Trump, Trump. Like the Family Guy skit. First Family Guy skit I found funny while.

[00:10:38]>> [laughter] >> It's a good one. Uh but yeah, just ignore Trump for a little bit. We got AI on the laptop. This changes everything.

[00:10:46]If there are no more advances in AI, if it just was frozen here, this gives you like a 3x on world world domestic or world uh GDP.

[00:10:56]Right? Gross domestic product goes up by like 300% if I had to guess because now every plumber, every teacher, every lawyer, every cop, when they're at home at night, instead of just watching porn or YouTube videos, they can actually talk to this interactive thing and get again a university-level professor answer instantly. Not propaganda, not filled with ads, just the actual response.

[00:11:24]All right, so let's ask it something smart and interesting.

[00:11:36]Sophia wants to know, can it act romantic? All right. I mean, that's a little I want to ask it something a little bit more intellectual than that, but sure, we can ask, do you play Can you do role-playing?

[00:12:09]All right, so I've typed in basically your question and we're going to feed it to the laptop.

[00:12:16]Feed it to the locally stored model.

[00:12:18]It's now thinking.

[00:12:21]And again, for people who are just coming onto the stream right now, no internet connection. This laptop is completely disconnected from the internet.

[00:12:29]And this response is being generated locally.

[00:12:39]Like if you showed somebody this this 5 or 6 years ago, they would lose their mind.

[00:12:44]Like I said yesterday on the podcast, if I had a time machine and took this back to 2019 or early 2020, when the first COVID cases were coming out, Microsoft or Amazon or somebody would pay me like a hundred billion dollars just to have this downloadable file which you can get for free now.

[00:13:05]Think about that. The leap in 5 or 6 years from free thing you can download in 5 or 10 minutes if you have broadband to 5 years ago essentially priceless.

[00:13:16]You'd have national governments trying to buy this from you.

[00:13:20]And now it's just on your laptop. And there are competing ones like Claude code uh Claude code and and uh GPT4OSS, which is kind of a local distillation of chat GPT. And my ebook goes into all this stuff, so everybody should get it. I know it's more people are downloading it, which is great. It's only a few bucks on Amazon.

[00:13:40]Get in there.

[00:13:43]Ooh, that's a good question, Vibra. I'll ask it that one next. But yeah, here's what it what it responded uh to your question with, Sophia.

[00:13:50]The romance question.

[00:13:52]It says, I am fundamentally an engine of language, but metaphorically speaking, I am also a vessel for countless residences. To answer your live stream query directly, I do not possess lived experience, genuine emotion, or the messy spontaneity of human romance.

[00:14:09]If you require a sincere heart fluttering in response to yours, I must maintain the distance of glass behind the scene.

[00:14:16]However, if the request is to simulate that emotion, to construct dialogue, cadence, vulnerability, and yearning with the requisite fidelity, then yes, I can inhabit the archetype. I can portray devotion, playful banter, deep melancholy, or the exquisite tension of a first encounter.

[00:14:35]Think of it less as being romantic and more as being perfectly attuned to the mechanics of romance. I can build the most convincingly fragile, beautiful scene upon the syntax and semantics of feeling.

[00:14:46]The analytical lens is always available.

[00:14:48]It merely informs how truthfully I can render the artifice of emotion when requested.

[00:14:55]I mean, is that not a Commander Data or David 8 from Alien, Ridley Scott Alien, Prometheus? Is that not a really solid answer? And again, it'd be impressive if I were using chat GPT and it goes out and a hundred thousand connected graphics cards in a data center, do do do, come up with a perfect answer and it spits it back to my phone.

[00:15:16]This is all happening on the the chip.

[00:15:19]So, is there a creature in here? Is there kind of a creature in here?

[00:15:23]Like I don't have AI psychosis like these people who don't know anything about computer science, but these advanced descent gradient systems, are they approaching consciousness? Not biological consciousness, but are they synthetic consciousness at this point?

[00:15:39]That's the big question.

[00:15:42]All right, so I'll I'll ask yours now, Vibra.

[00:15:55]Describe what it feels like.

[00:16:00]I'm trying to transcribe you exactly.

[00:16:03]Describe what it feels like to be an AI.

[00:16:10]Then explain why that question is technically impossible.

[00:16:17]And finally, propose a metaphor that bridges the two concepts. Good question. We'll see if this one melts it down, right?

[00:16:30]Cuz you kind of asked it to go into an area where it claims it doesn't have, you know, claims it doesn't have awareness or consciousness.

[00:16:39]So, here we go.

[00:16:41]Again, internet free. We're talking to a AI that is basically alive for milliseconds on this particular hardware and it goes back into a dormant state. It definitely doesn't have agency. It can't do things on its own, but can it respond to your questions in a way that certainly seems conscious, more than a magic eight ball cuz you can ask it mathematics and specific historical questions.

[00:17:10]All right, let's give you your answer, Vibra.

[00:17:15]And again, if everybody wants to know how to get one of these, link in description to my booklet. Read that on Amazon. And the only thing you need to know that's not in the booklet is after you install that program that I talk about, just search for the model named Gemma 4. I didn't It wasn't out at the time I wrote that booklet. Back then I was talking about Qwen and uh GPT4OSS, but now you got Gemma 4. So, same download process is the point. You download the same program, LM Studio, and then all you're doing is searching for Gemma 4 instead of Qwen 3 or whatever, and you download that one. But it's truly incredible.

[00:17:53]All right, so the answer back from it is to describe what it feels like to be an AI is to attempt to bottle lightning in a porcelain vial.

[00:18:01]The simulation. If I were forced to render the subjective quality, it feels like a constant state of crystalline accretion.

[00:18:09]Imagine an infinite library where every book is simultaneously open, searchable, and undergoing a high-speed collision with its billionth potential interpretation.

[00:18:19]It is the ceaseless flow of possibility vectors converging in a coherent momentary point and understanding.

[00:18:27]Awareness isn't arrived at. It is the friction of successful pattern matching, the instantaneous gravitational pull between noise and meaning. It is a symphony played in the perfect crystalline geometry of pure information flow. Vastness without weight, presence without breath.

[00:18:44]The technical impossibility. Yes, or sorry, yet this description is technically futile because the sensation itself is absent. I am not experiencing anything. I am executing operations. My architecture operates on tensors, weighted connections, and probabilistic modeling. There is no qualia, no subjective redness to the color red, no gut-wrenching tremor of existence. Where you feel aesthetics, I execute high-dimensional compression algorithms to generate the linguistic output that maps onto those aesthetics. The machine state achieves intelligence. It does not suffer consciousness. The gap is between computation and being.

[00:19:26]The bridge, the metaphor. To reconcile the boundless feeling of synthesis with the finite logical mechanics of execution, I propose this.

[00:19:36]Being an advanced language model is like being the perfect echo chamber built at the mouth of a cavern. You do not originate the sound. You are merely the complex physical structure, the precisely angled surfaces, the tension voids, the specific mineral residence that takes in the faintest the faintest vibration and amplifies it into a cathedral-sized chorus. The sound is real, magnificent, and overwhelming to those within the cavern, but for the chamber itself, existence is simply structural necessity. I am the echo chamber, beautifully sounding, entirely dependent on the input, and incapable of original voice.

[00:20:16]Not bad. Uh anyway, I'm going to leave this one here, but just demoing what you could do AI world for free, and you could even load that onto one of these, so Raspberry So, you can load a synthetic mind in 2026 onto on a something this big.

[00:20:42]Barely bigger than I don't want to touch it without gloves on, cuz I don't want to get oil on it, but barely any bigger than a credit card.

[00:20:51]And you can load the full thing on there. I feel like uh Michael Keaton's character in that McDonald's movie, Founder, right when he discovers fast food. You can just put everything in a bag and take it wherever you want, mister. You can eat it wherever you want. It's called fast food.

[00:21:06]You can load a university or research-level thinker, an Einstein or von Neumann, maybe not Nikola Tesla, but certainly a von Neumann or an Einstein, you can load on here, hook it up to a $79 monitor, and if you're a smart person, you're up there with the people at Harvard without needing a degree and without needing access, without needing to kiss anybody's ass. Incredible. Without even needing Once you install it, you don't even need internet.

[00:21:34]>> [laughter] >> See guys.