TLDR
Regulatory capture and market manipulation are reshaping AI access, with companies like OpenAI and Anthropic voluntarily holding back frontier models like GPT 5.6 and Fable at government request, creating a class system where the best AI is selectively rolled out to big companies and not to end users. This trend threatens to restrict local AI, force cloud dependency, and limit competitive capabilities for small businesses and entrepreneurs.
Key points
- GPT 5.6 and Claude's Fable were voluntarily withdrawn at government request, signaling regulatory capture and market manipulation in AI. Government steps are laying a foundation for AI regulation, focusing on contract, restrict, and regulate. Cyber security risks from leading models are cited as justification for control. The November election cycle will intensify this topic; viewers should voice opinions to local representatives. Big companies get first access to frontier models, creating a potential class system. Local AI and model downloads are advised as a countermeasure against regulation. The push to regulate may force users into cloud products and license-based access.
- Regulatory capture is similar to how CISA mandates software like Microsoft Windows in certain industries.
- Frontier models are being selectively rolled out to companies, not to end users, on an approval basis.
- This creates a non-market force where well-connected competitors could get better model access.
- Users should contact local representatives to voice concern about AI access and market fairness.
- Hardware makers like Apple and Google are positioned differently; Google releases open-source models (Gemma).
- The potential for license-based AI access is real and would enable marketing data collection.
- The time for action is now, before regulation is baked in.
Tools mentioned
Techniques
- Voluntary submission and withdrawal
- Selective rollout
- Regulatory capture
- Market manipulation via government pressure
- Approval-based access for companies
- License-based AI access
- Using cybersecurity as rationale for restriction
Takeaways
- Frontier AI models will likely be restricted to enterprise access, creating a competitive disadvantage for small players.
- Download and run local models now as a hedge against future regulation.
- Contact local representatives to voice concerns about AI accessibility and market fairness.
- The current trend of selective model releases is a sign of broader control to come.
Transcript (captions)
The news that GPT 5.6 is getting the same treatment that Mythos voluntarily submitted itself to after the pullback of Fable should definitely indicate to you that we are in unprecedented times of regulatory capture, market manipulation, and definitely we are going to see some big changes as far as local AI. The steps that are being taken right now are laying a foundation for regulation. And you also see what's happening at the governmental level as far as the response to cyber security from the leading edge models. And so far the entirety of it appears to be contract, restrict and regulate. This is going to be intensified in upcoming election cycles this November for sure. This is a topic that you absolutely can make an impact on also. And I'm going to give you some steps that you can take today to help you have your opinion on this topic be heard. So let's take a look at the regulatory sk landscape of what's going on right now with especially the voluntary submission and withdrawal and holding back of GPT 5.6 from OpenAI at the request of the government. So this sounds like a very friendly process. Hey Sam, would you mind holding back on that for a little bit? Oh sure, no problem. Let's submit this and we'll do a company by company analysis to onboard people. So if there is a ceiling that we've hit as far as what the public can have and it is based around cyber security threats, this points to a couple very interesting potentials for me. One, there is a huge potential that there is I mean let's just be honest huge cyber security risks that are associated with this. Will the companies be able to outpace the attack vectors and patch the attack vectors is a great question and we're seeing that with Mythos and the roll out of that that happened with selective partners with Anthropic. First off, it did appear that there was a very controlled roll out. However, Fable got yanked back in record time. That right there is, in my opinion, the indicator that while things may start off looking like a we're going to do this in our own version of a responsible manner, the federal government can step right in, talk to these people from a position of power and they are responding and that is absolutely not necessarily not by their own design. There is a huge potential that all the rattling that we saw from Daario, from Sam Alman about specifically the dangers that LLMs pose to you, to the government and to the world and to infrastructure in general are definitely the reason that we are here. There has been amazing amounts of saber rattling. Same products find themselves all of a sudden being allowed kind of with caveats and asterisks. And there is a huge impetus for that to be how things are handled from a corporate perspective looking at IPOs controlling market share and especially with regulatory capture. I don't think people actually understand how this works. And some people are like your your rage against Microsoft is unfounded. Like my rage against Microsoft is really their product has gotten really bad. One two there's no company that I'm ever I've never seen this before. it it's it's it's really when you work in IT, you're going to see this at some point if you actually go down the procurement pathway. And luckily, I've not had to do that very often, but I I did actually have to go down that pathway once. And I mean, when I say that I actually am a person that has a licensed copy of SA backed Microsoft Windows Server and Exchange, let me just tell you how expensive it is. And let me also tell you that there are certain industries that just have to have it. The reason why is CISA. CISA is a regulatory oversight and they provide guidance on what is allowed, what's not allowed on an IT and infrastructure level for people that are dealing with federal kind of tasks and they are immensely powerful. This is a really good also point for me to say you should go and check what the approved software is. There's a reason that Windows is so powerful. And while usually I say, hey, you know, the open market wins out. The open market wins out in fair competition, that does not mean thumbs are not put on the scale at massive volumes from special interests to steer things in not the market's intended direction or the market's natural direction. And any suggestion of that that is like otherwise is kind of crazy. And I think most people in the audience would agree with me on that. But I've seen some comments from some people that definitely points to there's a disconnect between the reality of what regulatory capture and how it happens, the steps that are used to get to there. Uh what it looks like and I mean this started a long time ago. If you look at nent technologies that are just coming into commercial market realization, that's the time that you really want to be doing it. Also as their capabilities are growing as Windows as an operating systems capabilities were growing they were able to get the ear essentially of many legisl legislators and that right there gives them a tremendous amount of power because it's basically one of the reasons that you have to use in certain industries things like Microsoft Windows and what if in certain industries you could only use certain things like mythos fable or chatgpt 5.7 rumored to have a function-wise step improvement over 5.6 and 5.6 has already been voluntarily yanked and will be selectively rolled out to companies not to you to companies uh on a approval basis that it's individual. So the control of AI is a concept that the government is already embracing. This is why market forces appear to not be shifting in the position in my opinion of what's best for enduser consumers in most regards. That's to say specifically that there is a very big you can use this at your job but you can't use this outside of your job. That is about to happen. In my opinion, we're about to see a corporate structure for the biggest best models and that's where the frontier is going to be very much confined. Uh will this impact research? Will this impact academia? These places may be able to apply, may be able to jump through whatever hoops are put up here and may be able to get access to the frontier models out there. But the days of 5.5 rolling out willy-nilly and 5.4 4 rolling out willy-nilly are probably the last models that that's going to happen uh that are you know considered frontier. This has a lot of implications for how businesses operate. Can you get approved should be a question. What if you are stuck on a 5.5 level and your competitors got onto a 5.6 or 5.7 level? What if they were just well connected and that's how they got there? See that's not market in force. That's actual manipulation and force. And that is exactly what can happen in this scenario right now that we're in. And this is exactly why you need to get off your ass today. Pick up a phone, call whoever it is. And I I know you might think, "This guy's in the United States. He only cares about the United States." No. Uh 25% of my audience is in the United States. The remaining percent of my audience is international. That is a tremendous amount of audience that is international. And while I don't understand the politics of every single country out there, there usually is ways and means to voice your opinion to a lower level person that actually can hear you and will probably take notes and pass things up the chain. Does it do any good is also a big question. And in the United States, especially during a midterm election cycle like we're in right now, this is a electiony year topic. we need to make it an election year topic. If anybody suggests it's not that you're talking to on the other end of a phone or that you're, you know, writing a letter to, I don't think emails and letters have the impact you think they do either. Most all of this stuff ends up just being tallied. You call, you write, you do something like that goes on a tally sheet somewhere. If there is a checkbox to be checked, there may not be a checkbox to be checked. So v voicing this as a primary concern that you have and that you are watching and that you will be publicly talking about on social media with your representative stances on things will get their attention. That will get their attention. I have gotten the attention of local leaders and house representatives by doing that myself. Not flexing as a YouTuber or anything like that. Just being I would say fairly annoying. So if you do make a habit of giving calls, giving your opinion, you are actually doing what is civically the right thing to do. Can you guarantee that this is going to have a political outcome that ensures that you have access as a business, as a small user, as a startup to the frontier? I can't tell you. You're in a desperate time and nobody's talking about it like the level of actual desperation that it is right now. Right now, this is your chance. This is the time that you can do something. When this is baked in, you're going to be up a really, really hard different kind of creek trying to paddle. That'll be a raging torrent you're going up versus the gentle stream that you've got to go against the headwaters of right now. But note what I'm saying here is you are facing a challenge. We are already in a position of facing a challenge. So, don't approach it like you're not in a position of facing a challenge. This is going to take work and effort. If we want to have very clear, very well understood guidelines that are fair across the board for how models are utilized, how models are accessed and at what specific points you and your competitors if you are a entrepreneur and let me just say going into the new world of LLM powered uh systems and capabilities and you know the coding that you can do being an LLM Being a entrepreneur is not a bad option. Being an LLM might not be a bad option either. That was kind of funny. I should have let that one fly. But definitely if you're looking at what you can do, the time is now. The time is absolutely now. I would say this is what I'm going to say. This is not me putting words in your mouth. You need to form your own opinions on this. I really despise people that tell you what to think. You really do need to think about for this on your own. But my own opinion on this is that we are creating a class system of access to fundamental knowledge and competitive capabilities for businesses, enterprises, especially small businesses and small enterprises that will be very hard be it arorous actually to probably comply with and to possibly see whether or not you can actually get access to them. This is before the co this is before the cost of accessing these systems is even addressed. Before the cost of the spend on these systems is even addressed. We have no serious conversation in this nation ongoing on this on a consistent basis. And whether we like it or not, the directionality of the United States on this will have fairly large sway because our businesses compete internationally just like China's. So, this moves on to the next topic, which is probably where you as a watcher and viewer of this channel, make sure to hit that like, subscribe, huge hat tip to all of our channel members out there. Thank you for everything you do for the channel. If you are a user of Local AI, there's definitely a push right now. Um, and I would say that push is probably going to intensify in the next month to two months towards a specific goal point. And the goal point is we are going to dramatically limit and or try to regulate and or maybe possibly even under the guise of national security cut off access to international models that we have deemed as using traces for training. Even if the training was like the cold start training that like all the businesses that use basically the most expensive models have been doing and stuff like that, it it doesn't matter. It really probably does come down to they are trying and they are trying and you can see the Department of Commerce definitely with what it's saying about Alibaba is definitely trying right now. I need you to actually think about what happens to you if you can't get the latest and greatest models without the threat that you're acquiring something that you shouldn't have. What if there what if there's an edict? Hey, by the way, that's a national security threat. I you you might blow this off as like not possible, but this is totally possible. What if what what is going to happen to you and your usage patterns of AI if you can't access something? Think about it. Think about it really hard. So, I think there's a couple steps you can take. Like I said, the biggest step you can take right now is contact. Don't go for the big fish. Like the big fish will never hear you, but definitely contact the small fish. the small fish, your US House representatives, your state representatives. If you think your state representatives are like less important than your US House representatives, you probably should approach that a little bit differently because those people actually have significant amounts of power and they're kind of that next step up. They have that aspiration. They'd like to go bigger in most cases also. And so they are very tied to you and your interests and they have a very good chance of actually listening to you. You can probably schedule a meeting and even go and talk to them gently, nicely, and everything like that, but definitely voice your opinion and your strong, resolute opinion that this is not the free market at play. And I think we're definitely seeing that there's a very, very large capability of this not being the free market at play. Suggestions otherwise, I think are a little bit outside the box of reality. What happens if, hey, you know, there's a next greatest GLM 5.4, but 5.4 for is restricted technology. So a lot of people are like, "Hey, you can just go to modelcope China, download it. You can go get Torrance, download it." Okay. Yeah, absolutely you can. Absolutely you can. And there could be threats of legal repercussions if you do. Are you going to be able to secure your IP address adequately? Do you think you should set up your own VPS that is possibly not in the United States right now that you have your own infrastructure? Let me use those words. your own infrastructure. I get so sick and tired of all the home libraries pushing VPNs and it's not your own infrastructure. Like what you're teaching people here is to help people get skills and like telling them to go get a VPN is not like that's an enduser consumer kind of move. Like you can go set up your own infrastructure. The push is on like never before to regulate you. And if you don't push back, you're going to get regulated in unforeseen ways that are definitely not going to be what you're hoping for. So, should you be downloading models locally right now? I think so. I absolutely think so. I saw somebody suggest that you should not be. I was like, "Oh, yeah." Well, you take advice from whoever you want, but I think the smart advice is have your models. Have maybe a couple variants of models. If you're thinking of expanding and adding a GPU here or there, think of that. Let's talk about hardware, though. Also, this all of this has so much impact on hardware. There's very few companies out there making tremendous amounts of money on hardware. Aides from what I can tell is like Apple is making a lot of money on local AI hardware. Like without a doubt, Apple's paying attention. Is Apple in favor with the federal government right now? Is Apple does do they have the ears? Are they going to land on your side or would they rather only like oh that hash check some uh can run as a local LLM on our system? Uh we're not going to allow other ones to run on our systems. If you think about it from a kind of a let's think about it from like a Gemma Gemini kind of standpoint, like the Google perspective, they're probably one of the biggest US-based interest companies that is still releasing open source models, actually releasing models out there to the public with the GMA lineup. Very good lineup also at that. But they might uh understand the game also pretty well and understand, hey, we can gain some market share here. If we embolden this, we're going to be in a different paradigm completely. We're going to be in a paradigm of control. And that is not where we want to be. That's the first step on the despotic dystopian timeline path which you go down. You're going to be upset if you're because I'll tell you what's going to happen. You're going to end up forced to use cloud products. Oh, that sounds like a fun time, doesn't it? I'm not inventing anything here. This is already what's happening with Mythos. This is what is already happening now with GPT 5.6. They are coming out as very very selective releases. The big companies get them first. So, it's not me inventing something to suggest that that is applied more and more down the chain. So, maybe there what what what are you going to require a license? Oh, man. They would love for you to have a license and to punch in a number to use your LLM. You know why they would love it? Let me tell you what first. They love it for the marketing because they know so much information about you. They can market to you better and they love marketing to you better. And I they're good. Like I've met marketers that think like marketing playbooks and they are good and they know that the cloud presents a very good option for them to monetize at a really really good rate versus sending something to you locally. It's happening now today this week and it is going to change the future. It is going to change your potentials whether you are an entrep entrepreneur now or whether you are thinking of becoming an entrepreneur in the near term which in my opinion you should definitely be thinking about becoming an entrepreneur now today and the steps you should take. You should be prompting your LLM. Let's come up with a business plan. Let's come up with a plan of action for how we approach our first product and let's get this sucker launched. But definitely when you're looking at the potentials that are existing today that we have, you've got to be looking at this with eyes wide open because you have more challenges in the pipeline for the next two months that will change the potential course of what is accessible and usable and what you can access legally possibly even. They're they're coming down the pipeline like right now. Like right now, make that call. Pick up the phone. Definitely hit like and subscribe. Drop your comment below. I look forward to what you have to say on this one. I think we're in for challenging times and I think one of the best ways you can do something about that is to tell yourself you're going to do something about it. Pencil in some time, make that phone call. Have a great day. I'll check you guys out later. Digital Spaceport out.
Jobs for this video
| Stage | Status | Attempts | Last error | Updated |
|---|---|---|---|---|
| summarize | done | 0 | — | 2026-06-26 22:01:43.149423+00:00 |
| transcript | done | 0 | — | 2026-06-26 22:01:27.896834+00:00 |
| metadata | done | 0 | — | 2026-06-26 22:01:22.405943+00:00 |