Subscribe
Share
Share
Embed
Hackers are now using ChatGPT for command and control, an autonomous AI accidentally leaked its own private keys, and Apple just spent $2 billion on a startup that might be able to read your mind. We're breaking down the wildest AI and cybersecurity news of the week, plus diving into some tinfoil hat theories you won't want to miss! Get briefed on the latest tech drama and mind-bending developments right here.
Join industry experts and thought leaders as we dive deep into how artificial intelligence is transforming cybersecurity, shaping defense strategies, and creating new opportunities in the digital landscape.
Hey, welcome back to the pod. Today we're going to be sharing some of the latest AI and cyber security news that's out there, as well as our unique spins or takes on the latest issues that are occurring. Uh, we got myself and the world's greatest co-host, Shelby, here. Uh, and you know, we're your eyes and ears for everything related to AI and cyber security. So, let's get briefed. >> Before we start, can I just ask you a get to know you question, Bryce? >> Yeah, do it. Throw it over the fence. What you got? >> Okay. Um, alltime favorite best vacay you've ever been on? >> Best vacay. Um, there's been a lot of good vacays. I'm not going to lie. I think one of the best ones was going to Kauaii. Is that the name of it? It's like the Northern Iron in Northern Island in Hawaii. It's the rooster island because there's a bunch of like chickens and roosters there. Have you ever been, Shelby? >> I have. Yeah. >> Oh, you're like, "That was my pick, Bryce. You just stole it." >> NO, YOU'RE GOOD. THAT'S a fantastic pick. >> I I just I got there. I don't know. We must have went not during the busy season because they like had a bunch of extra cars and stuff and so I like got an upgrade to like a convertible for super cheap and then we just drove around the island and >> you know we kind of went up the west side of the island. There's some super good hikes and then it was just real chill you know we stopped at some places near the beach and got some like you know great seafood and all that stuff. So that I just I don't know the vibe is also good. It's like very chill, right? So, >> and going during offseason is perfect. >> Yeah. Yeah. Sometimes it's a little risky, right? You might get adverse weather or whatnot, but I feel like people who end up living in Hawaii end up becoming like mini weather reporters, right? Where they're like, "No, you got to check this other app to like see when the rain's coming or whatever." And I'm like I'm like >> Oh, yeah. Or when they check their app to see like where the swells are. They're like, >> "Yeah." Yeah. >> Here's how the wind is doing and here's how the waves are doing. >> Yeah. Yeah. I feel like they're real more in touch with the weather than, >> you know, like I I barely know what degree it's going to be before I walk out the door. So, um, what about you though, Shelby? What's your favorite vacay? >> So, I will tell you it's called Iskaret. >> Okay. I'm probably saying it wrong because it's it's like maybe Spanish pronunciation, but it's like X and then K carrot. It's in Mexico and it's an all-inclusive resort. Not as bougie as some of the ones that are down there in like the Cancun area. But let me tell you what was so cool about this one. Okay. >> And I've got to say, I have sent so many people there. I had a good experience. And whenever people ask for advice, I tell them that. And I have sent several families. And so I'm going to need a discount code at some point because I've been out here promoting their their stuff for free doing their marketing for them. I've sent a lot of people. So >> tell them Shelby sent you. That's what you got to do. You're like >> what it means. They're like you've never heard of her. >> Um >> so it's just gorgeous. So it's like you want to have a vacation but you don't want to do you know a ton of homework that comes with with uh planning a good trip, right? So, um they have transportation from the hotel to the um or sorry, from the airport to the hotel for one thing, which is great. Um they and it's included. So, it's all included in this one price. So, the price per room is high. It's expensive to look at the tickets, but um it includes everything. So, it's all your food. Um and they have a number of restaurants. I think like the fivestar restaurants might be more. But um so all your meals like it's so cool to just walk around. They have, you know, I need to just send you some pictures to throw up on your screen. Maybe just let it talk for itself. But basically they have all these different parks. They have um what are they called? Like underground like rivers that you can swim through. >> Oh, that's cool. >> And they have the zip lines. They have um it's an part of Mexico where there's a lot of I want to say I think it's a limestone and so the stones are really porous and they've they're like all these tunnels and so you're just picturing like crystal clear blue water in like the most um tranquil settings um just like cascading waterfalls and all sorts of things. They have an optical museum or optical illusions museum that you can go play in. It's supposed to like trick your senses. There was like as part of that, you don't really need to wear a lot for this. Mostly you just hang out in your swimsuit and you go to the different parks. Um they've got like half of a zoo in one of them. There's like a jaguar and stuff. It was pretty cool. And just like it's not quite a roller coaster, but like you sit in this thing where you're face down and you're just like flying through the jungle on like a zipline. They had the what is it called? A mud like immersion river. Like you go through this like little mud spa thing >> and you're supposed to be like a lazy river of mud. >> Okay, I can get behind that. That sounds pretty. >> That's wild. It's hard to explain now that I think about it. I'm going to need more time for that, but you'll have to go check it out for yourself. Yeah, I really like when it's like kind of I know like the price tags up front scare people, but like I like when things are like kind of all inclusive and I like also that that's why one of the reasons that I kind of like doing cruises sometimes too, right? Because it's like the price might be high, but then you're like not as worried about like all these little microtransactions that are like eating you up while you're trying to enjoy your vacay. So, I Yeah, I really like the Disney cruise lines as well. Those like those things are pretty nice, but >> I've heard good things about it. >> Extremely expensive. Not doing that every month. Um, so let's see. That's cool. All right. You'll have to send me the details. I'll check this out. I've always wanted to go to a resort in Mexico. I've never really done it. Um, so yeah, I'm >> also I think we went during like 2020. >> Oh, perfect. No one was there. >> It was like we had ourselves. It's like the Disneyland of Mexico. It was so cool. >> Yeah. Yeah. That's nice. That's nice. All right. Well, um, a lot's been going on the last week. One thing that I just wanted to talk to uh talk with you about today is unit 42 pushed out like kind of like a longer white paper and so there's been a few news articles that have talked about some of the interesting parts of that white paper. Uh now the white paper is really focused on how hackers are leveraging AI systems. Um, and one thing that I thought was, you know, kind of interesting is that traditionally, you know, hackers would kind of set up a a command and control server and, you know, maybe they would go out there and they would hack a website and then they would kind of use that as part of their infrastructure when they are compromising companies. Um, one of the reasons for that is because, you know, if they install malware on a computer and then the malware connects back to some IP address on the internet that nobody ever goes to, you know, that's pretty easy for a security team to find. But if the malware is going to like a website a lot of people go to, then that's pretty hard for a security team to figure out like, oh, this is bad traffic versus good traffic on the internet. Um, and so one of the things that they showed was a lot of the generative AI companies like ChatGpt and Grock and you know, Gemini that they give you, they have like a free tier. And so they were able to create a strain of malware that would send prompts to like a chatbt- like interface using the free tier and then the the malware would communicate via just sending info there and then another piece of the malware would pick up the info and kind of relay it back to the attacker servers. So they're kind of using the chat chat interfaces as a you know intermediary to do command and control. um because you know those are trusted websites and anybody can sign up for accounts. Um you know it'd be fairly hard for a security team to kind of detect that. Uh so I I thought that was pretty interesting like kind of new way to do C2. And while there's been a lot of ways to do stuff like that in the past, um, you know, previous to this, I hadn't really thought about using the chatgbt.com interface to do command and control um, or write malware with that. You know, it just goes to show, you know, anytime there's a new service that gets out there, there's somebody out there that will figure out a way to abuse it um, or, you know, use it as part of like their hacking campaigns. So, it's really cool report. highly recommend people check it out. Um, but uh, yeah, that's kind of the TLDDR of it. Did you see any cool news articles this week, Shelby, that you want to talk about? >> Yeah. So, have you heard of Awakabot, if I'm >> No. What's that? >> It's um, okay. So, it's basically a science experiment, as are most of these um, projects that are happening right now. Right. So, it's an autonomous AI presented as um an experimental system that can be paused at any time and it's funded with a small crypto treasury treasury. >> Okay. >> It's got internet access and it's on got onchain um control. It's also meant to like build apps. they it was created by the Gitcoin team and they as along with it um kind of paired with it they also launched a token which trades only on like the unis swap v4 market um but yeah that's kind of what a wakaot is referring to and about a little over a week ago um they had to take it offline uh because it accidentally exposed its private keys to it >> not good If anyone was considering doing that, I would say don't. >> Just don't. Just don't. If you have a private key, keep it. >> That's not what we do, right? Uh so it was published to I guess GitHub. Um which was which was interesting and I feel like that's part of like what this experiment needs to do, right? cuz there are other um like AI agents that have been given onchain access, but usually in these other cases that I've known or have heard of, it's they're like more tightly controlled where humans are actually the ones executing the trades and managing the wallets because they're dealing with money, right? So, they were kind of like a little bit more hands-on with that. Whereas this one, they're like, "Here you go. Let's see what happens." Um, and so let's I just want to kind of like talk about what went wrong with it. I don't think it's anything too like crazy. It's it's kind of your standard um security hygiene kind of things. And I guess I'll say the good news is they only lost $2,100. So it wasn't huge. >> Only 2,000. So >> yeah, like that's not what that's not the amount I want to donate to. >> I feel like that would still be a bad day. like an unlimited kind of wallet here. Thankfully, it was it had a small uh it had a small crypto treasury comparatively. So, okay. >> If you feel like $2,000 is insignificant, leave a comment below. >> I want to be your friend. I want to be your friend. >> IF YOU THINK THAT'S NOT MUCH to you, just send it my way and I will give it all the value it or all the love it needs. All that money. Um >> I think I know who's going to be picking up the dinner bill moving forward. You know, whoever whoever comment below >> as far as theft goes, I've just heard of some really bad ones, so I guess >> Yeah. No, you're right. You're right. In comparison to other crypto thefts, that is a very low amount. You're right. You're not the like not crazy. Okay, so >> no, no, you're spot on. >> Let's break down kind of what went wrong here. There are kind of three main issues. The first one was secret management. We've got creds in plain text going across um JSONV files and get history. So like just put it in a vault, put it in encrypted store. Um stop doing plain text. Um the other issue was that the key was distributed in a lot of places. They found it in 18 spots which each have their own attack surface which just means your crown jewels are kind of just lying around in too many spots, you know. >> Yeah. >> Don't don't duplicate that. And then the second issue was um there was there were some issues with like separation. Um so there was no sandboxing when it came to execution. So the agent ran arbitrary commands from untrusted URLs with root privileges. So if you're thinking about doing that, don't >> always run as root. Always run. >> Follow me for more advice. >> Solid recommendation. >> Right click run as admin. Um, >> never sandbox anything. Never >> sandboxes just slow you down. >> Exactly. You don't want any of those overhead. What would >> Just don't even bother locking your doors. It's fine. Um, the other issue was that one compromised key had a lot of access. It gave access to the wallet, API approvals, and automated payouts. So, there are a few ways that their system was being abused. One was through like these untrusted URLs. Um and then another was through so that was like social engineering with that remote code execution. They got an SSH back door um through like a compromised third party but really they should have been not trusting those inputs blindly. Um the second thing was like the bounty system that they had for you know you find a bug please submit it so that we can go address it and pay you out for it. Um, I don't think there was any validation happening. You submit a bug and they would just pay out um without validating it. So, >> okay. >> If you're thinking about doing that, don't do that. Validate first. >> Submit. Yeah. Submit arbitrary bugs. Got it. Got it. Got it. I'll have my AI was that they didn't validate it or they didn't like rate limit it or something. >> Oh, they even better. No rate limiting whatsoever, you know. >> Yeah. maybe don't auto put too much automation or l you know just lack of checking on your on your bounties whenever you got a payout system it was all automated so um that was a slight issue and then the third problem was just like I will call it system design flaw don't set it to trust by default you know uh that's just not going to work well for you in the end so as of now or at least last I heard for a walkabot it was taken offline and they are trying to rearchitect and bring up the value of their coin again and kind of resurrect that. So that's the that's the latest of the Wakaot. >> That's pretty interesting though. Like I yeah I feel like a lot of the crypto stuff especially when you combine it with like AI capabilities just I mean it's a whole like whole new frontier out there. So >> right >> uh there's been some new there's been some new capabilities coming out in the last week that you know I guess we should talk about. So the um so anthropic they have the claude models and historically um by historically meaning like over the last month you know a lot of people have been using opus which is kind of their bigger model um opus 46 to be exact to do a lot of coding tasks. Well, they have now released uh their sonnet model which is a faster version and it actually on some of the benchmarks actually beats the opus model. So like specifically on computer use benchmarks so which is uh sonnet 464.6 six ranks really high which you know a lot of people are using these models these days to like automate desktop application controls like that's like the whole cloudbot ecosystem right where you're you know taking a Mac mini and you're installing cloudbot on it and then cloudbot is able to you know open applications type things in the applications you know automate a lot of the workflows that are out there well apparently sonnet even though it's faster and cheaper to run uh is now even better than Opus on a lot of those tasks. There's still some benchmarks that Opus beats it on. Um but you know the computer use one it was kind of surprising that like the faster cheaper one is like actually better at it on some of the benchmarks. So uh so you know it seems like Anthropic's really um you know doing a great job there. >> I I I did set up my own cloudbot over the last week. Um, I know I know we talked a little bit about how secure that was before, >> but um, this is what I did to try to make it more secure. Like I took an old MacBook Pro uh, not Pro, MacBook Air. It has only 8 gigs of RAM, so it's not it's like a very old tiny one. It's got an M1 chip in it, so it's got the ARM architecture. Um, but yeah, it's it's pretty it's, you know, pretty ancient now. And then I formatted it so like there was no remaining information of mine on it. And then when I did a new install, I created brand new accounts. So like I didn't log in with my Apple account. I just created like a whole new account for the Cloudbot and all that. And >> so it's like a dedicated device at this point. >> Yeah. With dedicated accounts. Like I I would not even log into like a password manager on it. M >> so like I would just copied and paste like I cop would open the password manager on my phone and then like type it in >> if your if your bot if your clawbot got mad at you and it had control of your password manager it could really ruin your day. I know a lot of people have given it access to like their one pass like they have all their passwords in one pass and then they gave it like API access to the one pass or they installed the one pass you know guey on the box and >> I would not advise that man that seems really risky I mean cuz he could just go in and be like oh I don't have access to this thing well the password's just in one pass let me just go in there and grab it and log in you know um and so we'll see I the one question that I really have about the Cloudbot right now. Uh maybe I've been running it for about 48 hours and I have it set up so I can text it on my phone over like a uh private Tele Telegram conversation. Um not over text messages cuz I I was worried there's like people who set this thing up so you can just text it like give it a phone number, >> right? >> I'm like what? Like I just >> the interface, right, for interacting with it. >> I was like what if somebody else just starts texting my QuadBot and tells it to do something? I was like that seems horrible. So, I set up a private channel in Telegram just for us to to talk back and it does an okay job. I mean, I, you know, not everything's perfect. Um, I had it build a couple skills to essentially, you know, do things more reliably. Um, that is one thing that I like I I do think helps a lot is once it figures out a task and like does a task for me, then usually the next thing I tell it to do is like build a skill. So next time you do this, you don't have to figure it out again, right? >> So anything that I think it's going to do more than once, I basically have it make a custom skill for. >> Nice. Are you publishing your skills or this is this just for you? >> I mean, I have not today, but I'll hook you up, Shelby. >> Okay. Okay. >> And I No, I I'm sure there you know, I know there's like the skill market where people can download other people's skills. I just I don't want to do that. Like I I feel like then I've got to vet other people's skills because skills can contain like binaries and scripts or malicious prompts. And >> I'm like I just I don't want to deal with trying to figure out, you know, cuz I know eventually I'll just install some skill that's going to be malicious and it's going to just hose me over. So I feel like if the bot can just I mean the bot can code, right? Like Claude can code. So just code skills. It's not that hard to code, right? So, I um you know, I'm sure some of the skills that are on the skill market are more complex and so they make more sense not to like recreate the wheel, but I I've been enjoying it. I've been enjoying texting it on my phone. Um >> that's cool. >> And mostly just having it do like random research tasks for me right now. But, um you know, I uh I do it all. I'm I'm also having it monitor a couple like chat rooms, right? Like I joined it to a couple chat rooms so that so that it's not talking. It's just watching them and then it texts me like, "Hey, idiot. These people are asking >> they're talking about you." >> Yeah. Well, they're asking a question or they're talking about you. You should probably respond like so rather than me trying to >> the last six months in conversations. >> Yeah. rather than me trying to scroll through the all the stuff. It's like this person asked this question. Um, and so that's that's been kind of a timesaver as well. >> I do also want to mention that uh Grock from XAI, the GR K um you know they did release a new model too. They released 4.20 Grock 420. So, leave it to Elon Musk to make everything a joke, right? So, uh so I haven't been there there's not a lot of there's not a lot of data on the benchmarks. Um I did see a couple benchmarks on Gro uh 4.2, too. But um and the it was good on the benchmarks, but uh you know, I've just been testing it on anecdotal things and um not not so much coding. I'm still using mostly the anthropic models for coding, but I I've been testing it on like research task or like I've had some, you know, interpersonal problems. So then I like got its feedback on what I should do and yeah, it seems pretty solid. Um, you know, so I know that you can get access to the Grock uh chatbot sometimes for free if you have um you know a Twitter account or you have or if you're on any of the paid versions of Twitter. I think they give you Grock for free. So I don't know. It's it's another tool that's out there that what I think 4.2 is is considerably a step up for it. So, I in the in the world of grinds my gears this week, I'm I'm not happy with Perplexity, right? Like, uh, Perplexity changed their tiering structures and >> Oh, >> honestly, it's they like implemented like a 200 a month tier. I don't think it's actually 200. It's like $180 a month tier. >> Okay. And then they changed the pro tier, which is like $20 a month. They just made it so like you basically can't you can only do a couple deep research um each month uh which is the feature I kind of use the most on Perplexity. Um, I really liked I really like their deep research feature, but like if I can only do five deep research things per month, that that's the other thing that kind of annoys me about it is like I go to their website cuz I'm like, well, what are the new limitations? Like how many of these things am I allowed to do per month? And it's like it's unclear other than the fact is when I'm in the when I'm logged into the account, it's like you have one one left, you know? >> I'm like, if it's the beginning of the month and I have one left, I was like, this is not going to work for me. like I just need to figure out another solution to be able to do the same functionality. So, >> so I've been looking at like Perplexica, which is kind of like an open source equivalent >> and because like I don't I don't want to pay $180 a month. That's outrageous, right? Like that's that is >> I mean all that. That's Wow. >> But like if they had like a like a tier that was in between, not not even in between. Like if they had like a $40 a month tier, but I got more deep research queries, like I would probably pay them that. Like that doesn't seem outrageous to me. >> In fact, pro tip for anybody who's listening, I really think like $40 a month is like the sweet spot. Like where people would be willing to pay it if you're getting a better service, but you know, they're not going to like just drop off and find alternatives, right? Anything that's above $100 a month, like I mean, you pretty much price everybody out of it. So >> yeah. >> So >> I was going to say when you talked about setting up um your cloudbot >> Yeah. >> I heard about one called Oama which is >> Okay. Yeah. >> You know it's um it's another open source framework where you can download and manage your own LLM locally. >> Yep. >> Um and last week I saw that some research was done. is a joint investigation with Sentinel 1 and Sentinel Labs or sorry Sentinel 1, Sentinel Labs and Census and they found um 100 Yeah, 175,000 unsecured Olama hosts. >> Oh, I believe it. Those are everywhere, man. I I mean, I've set up I've set up Olama servers. Yeah. Like I I pretty much install that on every laptop I own, right? So, I I got to imagine there's just a ton of these on the internet at this point. So, wow, that's that's incredible, Shelby. I I didn't really think about that, but yeah. >> Yeah, >> there's kind of like an And you kind of when you're using Olama, you kind of need to connect other things to it because it's kind of just like a demon that runs in the background. So, typically you'll do like open web UI with Olama and if you were to set that up on like a AWS server. Yeah. I mean, and you don't secure it, anybody will be able to get to it. That's not great. So, what? Yeah. Yeah. I I do. Yeah. I'm not I'm not surprised that there's so so many uns if >> anyone here is listening to it, go check your server. Make sure it's still yours. >> Yeah. One thing that I did see this last week that was pretty cool was uh so Olama you use that to run models locally, but some of the models you run locally still have guard rails, right? Like they don't like if you're like, "How do I build like a nuclear weapon?" They're just going to tell you no, right? Um someone released a project. It's under the name Heretic. And uh like if you got a couple GPUs or you can spin some up in the cloud, it will you can give it the question that you want the model to answer that it's telling you no about it will go through a process of figuring out which weights inside the model are causing it to tell you no and then it will tweak those specific weight values. So it's essentially a way to get like open source models to be jailbroken, right? So um >> Oh, cool. >> Yeah. Yeah. So I I just thought typically they said in the past that if you modify the weights that they would kind of have un un like unknown consequences or unintentional consequences which would break the model and like make it kind of unusable. >> But I I'm not quite sure what Heretic is doing under the hood, but it's figured out a way. It's like per session. Like if you were to break the weights or like modify it, it wouldn't affect it for everyone. It would just for that user perhaps or >> um the weights are specific to the model that you're running essentially like >> when you send text into the model, it goes through a tokenization process. Um and then the tokens are given um values right as they relate to each other. Um so there and those are called vectors typically. Um and so um so like basically there's like if you have the word like um ice, right? Then you know you would have like a a value to the word cream or or something else that's related to it. >> Okay. so that you would know like hey these things are usually related these two concepts. >> Um and so what it's trying to do is figure out like what are these connectors inside the model that are making it tell you no and then lowering those values so that it doesn't take those paths anymore. >> Um so it's not something that's like a per session thing. It's like something where you're like modifying the files on disk to like permanently be broke jailbroken. Right. So, >> oh, interesting. >> Um, anyways, that was kind of a sidetrack, but and now there's some bad guy out there about to spin up some GPUs and jailbreak some models. But, uh, >> he's like, "Bryce gave me a great idea." >> He's like, "What's the name of that tool again?" He's like, "Where is rewind? >> The link's below." No, just joking. It's not. >> But I'm sure if you >> ethical choices here. >> Yeah. I'm sure if you look hard enough, you'll find it. But um um you know, it's just the there's always like this cat-and- mouse game, right, between like you know, defense and offense, right? And kind of going back and forth. And I think we'll see a lot more of that moving forward. Uh yeah. All right. Sorry, I know I got sidetracked once again, Shelby. Is there is there anything else you wanted to share? >> Enjoying the ride, you know, me on this journey. this journey of what happened in the worlds of Bryce's mind in the last 48 hours. You went all the way from Claudebot to uh jailbreaking models apparently. >> I did I did feel like I got just a little glimpse into your mind and how the connections get made >> and you're like and I do not want to be there. You're like get me out. >> It's a web of ideas all connected. >> Yeah, >> I do have a story for you. >> Yeah, go for it. Hit hit me up. Yeah. Okay. So, I I only found out about this today, but in 2025, >> so just recently MIT created a system called Alter Ego. >> Okay. >> And it's a wearable device. It goes on like your jaw and neck. And these nodes detect neuromuscular signals sent from the brain to speech muscles. Okay. >> Okay. Okay. >> So, I'm just going to stop that story right there cuz that's the end of that one, but we're going to come back to it later. You ready? >> Wait. Okay. All right. You're going to come back. Okay. All right. All right. I WAS LIKE I WAS LIKE just barely following to begin with and now I'm really lost. But go. >> Yeah. It we'll we'll come full circle. Okay. Um Okay. So, let's talk about Apple. Do you know that they just had a giant acquisition? >> They they finally bought something. >> Yes. Yes. And um what did they buy? >> Because it happened so quietly, right? Like >> they never buy anything. I feel like they're the biggest company that never that has the most cash and never buys anything with it. So >> So yeah, let's talk about this. >> Call them screwed. >> Um >> that's basically the way they operate. I read some commentary from an Israeli column something um and they were talking about how the way that Apple chooses their um their acquisitions is never to just catch up but they're always trying to like revolutionize and they're not looking for just um like apps they're looking for capabilities right to fundamentally expand to compensate for some sort of you know you know they don't have that in-house. So not so in like 2014 Apple had their biggest acquisition that was of Beats. >> Oh that's right. That was a good acquisition. >> Yeah. Yeah. Um so this one they there's not a lot of like details published about it. Um it's a little company called Q.AI. >> Okay. Um, I call it little. I actually don't know how big it is, but it is I I saw it labeled as a startup somewhere and it was relatively unknown. I went to their website, Q.AI. >> Uhhuh. >> Check it out. There's nothing there. It's so weird. It's a It is the most blank website I think I've ever seen since I made my first one in just plain HTML back in uh 20 years ago. So, um yeah, very interesting. It's a It's rep um not reported to be rumored to be about $2 billion. >> Two billion dollars and they don't even have a website. >> All the money >> $2 billion and they don't even have a website. >> To this company to this startup company that is Israel based that has previously not >> it gotten a lot of attention. It's not been very shiny and yet uh here we go $2 billion later. So, Apple has not made any specific comments about how it's going to be used this technology. However, um a quote from borrowing a quote from Apple's one of their senior VPs, he said, "A remarkable company is pioneering new and creative ways to use imaging and machine learning. So, it doesn't really give us a lot to go off of, right? So, we're going to talk about some speculation now at this point because we've kind of exhausted what we know about it >> and I really like it." So speculation number one, someone says AI powered wearables. >> Yeah, that's >> that's an option. I feel like that would get heavy. Can you imagine your arm holding a whole like processor on it? Anyway, I don't know. That seems like they would have to do some work to get it >> light enough. I think the general concept, and this is just all rumor, >> okay, >> is like the phone, like you're you're going to need to buy like a iPhone Pro and then that's going to have the processing capabilities to run the wearables. >> Like that's the rumors that makes more sense. >> So, so you wouldn't need like a whole another like CPU on you, right? That the phone would be that. And I think that's where they're going with because they have like the normal iPhones and then they have the Pro ones. And the I mean there's there's some differences, but like there's not like a ton. But if they like gatekeeped a bunch of AI stuff behind the pro label, I mean they could probably jack prices up and make even more money. So I mean that's what I'm guessing they're going to do. But go ahead. Sorry. I'm >> here's the next speculation. >> Yeah. Go. >> Um this is going based off of what patents has Q.AI been filing? >> Okay. Um they their patents indicate that their tech has is being used in headphones or glasses using quote facial skin micro movements to communicate without talking. >> So I can do it with my eyebrows moving forward. That's what you're telling me? >> Yeah. So what we're looking at right now is using AI to evaluate and understand silent speech. >> What is silent speech? >> Thank you. That's a great question. >> I I REALLY DON'T KNOW. >> I HAVE no idea. >> I'm sorry. What? Okay, so we're going to start simple whispering. You know, maybe. Okay. >> You know, I don't know if my phone can fully understand what I'm saying if I'm whispering it to someone >> across the room or something like that, right? So, whispering could be one. Um, and then it starts to get a little stranger here. So now we're talking about involuntary micro movements. So let's talk. This is where I'm going to loop it back to MIT's invention. The nodes on the face as people were wearing them. It was evaluating neuro signals um neuromuscular signals. So communication from the brain to the talking muscles before you say something. They knew what was going on. >> Sounds like it can read your mind. That's what it sounds like to me. Shall we? Is that Are you telling me they can read my mind now? >> I'm not telling you that. >> I feel like >> I'm not going to an Apple phone for a while. Um cuz they used to be well known like 10 years ago or whatever. They used to be like they're, you know, the biggest draw for me with the Apple with the iPhones was that they're so dedicated to their user base in providing them privacy. True. >> I'm just going to end my sentence right there. And now they're looking, you know, so the te the speculation is that Q.AI that their breakthrough is that it's they've used um optical instead of having a wearable. So if your devices could tell by looking at your face the involuntary micro movements of to understand like your internal dialogue before you say something or if you mouth something or if it gets I don't know if it can get to this level or not but if you think something can they accurately detect the words that you are thinking right so >> I guess >> I >> yeah go ahead >> go I was going to say, do you think that's like even possible for one? >> Uh, yeah. I think it's definitely possible. Okay. >> Um, like I I I guess I could see this being making a lot more sense now that you're describing it cuz it feels like it's going down the path that Apple's already been going down, but largely has only been implemented in the VR unit, right? Like in the VR unit, they really didn't want to have controllers. They wanted you to be able to control everything with like your your eyes, right? And >> I don't know. I don't know >> exactly. >> I don't know if it's just me, but like I am way too freaking twitchy for that VR unit. Like I cannot like it's like look to the lower left corner and I'm like like I cannot get the mouse to go there no matter what I do. I'm like it just will not go that far down. And like I I So I mean if they can get it. >> I've never tried those. >> Oh, it's I mean it's great for two minutes and then you're like tired of it being on your face. So I um >> it's just So here's my beef with it. It's just too heavy, right? Like like if they built the exact same thing in plastic and then they also put a counterwe on the back. Um kind of more like the design. I don't know if you've seen, but Steam is Valve or Steam is releasing a VR unit and they have like a they have a counterwe on the back that's like the batteries on the back and then the units on the front. Like if they had like a better balance or they changed it from the metal to like a lighter plastic. I mean, I probably would have bought one. Let's just be honest. I I own virtually every VR unit that is on the market, right? And I use them each for about 30 days and then they collect dust. So, I don't, you know, I like the tech because it's fun and it's like new, but you know, in good in good conscience, I cannot recommend like the Apple VR to anybody. It's just like the for at least I I've had my buddies come over and try it and they are able to control the desktop a lot better than I can, but for some reason my eyes are just like way too twitchy to do anything meaningful on that device. Um, >> maybe this will be an improvement on that. I'll tell you some of the potential use cases. Yeah, >> that are actually being used right now, not necessarily by Apple. But um the best I think the most virtuous um use of this kind of tech that I can think of is for people with disabilities. Um right because if you can type by thinking that would just open a whole world for them. That would be amazing, right? Um so before I demonize this whole idea because I really hate this acquisition. I really hate what the thought that, you know, maybe Apple in some number of years is going to be using their tech to monitor or like think like see our thoughts with accuracy. Um, but it could really improve some people's lives, right? >> Yeah. one. >> I think there's a time and the place for it, but I share your concerns like I don't >> I don't want it like knowing my thoughts and then like using that for marketing activities or things like that, right? Yeah. Yeah. >> That seems >> or another one is um so NASA is using it for um microphones for communications because they put them they put their astronauts in situations where traditional microphones might not make any sense, right? Um and so it being able to interpret micro expressions and put that into words is really helpful. Um, also heard rumors that Israel is currently experimenting with military uses so that their military forces can communicate without speech. >> So, >> I could see that being extremely useful like in special forces where, you know, you need to talk with your teammates, but >> you don't want to tip off that you're Yeah. that you're at a specific location or whatever. I mean, that seems that seems to have a lot of military uses, I would say. For sure. Yeah. So >> yeah, for sure. Um, and then so yeah, I guess just in conclusion, it looks like Apple is looking to really flip the paradigm. They're calling it, or not they, I just saw a reference to this like the invisible interface, right? Um, really just changing the way that humans interact with our devices. And I hope it's for purposes that just kind of suit the humans and isn't just for monetizing off of our thoughts and marketing to us when uh when our smartwatch says that our blood sugar is low. >> Little calm A, little calm B. I think that's what it's going to end up being. >> Are you are you aware that Apple's they have an announcement event coming up on March 4th? >> Oh, so that'll be soon. >> Yeah. So, um it's unknown what they're going to, you know, talk about at that event, but um some of the hardware products that they have now could probably use a refresh. I I don't So, some speculation that that might be part of the event. Um there's also speculations that they might start showing the wearable devices that they've been working on allegedly. Um, you know, allegedly they've been working on a set of glasses that is like a meta glass competitor. >> Um, so we'll I mean we'll see what they release at that event. Um, uh, I, you know, personally, and I guess they've, we've already talked about this a little bit on the show, you know, they're partnering with Apple, not Apple is partnering with Google to leverage Gemini to release a smarter version of Siri. So, you know, potentially that might be part of the event, too. I I really think the biggest thing that Apple could do just to like make customers happier is like just make it so you can talk to Siri and it's not horrible. Like if they'd replace Siri with like a Cloudbot equivalent and then had that so you could control your laptop just by using voice. Think how futuristic that would be. Because like you and I know what Cloudbot is and you and I know like we could set it up, we could use it, but think about like you know you know your parents. I mean, they might own AirPods, but they're never going to own like a Cladbot. And they might own like a Mac laptop, but they're never going to own like a Clad they're never going to set up a Cloudbot thing. But, >> you know, if they had something like that that was built into the OS and Siri was actually useful, >> um, you know, I think, you know, they could turn around their AI story pretty rapidly. I, you know, I feel like most people feel like Apple's not made the best calls when it comes to in the AI space. Like a lot of their attempts to do AI related things have been kind of like fails. So I I got to imagine some aspect of this event is going to be >> to address some of those, >> you know, shortfalls, right? >> Yeah. >> But I don't know. We'll see. Maybe they'll just release a new Mac Mini or something and call it a day. Now I'm curious. >> We will see. Um, >> funny. >> Okay. Well, I think that's all the news we got for you today. Uh, the I just want to take a moment to before we go into the fun segment of the show, right, where we talk about fun things. So, stick around for that. just say, "Hey, if you haven't um you know, taken the time to share the podcast with somebody else, uh you know, you know, take the time to share it. If you like it, you know, like and subscribe and share it with your friends cuz, uh, you know, that's really the only way that we grow and we can help more people." Um, so with that being said, it's time to move into a new segment of the show that we're calling the tinfoil hat segment of the show. Tinfoil hat time. Shelby has a tinfoil hat. I failed. I was not able to make one in time, but I will have one for next week, a tinfoil hat, so we could do better on our tinfoil hat segment of the show. I um I want to talk about my favorite conspiracy theory. So, there's a lot of great conspiracy theories out there. Uh but I've uh I've been in multiple arguments about the moon landing and it being totally fake. And recently there was a on the I don't know if you've ever listened to this, but there's like a Danny Jones podcast, which is just a podcast with like it's pretty much just pure conspiracies, right? Um where they actually had a debate on there between one of the astronauts that went to the moon allegedly, and someone who's like kind of the leading expert on why the moon landing is fake. And I got to say honestly after I ended up watching the whole podcast that I definitely like the astronaut more than I like the conspiracy guy. No offense, but um you know I you know the astronaut seems convinced that he went to the moon. Like he didn't really seem like he was faking it. >> So this is my latest theory. I'm just going to give you my latest conspiracy theory that when the astronauts come back from the moon trips that they get brainwashed so that they believe that they actually went to the moon, but the reality is they didn't actually go to the moon. There you go, Shelby. So, that's my latest conspiracy theory. That's my latest theory about conspira about a conspiracy theory is that the astronauts that came back from the moon truly believe that they went to the moon because when as part of their debrief process they got brainwashed into believing that they went to the moon. So like it's like a perfect cover story because they really think they went to the moon but human memories are awful. I don't know if you've looked at any of the studies about human memories, but they are very unreliable to the point where some experts believe like, you know, you expert witnesses in court really shouldn't be allowed. Like that's how bad like studies on memory like some experts on memory, human memory think it is. So, >> wow. >> So, anyways, if it's an interesting podcast nonetheless if you want to hear people argue about the moon landing for 2 hours. Uh, so we'll put a link to that below if you want to check that out. Um, I uh I got a I got a I got a a methodology to my madness, which is like I will get in a hot tub and I'll just lay there and then I'll put on the Danny Jones podcast and listen to some great conspiracy theories and Uh that's a good way to unwind. It's a good way. I think most of it is kind of, you know, not on the up and up, but uh there's some real crazies out there, but it's pretty entertaining. What about you, Shelley? What's your conspiracy theory for today? >> Well, I'm going crazy because I can't find the original quote, but I will have to summarize it for y'all. And if someone can find it, please uh enlighten us in the comments. But basically, um have you heard of the Monttok project? >> No. What's that? >> Um, basically it's what the Netflix show Stranger Things >> is based off of. >> That's based on a real thing. >> Well, right. It was like this secret government like research project thing. It really involved um like experiments on children involving psychic warfare, >> time travel, and interdimensional contact. and like psychic abilities and things like that. So, um but what was very strange and it's even like locked down like this was like old news, right? This was decades ago. Um but for whatever reason, like I don't know, there's still a lot of fun theories flying around about it. And my favorite was that some guy, someone please tell us in the in the comments, but some guy said who was like involved with the Montalk project or had some sort of relationship to it said that that that Stranger Things was more of a documentary than it was fiction. >> I could see that. >> That sounds >> honestly exciting because that stuff was all like firmly in the arena of fiction in my mind. But to hear that quote makes me very curious like what does he know? What does this guy know? Huh? >> Yeah. I mean probably a lot. I mean I I don't know. There's some weird stories out there of uh like remote viewing capabilities where do you know what remote viewing is? Like where you can like >> where you can like see something but >> or like project. >> Yeah. There's been some weird like edge cases where like you know the government was funding remote viewing programs and the remote viewing programs were able to produce intelligence and then they were actually able to like collaborate the intelligence with like people on the ground which is like mindblowing in and of itself. like you know that like the odds of that working out just seem like so low that like >> um >> it just feels like I'm like in like watching a magic show and I'm like how do you do it? How did this magician trick my eyes? That's what it feels like. >> Yeah. Yeah. And it's like okay it's one thing if it like they're able to guess something that's accurate once, you know? It's another thing if they're like consistently doing this like several times, right? So I I am convinced that we are all living inside a computer simulation and that the code inside our computer simulation is not even good. It's like got bugs everywhere. It's got bugs. >> You know, we might need to start a whole other podcast for this. >> Yeah. Yeah. Yeah. >> All right. Well, more more if we ever do a tinfoil hat section again, we'll get deeper into this uh matrix conspiracies. Uh thanks everybody for uh staying tuned. Uh that's all we got for you today. Um you know, stay safe out there, stay informed, and we'll see you in the next one. >> See you.