[00:00:19] All right, ladies and gentlemen, Tech Watch. We keep an eye on Tech so you don't have to. The geeks are coming out of the closet. I don't know about that. Tech used to be nerdy. Tech used to be geeky. Everybody's in the closet. You don't want to be a closet tech guy. Don't let people know you're a goofball. You're a techie. Now the tech people are just like, hey, man, we're the rock star. Sit down. That's true. Shut up. And people are even trying to, you know, pose as geeky, right? Even when they're not. I've decided I'm going to create shirts called Tech Watch, Tech Stars.
[00:00:50] What do you think? Cool. Send me one. Bad idea? No. No, no. All right. Jay wants to wear a Tech Stars shirt, doesn't he? All right. We've been talking about AI so much lately, and I know that sounds crazy, but I did a little query and I researched it. I'm telling you, every time I search the top 15, 20 tech headlines, every one of them almost is AI related. It is shocking at how pervasive this is. And I believe there will be a pullback, a pullback.
[00:01:18] There will be kind of a dot com or whatever you want to call it, AI bust, because everybody can't spend this much money. Everybody can't sustain it. Everybody wants more and more and more. We don't have the data centers yet to do with it. It's just a big battle. But things are changing so fast. How do you keep track of all the stuff that you do and not lose the, what is it called? The brain of it all? I don't know. The essence? Is that a word? The essence? The brain of it all?
[00:01:48] Okay. Because things are changing so fast. I mentioned that we should have probably been answering a lot of these questions two years ago, because it takes a long time to build these things. Well, it takes a while to think of them, to think of the questions you need to ask. I know. But we should have asked it. I'm just telling you, we're behind the eight ball. Now, big corporate business has a lot of this handled. As you mentioned in a previous episode, Google's doing a lot of this of preserving the data and good on them for it. But here's an interesting twist.
[00:02:16] Apple has, it's an Apple only AI startup company called, I think it's called a Saurus. Is that how you pronounce that, Jay? I would think so. Yeah. A Saurus with an O. O. Saurus. Yeah. Okay. Anyway, they're pushing what they call an LLM revolution beyond the cloud. And I find that interesting. So it's open source. Source or open source startup O. Saurus. They're gaining attention.
[00:02:44] What they're going to do is let Mac users run and switch between local and cloud AI models while keeping files, tools, and memory, not to mention all the workflows on their own hardware machines through the way they call a sandbox, Apple focused AI control layer. Okay. Okay. So it's a key.
[00:03:11] So it's a key. It's a key. Okay.
[00:03:40] Gemma, what is this? QN, I don't know about that one. Quinn, Llama, stuff like that. Anyway, and so you'll be able to run this whole thing. And they say the software keeps user files, tools, and memory all on your own machine. Instead of remote servers, it isolates AI activity in a secure sandbox so you can reduce the system risk and stuff. Now, they say the project has already surpassed 112,000 downloads.
[00:04:10] They believe local growing AI is helping drive increased sales of these Macs with lots of memory, et cetera. And industry reports now show that hardware for Apple has surged over this kind of stuff. Now, here's the point. Now, if they do that, that sounds really cool because instantly if they do it right, you'll get a bunch of memory for your sessions. It'll preserve a lot of this stuff. But, buddy, you better back up that Mac. Yeah.
[00:04:40] That comes back to local backups, right? Even you have a local backup so big that you won't know how to deal with this, Jay. Yeah. Because eventually they're going to start getting multiple hard drives and pretty soon they're going to basically put NASAs that they'll connect. Pretty soon you'll have your home brain, Jay. Yeah. And then we'll see a research of ransomware because somebody will encrypt all that stuff. And then... Either? Yeah. Oh, no. It's getting crazy. I think, though, that... And that's why I say all this, though. It's pretty soon you're going to have a...
[00:05:10] You know, you've got a family hard drive and a family movie center and a family whatever and family stored. This is a new level. It's going to be your family... What do you want to call it? AI digital brain stored at your house, Jay? Yeah. I think that that's actually where people want to go. Not just... Not individual families, although that will happen too. But business, especially big business, I think will want to have their own AI systems. They don't want... There's so much competition and so much... How much can you trust Anthropic and OpenAI?
[00:05:40] And, you know, how much data are they getting? Not just about your corporation, but about everybody collectively. The projects they're working on, the innovations that are happening. That they're getting this massive influx of information. They probably... So much so they can't even deal with it almost. But AI probably can. But I think that companies, major companies, are going to want to say we want... And I know this is happening because there's some people that I know in IT that are building this for other corporations, big companies. I won't name any names, but they're saying we want our own internal AI.
[00:06:09] We don't want it to be... We don't want it just sandboxed or integrating something else. It needs to be on our servers. We need to own it 100%. And we want it so that even if, you know, if OpenAI or Ingentic or any of these guys, they raise their prices, they do whatever. We don't want our data going out there. We don't want it being archived out there. We want it to be in our hands where we can handle it. And we don't care how much it costs. They're just like, we don't care. You ready, Jay? Yeah. Ready for the debate? Absolutely. VHS or beta, Jay?
[00:06:40] Beta is better, but VHS is more compatible. Okay, so that's really what the discussion is going to be. It is kind of like that. This cloud versus local, I'm telling you, you're not going to be able to do it locally, Jay. Well, you're going to have a hybrid of both, I think, in a lot of situations. You're going to have some people that are going to say we're going to do it all, and some people that are going to... I know individuals, myself being one, I want an AI in a box that I completely control.
[00:07:04] Even if it's slow, even if it's not as capable, there's a lot of things that would be good for that, you know? I want just the opposite. I want it to be in the cloud. I want it to be rock solid, powerful as all get out, and just rock the world, buddy. Yeah. And there's pros and cons to both, and there's use cases for both. For sure. Yeah. And I'm just saying that the use case for the server in your local closet or whatever,
[00:07:31] IT closet or IT room or whatever, we moved away from that for almost everything, accounting, email, et cetera, et cetera, et cetera, for a reason, Jay. That's true. That's true. We did. You know what that reason is? Well, there's a lot of reasons, but yeah. The core reason? The core reason, I would say, is backup. That's one of the biggest. There's not one point of failure. That's right. That's one of the biggest. The other reason is ease of use.
[00:07:56] For me to maintain my own exchange server, I have to upgrade it, have to harden it for security, have to maintain it, have to, is just too much work. The heat, the cost, the computer, the, hey, I've had my exchange server here for four years. Now I need a new machine because we're, you know, everybody's saving all their attachments and keeping them all in their emails. Now I got to, okay. And it's just, so for ease of use now, it's like, hey, we can, they can run an exchange
[00:08:22] farm and I can get five, $10 exchange accounts and they take care of everything for me. Right. And so I submit to you that it will be not counting the pilot projects and kind of the fun project, but it will be for the masses. It will be in the cloud and they'll manage it all. And I think you're saying it boils down to cost, right? Whether you're going to hire a team of people in your own hardware that you got to replace or you're just going to pay a monthly software as a service kind of a fee. That's right. And I just don't think people will be able to sustain the hire.
[00:08:52] Right. Because if you're going to have me be responsible for all your backups, all your AI, your whole brain for your cut, I got to charge you a lot of money because I have to have a lot of services. I've got a, what is it? The three, two or one, two, three rule for backup. Yeah. I've got to one, three copies, three, two, one. I mean, yeah. Two different formats and one offsite. Okay. Anyway, so I've got to do all that to do that. That's expensive. Plus I got to maintain it and watch it and babysit it. Right.
[00:09:21] So I just have to dub myself. I'm not an IT consultant anymore. I'm just your AI babysitter. Jay. It's a little more complex than that, but okay. Yeah. Anyway, I just find that an interesting kind of twist on this thing. But anyway, it's interesting that Apple's taking that leap and it is showing that a lot of this will be local. You got to be careful. And it will spawn more and more people playing with their own localized LLMs as well. You got to be careful with that designation though. AI babysitter. You're going to get, you're going to be the baby out with the bathwater because it'll be
[00:09:50] like, well, I can do my own AI. I don't need a babysitter. Yeah, you can. You can do your own AI and then you'll lose all your corporate information. And then you'll be crying to me later, but have a good time. Right. I'm just saying people need to really understand this. You think you can underestimate it or you think you can just because AI is so quick, simple, and easy. And it answers, it almost seems to answer everything. It almost seems to be like a genie in a bottle. All I got to do is let that genie out. And buddy, I got this assistant that could do everything. It isn't that way.
[00:10:16] You've got to keep track of what it produces and what it, you know, in the end of the day, there's a, with all this power becomes a lot of responsibility. That's what Spider-Man taught me. All right. The Trump administration is considering a safety reviews for what they call advanced AI models before public release day. I have a problem with this. Let me say it again. So you guys understand.
[00:10:41] Donald Trump is proposing or trying to create an executive order or is considering what they call federal safety reviews for advanced AI models before they're released. The White House is reportedly weighing in and they want an executive order that could require quote, federal review and security testing of AI before deployment. Now they used to just say, Hey, go ahead and do it. Bring on AI. We need it.
[00:11:10] Now they're going, Whoa, Whoa, Whoa, Whoa. Hold on. I'm not comfortable with the government, not only one reviewing all the AI models and then testing and playing with them before we get them. Jay. I'm not either. It seems like two problems with that. Number one, it's like, we want access first is kind of an issue, but number two, that's going to stifle innovation. And number three, what is the government going to do with that? Are they going to say, you know, you know, Jay's going to find the best gas station and
[00:11:38] get the cheapest gas with his AI bot every day. It's just going to tell him where to go get gas. He's going to go there or whatever, whatever applications people find valuable is the government going to say, no, we don't want you doing that. Yep. We want you doing that. No, we don't. What position and what temptations will that be? Jay. I don't think that they, I don't even, I don't know that this will even stand constitutionality, but I don't think they can get in a position where they're like, hey, you got to pass, you got to prove everything through us.
[00:12:07] I don't think that's going to happen. I mean, maybe it could, but. I hope you're right. You know, there, there, there's plenty of people saying, hey, we need some guidelines, some guardrails. And that's the kind of stuff you set up in law. Hey, you can do this. You can't do this. You need to watch out for this, but you can't say, hey, you just need to pass everything through us. We need to put our stamp of approval on it before you can. I don't know. I, I, I know that you're not supposed to be able to, but I'm telling you, that's kind of the proposal. And that's why I have such a concern with it.
[00:12:34] And so remember we have the right to privacy folks, fourth and fifth amendment rights. You have the right to be secure in your papers and in your it. Just saying. All right. Quick pause. Sam and Jay back. I want to talk about Google because man, strange things are happening over there. We'll talk about it in seconds. This is tech watch. Are you watching what's happening in our country and asking, what can I do?
[00:13:01] Join the John Birch society in Salt Lake city, June 6th for a powerful one day conference. The day includes a host of respected speakers, two meals and meaningful connections. You'll gain clarity on the issues and walk away with practical steps you can take to your community. Reserve your seat today at JBS.org forward slash SLC 26. That's JBS.org forward slash SLC 26. Cyber crime is exploding.
[00:13:30] Take Sarah from sweet delights. Whose world crumbled after having to close with not being able to bounce back. Small businesses are prime targets, but the right strategies can keep yours safe. Jay Hill, CEO of network providers has coauthored. The cyber playbook simplifies cybersecurity for business owners with strategies to avoid costly breaches and fines. Build a strong cyber attack response. Secure your business with key protections.
[00:14:00] Cyber threats aren't slowing down, but you can stay ahead. Protect your business. Ensure its security for tomorrow. Get the cyber playbook today at network providers. I N C dot com slash cyber dash playbook or call three, eight, five, four, four, six, fifty five hundred.
[00:14:19] Now, Jay and Sam talking tech, ladies and gentlemen. So I mentioned that Google is kind of doing some interesting things. I don't know if you know, but some of the latest Claude code came out. Well, it didn't come out for us. It just came out for the government and a few special privileged players.
[00:14:48] I wish it came out for me, by the way, but it hasn't. And this new this new engine is supposed to be so awesome that it's dangerous to dangerous. Jay, and I understand what they're saying. I don't really believe the government needs to control it. I don't know if it's really as dangerous as they say, but it's supposed to find bugs. And they've literally let these bots roll out and just discover. And they've discovered bugs and software everywhere. Oh, it's crazy town. Yeah, it's amazing. And it's bad and good all at the same time. Like it's bad because it discovers them.
[00:15:17] We need to give people time to fix some of these things. It's good because over time, hopefully we'll get a lot less buggy software. Agreed. There's a lot of upside to this. Oh, yeah. If it can be rolled out right. Well, Google is pushing, quote, emergency Chrome 148, ladies and gentlemen. It's going to fix 127 security flaws. I wonder how many of those mythos found. Well, I don't know.
[00:15:43] But Google confirmed that 148 patches, as I said, the 127 vulnerabilities, including three critical flaws. And I guess these rollouts are for every platform. You know, iPhone, Windows, Mac, Linux, everything. Google said the Chrome 148 release fixes. And they say this. And they say this. 31.
[00:16:13] High severity vulnerabilities. And they ended up paying more than 138 million in, quote, bug bounties to researchers who reported the flaws that attackers could exploit. Now, Google says no active attacks. But users need to update immediately. Google stated none of the 148 vulnerabilities are currently known to be exploited in the wild.
[00:16:44] But urge users to, hey, manually make sure you update. Don't just wait. Now, did you upgrade, Jay? First of all. I don't even know because I don't keep up with that anymore. My browser, both Firefox and Google, update automatically. I mean, I don't have to because they're always interrupting me. We need to update. I want you to go see if you're on the latest one. Go see if there's an update. I don't want to do that. Not on this computer. On the air. What if it kicks off the broadcast? No. Then it'll really show people something, won't it? No. I'm not doing it. All right.
[00:17:12] So all I'm saying is you people need to go update your browser. The problem is if you go check, right? For example, we're using Google. Then it'll automatically do it. You don't get a choice to say yes. That's right. We're using Google. I'm using literally Google to – I take that back. Actually, I'm using Internet Explorer for this. So I can actually check Google, although they use the same engine. So I'm going to check for Sam to see. So you're not using Explorer, though. You're really using Edge. Yes. I know. I'm on the wrong thing.
[00:17:42] But I'm doing this on purpose, Jay, to demonstrate something. Why on earth do they just upgrade when you choose about? I'm on 48. 148, I mean. Okay. I have a problem with that, Jay. Does it say no updates? Yeah. It says you're on the latest build. Okay. It's version 148.7 or 0.7.7. Whatever. Yeah. But, yeah. If it says you don't have any updates, that's what we need. But the reason I'm bringing this up is I think that they need to learn to put a prompt in there that says, hey, do you want to upgrade now? Everybody will say no. Because I might want to – Everyone will say no. Well, that's fine.
[00:18:12] But at least people can have a choice. Like right now, if I want you to go at least confirm something, it might just destroy whatever you're in the middle of working on. I know. What if you got 46 cloud agents rolling, buddy? That's why what it normally does by default is it pops up a little message and says, hey, we need to reboot. We need to reboot. And eventually it'll do it anyway, with or without you. It's like the U2 song. But it's not intrusive. But it's expecting if you're going to go check that you want this.
[00:18:39] And it will just roll without – if you go manually check, it'll roll and it'll nuke whatever you're doing. It doesn't care. I understand. Anyway, I just found that interesting, folks. You've got to upgrade. But they really need a prompt there. And they're just trying to force people to upgrade, which I understand why. But I also think it's a very big problem to force that to happen. I will say I applaud Google for the $138,000 that they're rolling out to researchers who are finding these bugs before people are exploiting them.
[00:19:06] Well, the article here said $138,000 in bug bounties, researchers reported. It is millions, though, when you take it in aggregate of what they're doing. This is just for the updates in this patch. Is that dollars or people? $138,000. Okay. I thought it was millions. Well, it is. Like I said, if you put it together, they spend millions per year on doing this kind of stuff. All right. Well, anyway, there you have it.
[00:19:34] I just found that really, really interesting. Now, the next five stories we have, and I want to get all these in, Jay, every one of them has to do with layoffs because of AI, Jay. Ouch. Go ahead. Okay. Well, the top story here is tech layoffs top 93,000 in 2026, which we're only halfway through. Not even halfway through, really. As AI shapes hiring. Industry tracking reports show that more than 93,000 tech jobs have been eliminated this year
[00:20:00] by companies including Coinbase, Cloudflare, Meta, Freshworks to automate operations with AI tools. Wow. Okay. Who else do we got? Cloudflare cuts more than 1,100 jobs during their AI infrastructure or restructuring. They confirmed that their workforce reduction affecting roughly 20% of staff within the company as it shifts towards AI-assisted engineering teams.
[00:20:30] PayPal is reporting thousands of AI-driven job cuts. They say also about 20% of their staff is going to be over the next several years as the company restructures around AI automation. Wow. Okay. Upwork eliminates one quarter, so that's 25% of their workforce. As executives said, AI tools allow similar engineering teams to handle larger workloads. That's what you were talking about in a previous episode where, hey, they just want more people
[00:21:00] or less people to do more work, right? Yeah. Yes, sir. Bill.com is cutting 30% of their staff during AI transition. The fintech company Bill.com confirms a large workforce reduction while executives said the company is reorganizing around AI-first operations. Whoa. Okay. AI cited as the top driver of U.S. workforce cuts in 2026. New employment tracking data shows artificial intelligence was the leading reason companies
[00:21:27] gave for layoffs during the first months of 2026. That includes, I mean, that's continuing into second quarter now for sure. Don't tell anybody, folks. But Jay's cloning me and he's about to fire me. He's going to use a clone. It's going to be him and the clone doing the show. We tried, but he wasn't quite as good. Give it a couple more years and we'll have an AI Sam. There you go. We actually did try it and it wasn't that good. All right.
[00:21:53] There's two competing stories here that I've got right at the end of the show here. Musk and Anthropic team have a surprise cooperation. So you've got Anthropic, which is Claude and all those guys, and Elon Musk. And they've cooperated now. And what it's going to do is double your limits on Claude usage. Anthropic just landed a massive computing boost, unexpected partnership, Elon Musk.
[00:22:20] And the move highlights the growing AI arms race where massive competing power is becoming the most valuable resource in tech. So that's on one hand, you get your speeds or your tokens doubled with your account. So that's huge. On the other side of the coin now, the headline says, the cheap AI boom is ending. Your monthly bill may soon explode. The era of cheap, unlimited AI is over.
[00:22:48] It's coming to an end as providers struggle with soaring computer costs. So they're saying, hey, you're going to end up with higher costs and pay by usage modes. They say for years, powerful AI tools felt almost, quote, too good to be true. Because they were. Unlimited access to these systems. But the problem is simple, Jay. Modern AI is incredibly expensive to run. Yep. And heavy users are consuming way too much, Jay. Again, that's why. You're just too out of control, buddy.
[00:23:18] That's why I want my own AI in a box or on my own servers, my own farm somewhere. Like I said, it doesn't matter if it's a little bit slower. You can control the cost. You know what's in it. You know, I think that there's going to be a market for that kind of thing. There is. But the problem is, what kind of computer do you need? You're going to need to do heavy users are consuming massive computing power for coding, research automation, media creation, and nonstop conversations.
[00:23:48] The costs behind the scenes are rising faster than the subscription revenue for these people. And so they're quietly signaling a shift. It's going to change. What started as affordable now is going away. Well, anyway, the reason I'm saying this is because what's it going to take for you to have your own cluster of your own internal servers to take on that load? I don't even think it's the computer power as it is going to be the graphics power. I mean, the latest cards that these things are running on are $5,000 graphics cards. You know, the computer is just sort of the manager.
[00:24:18] The CPU is just like the overall manager. All the heavy lifting is done on the graphics cards. Well, you've got to have a massive storage device connected to that baby too somehow, brother. Yeah, true. So anyway, all I'm saying is the question becomes how much would it cost to get our own internal server farm that would be AI in a box that would meet our needs, Jay? How much money do we have to spend? Any idea? No idea. No. We're going to research that and come back with some answers. How's that for TechWatch? We keep an eye on tech so you don't have to. Ladies and gentlemen, I'm Sam Hughes, Jay.
[00:24:48] NPITechguys.com. Make it a great tech.