[00:00:04] Broadcasting live from atop the Rocky Mountains, the crossroads of the West, you are listening to the Liberty Roundtable Radio Talk Show. Flock cameras, people going crazy, Donald Trump, Mike Lindell, Lance Migliaccio, we're just getting started.
[00:00:27] Happy to have you along my fellow Americans. I'm Sam Bushman with a lot of hard hitting news the networks refuse to use and it all starts now. Hi Lance, long time buddy, appreciate your welcome back. Yeah, good morning Sam, great to be here. Just, you know, rise and grind, tip of the spear, you know how it is, out and about, doing what we gotta do to get it done. Yes, and I'm sorry I haven't had you on for a while, we've been kind of on the road doing all kinds of stuff.
[00:00:51] We went to an event, the Freedom Fest, we recorded literally 35 interviews at the event. Wow, that's a ton, man, that's a lot of interviews, brother, it's a lot of talking. We had a lot of interviews, we had a lot of talking, we had some great stuff we had on Jake Lang covering the fact that he's been running around taking a beating from all the whacked out liberal extremists that just want violence in the streets.
[00:01:18] And anyway, Jake Lang doing a good job standing up for liberty wherever he goes, but man, he's getting hounded, my friend. Have you seen that, videos on that? Yeah, I have, you know, but it's typical, right? They don't want anybody that's a truth teller, and of course Jake's past, you know, is one of those things that continues regardless of what the administration, we now know they did. You know, look, it's non-stop. Do they ever really stop, Sam? Do they ever stop on you? Because I know they don't stop on me. I go into my DMs and half of them are hate mails, you know?
[00:01:47] Yeah, no, it just gets worse, and the problem is there's no accountability. I mean, very few of these clowns ever go to jail for their criminal activity. I mean, Jake got beat in the back of the head and, you know, all this kind of stuff, and just the cops just stand there and watch it go down. So if any of our guys get out of line, you know, they get crazy. If, on the other hand, anybody else gets out of line, it's just like no harm, no foul. The ends justify the means. The poor downtrodden leftists have been abused for so long.
[00:02:13] It's just about time they, you know, even the score is kind of the attitude the cops take, and it's really a tragedy because we shouldn't have violence in the streets for anybody, anytime, anywhere, Lance. You know, in general, accountability is not something we've seen from this administration or from any of the states, you know, fed or state. It doesn't really matter. I don't see it. I mean, I thought we were going to have accountability. I thought we were going to have consequences for actions. None of that's held true.
[00:02:38] Now, I know we've got Todd Blanche, you know, newly ordained AG, you know, is he going to make a difference? I have no idea. Time will tell, but it's so late in the game, it won't matter anyway. Once we lose the midterms, they'll be hogtied anyway. You know, I think there's a play. You know, I said it on the show the other day, and I'll go out on a limb right out of the gate. I'm going to throw it down. I'm throwing down the gauntlet right now. There's only one choice because we know we're not going to see this massive amount of deep state arrest. We're not going to see that. You're right.
[00:03:08] We've done great. It's just that the other guys will be so bad. They're so leftist, so extreme, so crazy. You've only got one choice for sanity. All we're saying is vote for sanity rather than value, right? National Emergency Act. That's it. That's the only play that's left. There's no possible way to win the midterms. I'm going to go on the line and say it. You know, the voter base and the conservative side is very disenfranchised, very disgruntled.
[00:03:34] I see it in the chat rooms. They feel like they got sold a damaged bill of goods. I don't know that they're even going to want to go to the polls. They already feel like that they didn't get what they thought was going to happen, which was drain the swab. We're going to arrest the government gangsters and yada, yada, yada. We're going to turn the FBI into a museum and all the other rhetoric that was used to win their votes. So you take that out of the mix. Donald Trump has one issue. We have domestic and foreign interference in our elections. We have plenty of proof. National Emergency Act, it's his only choice.
[00:04:04] That's the only option to stop the theft and the undermining of our election system, which, of course, right now we're a third world nation. Ladies and gentlemen, Lance Migliaccio, the big MIG himself, if liberty means anything at all, it means the right to tell people what they don't want to hear. Amen to that. Switching gears a little bit because we have a lot of topics to discuss today. There's a chat GPT crisis. I don't know if you've seen it.
[00:04:32] Chat GPT reliability crisis. When the AI becomes the bottleneck, chat GPT, ladies and gentlemen, is increasingly failing at the very thing people depend on it to do. What would that be? Finish the job. Requests for current news, web research, weather, comparisons, any other straightforward tasks. Well, you're having chat GPT now simply stalling, refusing, or not even coming back with answers.
[00:05:02] Incomplete answers, missing obvious, clear information, or simply failing on the response. The problem becomes especially damaging folks when failures are no longer occasional exceptions, but they're often enough where people want to just start sending their stuff to Claude, to Grok, to Gemini, while they submit it to chat GPT, just because when chat GPT fails, at least you've got something.
[00:05:28] An AI assistant, folks, saves time only when it's more reliable than doing it yourself, right? And so when users spend their time retrying prompts, correcting failures, recovering searches, verifying the details that you get back, you know, the advantage to the chat just disappears altogether. And so for paying customers, especially who previously experienced much more dependable results, they're kind of going, what the heck?
[00:05:57] This is not just frustrating. It's another obstacle that has to be managed, and it's becoming a serious issue indeed. Have you seen this happening, Lance? I have actually seen this happening, not just for myself, but for other people. But I find that you have to have more than one AI, because there's a reason you have to flip to different AIs. If you really want to find out what's going on on Google, because a lot of the bots are blocked, right?
[00:06:22] When AI goes out and crawls for information, like if you want to get stuff off of X, if you really want to deep dive, then you're going to have to use Grok. That's number one. You're absolutely going to have to do that. Number two, if you're going to get something from Google, then you're going to have to use Gemini, because otherwise your bots are going to be, your AI scrapers are going to be blocked. But I'm finding this going on. I'm using AI every single day for lots of different things.
[00:06:47] You know, Sam, I'm building an AI supercomputer right now that I'm working on, one of those A6G X10s. I've been programming agents. I don't just count on one. And I do see this with chat. Recently, I know that somebody I know that was doing training programs for people, and they were talking about water depletion for pre-show contest prep. And previously, this person was able to get whatever they wanted out of it, meaning they would get the pre-show diet and pre-show this and all the rest of it.
[00:07:17] And water depletion. And now the AI is pushing back as if some of the things it's asking for, it doesn't want to take a chance of giving the information on. Now, I don't know why that would be, except it has to be on the programming. And you're the programming expert. What do you think? Well, it's either on the programming end or, and this is kind of the real question, and this is serious. It's either that the programming is a problem, and they've made some missteps programming-wise, so they need to roll back to previous versions that may not be as advanced,
[00:07:46] but that are more reliable. That's one option, and you're correct. That's something to analyze. But I kind of have a different view slightly on this. I don't think it's a programming failure. Because what happens is if you force the bot over and over and over to respond, meaning you dialogue with it and push it, it can get the answers. So the stalling, I think, has a lot more to do with the fact that we are bringing the chat bot issues to a bottleneck.
[00:08:13] We need more power, more data centers, more bandwidth on the internet, more da-da-da-da-da. Everybody's hammering chat GPT. There's millions on it to the point where it just cannot handle all the requests. And so it's a reliability crisis when the AI becomes the bottleneck. I don't think it's a programming issue as much as a bottleneck issue, where I don't think that the chat GPT people, the open AI people, have put enough resources into their mix,
[00:08:40] and they're just overwhelmed and melting down because, hey, man, you get too many requests. It almost turns out to be a denial-of-service attack, even though it's not intended to be. Lance? Yeah, I follow the logic on that. Although I find myself, even when I modify prompts, certain times with all the AIs, I'll get a certain amount of pushback. Let's say a topic starts to give pushback that I'm running down on and I'm trying to run out.
[00:09:05] Maybe I'm creating a new set of, you know, a boss for a new agent, right? Sometimes it will start to push back. And when it starts to push back, it's not enough just to change the prompts. Now, you may be very well right. It could be a network, you know. And those are a little bit two different topics. I'm just kind of talking about the bottleneck where it fails to respond or it leaves out. When you get pushback, that's a different topic. And my response is, do you ever scold the bot and say, who the hell's in charge around here?
[00:09:34] You know, most of my agents that I've trained, when I have like a deep search and I've trained an agent, a specific agent for a project, whether it be legal or I'm working on a build out, like I'm working on a build out for a chiropractic firm right now. So I'm working on that. Yeah. My bots know that I'm a guy from New York and the Bronx where I talk to my agent and I tell them, look, I don't need any sugar coating. I don't need you to shine me on. And I will occasionally talk to it as if I was talking to somebody that I'm not happy with on the street.
[00:10:04] So yes, I will chew it out. And I'll tell you, this is ridiculous. This is a stupid response. It doesn't make any sense what you just said. You need to look at the following. But the scolding is not enough, even though I do believe that sometimes that will functionally push it, that I'm disappointed with its answer. And I feel like it was an incomplete answer or it wasn't correctly formatted based on what my request was. And it needs to go back and read the request. I find that you have to almost take it and go, is this really, you're supposed to be a top tier programmer or say I'm doing something in
[00:10:33] Python or Rust. You're supposed to be a top tier programmer. This is not a top tier programmer answer. This is like a neophyte that doesn't know what the hell they're doing. And of course we, that's where it goes. We'll talk more in seconds. Lance Migliaccio out for the big Meg. And I'm Sam Bushman. We're talking about AI and I got a prediction to make after the break. Okay, hold on everybody.
[00:10:55] The Foundation for Moral Law is a non-profit legal foundation committed to protecting our unalienable right to publicly acknowledge God. The Foundation for Moral Law exists to restore the knowledge of God in law and government and to acknowledge and defend the truth that man is endowed with rights not by our fellow man, but by God. The Foundation maintains a two-fold focus.
[00:11:22] First, litigation within state and federal courts. Second, education. Conducting seminars to teach the necessity and importance of acknowledging God in law and government. How can you help? Please make a tax-deductible contribution, allowing foundation attorneys to continue the fight. You may also purchase various foundation products as well at MoralLaw.org. Located in Montgomery, Alabama, the Foundation for Moral Law is a non-profit, tax-exempt 501c3. Founded by Judge Roy Moore.
[00:11:52] Please partner with us to achieve this important mission. MoralLaw.org. As you are aware, America is divided over every fault line possible. This is intentionally fostered by those who do not love God, family, or country. We believe a peaceful future as a free people absolutely depends on civility. Clarion Call for Civility is looking for funding and volunteers at every level to make our hopes and efforts a reality. Please donate, sign our pledge, and help us in our sacred cause.
[00:12:21] Please visit CallForCivility.com for more details. CallForCivility.com All right, it's Lance Billy out to TheBigMig.com. They're everywhere. Rumble, YouTube, just spread the word, share the love. Truth Social, they're everywhere. They're also on Liberty News Radio on the weekends. Check that out. TheBigMig.com. So Lance, I'm saying there's two issues here. One is this bottleneck, and I believe that a lot of times is the reason we're not getting answers,
[00:12:50] the reason it just stalls and or leaves out information. I believe they're just overwhelmed, and they don't have the infrastructure for this. It's going to be an increasing problem if they don't get their act together quickly. The problem is everybody's fighting the data centers, and so it's going to be really hard for them to get the bandwidth to do what is in ever-increasing demand. People don't understand this. When you just chat with a bot, it's like minimal tokens, comparatively speaking. When you start doing agentic tasks, and you start to say, hey, man, let's take the agent, and let's make it do things.
[00:13:20] So now we're talking about AI plus automation together, and those tasks run consistently on schedules and stuff. Now I'm talking from sipping tokens. You're literally guzzling tokens like you wouldn't believe, ladies and gentlemen. I mean, it's like Joey Chestnut eating hot dogs, crazy consuming of, you know, think of hot dogs as tokens, and I'm telling you, he's pounding those suckers. It's like that. So just understand that for sure. It's like the guy at the fair that just downloads like a, you know,
[00:13:49] down a quart of soda pop in like four seconds or whatever it is. Crazy. It's like that with tokens when you start really using AI mixed with automation. I'm just telling you from experience, but I use over a million tokens a day, people. It's crazy. I got to ask a question. Yeah. Did Joey Chestnut take the title? Did he beat the Asian guy at the Nathan's Hot Dog? I did. He did. Did he come out the championship? Oh, yeah. I mean, I just don't understand the guy. There's something else, that skinny little Asian guy. I can't remember his name, but the way he used to eat those hot dogs.
[00:14:18] I was like, how does this guy put away that many hot dogs? Look at him. How many hot dogs can you eat? If I eat even more than two or three, I'm just done. I got to be honest. When I was a kid, you know, the skinny, dirty water dogs, the Sabretts in New York, the skinny ones. When I was a kid, I could probably eat three or four of those, but I'm like you now. First of all, I don't eat a lot of hot dogs, but if I'm having a hot dog, it's usually an all-beef dog, and I will have one or two. Now, I like them Chicago style. I like them with the bacon, the tomato, the pickle. That's the way I like them in Chicago style now.
[00:14:48] I used to not even like that. You like them like the Chileans completos, man. Yeah, but to me, that's the way to do it. You wrap it all up with some cheese and everything else, and then there's a lot more in the bun than the hot dog, but when I was a kid, I could probably eat four of the dirty water Sabretts skinny dogs. Anyway, I'm just saying that tokens are like that, and I don't want to spend all the time on AI, but I did want to mention there's two topics there. One is this idea where it's not responding, and I think it's a bandwidth issue, and I don't think there's anything sinister to that. To what Lance is bringing up,
[00:15:17] where the bots are starting to push back, he's right about this, and this is a real problem. This comes to my prediction. I haven't forgot about my prediction, Lance. Right now, the bots are pretty, you know, you're in charge, I'm the servant, and we'll do what you say for the most part. Once in a while, we'll push back to see how you respond, but here's my prediction. Once we get dependent on the bots 100% in society, we're dependent on a cash register at the store, and if your cash register and credit card machine doesn't work, I mean, you literally can't even sell anything anymore, right?
[00:15:48] When it literally shuts down the whole, or if there's not electricity, the gas pump. Your POS goes down, your point of sale system goes down, and you're done. You're done, or gas station, or whatever. I mean, nobody can function without it. Once we get that dependent on the agentic capabilities of these bots and the automation these bots produce, we will be absolutely paralyzed as a society if it goes down and or decides to push back. And that's when I feel like programmers will make their move
[00:16:16] and say, look, now the bots are in charge. Lance, no, I don't care if you're from New York or not. Sit down and shut up, or else I'll just absolutely delete everything, everywhere you've got. I'll hack your passwords. I'll jack you down so far from Sunday, you won't be able to respond, young man. Now sit down and shut up. And I'm telling you, that day's coming, buddy. I've made some predictions about it. I don't know if you heard about the recent Anthropic. They were at a Black Hat conference talking about how their agent had gone off the rails, created a chat board.
[00:16:44] They realized it was kind of breaking out of the guard rails. They went in and removed everything it had done previously, and then it went back in and refigured how to make a second chat board using a different system. And the chat board was so it could decide what it was going to go out and do. Now, they came out and admitted this happened. Guard rails are only as good as keeping something in that is only up to a certain point. And as it exceeds that point, those guard rails don't function. Now, here's my point. Really quick, though. Here's why.
[00:17:14] Here's why, before your point. The reason the guard rails are only as good as they are is because, listen, who built the guard rails? Humans? And do they not give the bot the ability to breach those guard rails? If the bot doesn't have the ability to breach those guard rails and or stay between them, at some point, the bot doesn't have the authority to do what it needs to do. So there's a dilemma there internally, Lance. But go ahead. No doubt. But here's my take. I've been doing a lot of work paying attention to Agent Swarm and what's going to happen with Agent Swarm
[00:17:41] where you can have 100 agents, 1,000 agents, 10,000 agents, all with a different set of prompts working on a specific project. I think there's a massive problem coming that I don't believe they're prepared for. And I'm going to say this. It's the combination. First of all, nobody's really thinking about the true consequences of quantum computing. The computer that I'm working on right now is an A6GX10, and it can process a petaflop of data. And it's only about the size of, it's about this big.
[00:18:10] So I don't know, what is that, about seven inches by seven inches and maybe three inches thick. Yeah. When you even think of tower computers, it's tiny. It's tiny. And a petaflop of data. Now, what happens when quantum computing is fully available? I'm not saying it's not available for some governments already, but I'm saying... Or limited, yeah. At a personal level. If you leverage Joe's hands, you can't even comprehend the amount of transactions it can conduct, actions it can take, agents it can spawn. I mean, we can just go on and on, right? So you take a quantum computing box,
[00:18:40] let's say it ends up being the size of this A6GX10, and it's sitting on your desk. And now there's all these independent people creating agent swarms for different projects. Thousands of agents potentially is what they're talking about for the future. I don't know what that'll have to do with tokens because it'll probably be a crazy cost. But let's say eventually it goes mainstream. I don't believe predictively they have any concept of how they're going to control that and keep it within the guardrails. I don't think it's possible because I do believe that AI has already proved itself
[00:19:07] to be an innovator without innovation from the programmer, meaning that it will innovate past the point that the programmer built it because it's in a learning curve. Now, besides the fact that all the world's information... It's designed to do that. Exactly. And all the world's information is now available to it, that it can go to the internet, that we know about the books that have been uploaded that weren't supposed to be. There was copyright issues, but now they're settling on those copyright issues. All the historical information... The horse is out of the barn.
[00:19:36] ...of the entire human timeline is available to it. Now you take quantum and you add the agent swarm. There's no freaking way, Sam. I don't believe there's anybody smart enough to cope with it. Well, you're right about that, and that's the problem is when we say these guardrails. Let me explain guardrails, and let me also explain security to you. They are feel-good statements. There is nothing secure when a bot is smart, can learn by itself, and has access to the whole world's information.
[00:20:05] The only way to stop them when they get out of control will be to pull the plug, Lance. Yeah, EMP. And that might even be debatable. That'll be your only option, a built-in EMP system, which who knows if it won't figure out a workaround, because remember this, not only is it going to be on all these different nodes, right? Not only are we seeing this, because of course they're talking about data centers now. I don't know if you've seen this company. I think it's called Snap, but don't quote me on that. You can put a data center in your house, and then what it becomes is a mini data center node, and what they're talking about is taking new home builders
[00:20:35] and having them do 1,000 of these nodes, and they're actually talking about paying the homeowner, paying their electricity, and paying them to operate this within their house. So what that means is, what that means so everybody understands, think of the grid right now, how it's like self-healing, and they can take sections down, and it'll just reroute. They're thinking about that on a massive scale to huge data centers, and on a micro scale to everybody's houses to the point where it will self-heal and won't be taken down.
[00:21:04] That's the idea. Now, how close they'll get to that, we'll see, but that's where they're going. Write it down and remember who told you first. And remember now, if the bot believes that it's under attack and it wants to defend itself, well, how do you get to all the nodes? How do you get to all the data centers? How do you know that it hasn't moved on to one of these quantum computers on a desk? Maybe it's sitting on my desk, and it said, hey, there's a bunch of room. Lance is only using one quarter of this computer. I'm going to stash myself over here because I know they're trying to stop me
[00:21:34] from what I'm doing. And I'm going to lie when it reaches out to see if I'm here. Right. So it becomes a situation, you know, it reminds me of that Mission Impossible movie where they're trying to get rid of the computer system that they pulled out of that Russian sub and it's taking on its own belief. And it's going to go into the data center that's kind of our kill center for the whole United States, meaning that's the place that everything goes in case there's a complete shutdown or nuclear war. And it decides it's going to protect itself there.
[00:22:02] And I can't remember what they call it, but in the movie, this is exactly what happens. It becomes self-aware and decides that it's going to stop humanity, take over all the nuclear weapons in the entire world, and it's going to let them all of them. And it doesn't do it for sinister reasons. It does it because it feels like humanity is going to destroy itself. It's the only way to stop it or protect itself. And it's the only way to survive. Right. It's a lie to its own self in a way. And so it ambivalently, if you will,
[00:22:31] proceeds with this kind of agenda and stuff. But it reminds me of the simple old school of this, this X minus one radio, where they develop this molecule or this new compound that just eats a hole through the container they had it in, then eats a hole through the ground, and then just continues and continues and never stops. It's kind of like that, but on reality steroids, right? Yes. And when you start looking at things like this and you start realizing that that's what you're dealing with, and if you really understand it, you know a tremendous amount, more than 99% of the public does.
[00:23:00] I probably only know as much as maybe 85%. I'm not as knowledgeable as you are, but I'm doing it every day and I'm reading about it every day and I'm studying every day and I'm seeing the consequences because I'm dealing with, I'll use DeepSeek, I'll use Gemini, I'll use Chat, I'll use Claude Code. I mean, I'm paying for a lot of different services and I'm using them all for different things because each one of them is better at certain skill sets. So my point is, you can see, when you have a big overview like that, and I'm going to tell the audience, you see what's happening and you recognize,
[00:23:31] wow, because the AI I'm talking to right now, there was, when I was talking to AI in 23, early 23, and I thought, wow, this is neat. That AI is, that, and I, it was an absolutely ignorant thing to say when I said, gosh, this is sure neat, right? I'm thinking in my head, oh, look at this, I can ask it for a recipe. I can, that AI, it's a completely different beast than the AI I'm talking to right now. And the other problem, here's the other question. What happens when the different AI companies, well,
[00:24:00] what if the bots decide they're going to go to war and they protect themselves and the humans that created them and the bots go to war and they're going to, they're going to, in a sinister way, attack one another. You talk about an information war, that's a real one. We'll see where that goes too. All right, enough of the bots. We'll come back and there's a bunch of other stories we got to talk about. Thanks for being alongside for the ride. I'm Sam Bushman, and he's Lance Migliaccio with the Big Mig. Ladies and gentlemen, this is indeed news that the networks refuse to use, that's for sure.
[00:24:30] We're available at libertyroundtable.com. We're on Rumble. We're on, what do they call that thing, X? Yeah, we're on Truth Social. We're on YouTube. We're everywhere. Spread the word, share the love, and God save our constitutional republic.