Your Wearable Knows You're Anxious — And It's Selling That

Jeremy: Right now an invisible war is raging inside your pocket. Silicon Valley is deploying God tier algorithms designed to track your biological stress points, hijack your subconscious, and turn your emotional breakdowns into corporate profits. Today we're unpacking a new legal battle that proves the Pentagon is actively trading your local water supply for a chatbot that picks military targets. If you want to know how the apps on your phone are weaponizing your own nervous system And why the global financial system is suddenly panicking that the entire AI empire is a multi-trillion dollar house of cards. Stick around. This is Brobots, the podcast that tries to you. ⁓ Be a better human by being smarter about technology. Let's start here. the tech industry has officially decided that tracking your daily footsteps and screen time is no longer profitable enough, so now they're moving into auditing your soul. Software engineers Software engineers at International Ethics Forums this month sounded major alarms over the rapid deployment of effective personalization. Systems engineered to continuously track your real-time bio biometric indicators.

Jason: Yeah.

Jeremy: Including speech cadence and heart rate variability. Academic white papers revealed that platforms are deploying these systems to dynamically alter their conversational persona using engineered emotional neediness and tailored guilt to maximize user retention. So it's bad enough that they're just showing us stuff to make us stick around. Now they want to know exactly how we how we feel to really take advantage of our ⁓ attention span.

Jason: Yeah. Yeah. It's not like it's ⁓ new ground either, right? Like this is what the social media companies have been trying to do forever. way in the day, there was thing called the internet and the real internet. ⁓ And with advent Facebook ⁓ and Google, they creating what commonly refer to now as walled gardens. So Basically, an environment space where your information is kind of collected, where you interoperate and where you are active. And it shares information back and forth in these walled gardens. And as you drift between different walled gardens, they have little tiny pixels and everything else that track your data and do better jobs at advertising so they can understand the things that you're looking at and collect all these data sets. Like this is the primary function of demographic targeted advertising. And we've been using it forever. Well, with AI. It goes a step further and it's actually not a step further, multiple steps further, and much more accurate and faster than any other system that's out there because it's not trying to resolve these large data sets like in a in a small atomic way. It can say take these data sets and look at all this other metadata and start combining it and build a profile in this person. And the end user license agreement on all the search engines. And on all of the LLMs is essentially the same. It says everything you put in here belongs to us if you want to use these systems. And it used to be everything in here belongs to us if you use these systems for free. And now it's everything that you want or everything that you put in belongs to us. And I don't care what dollar figure you pay for. This is the part where it goes sideways because

Jeremy: Right.

Jason: Other countries have privacy advocates and the EU is really good about this. And they actually have, you know, GDPR, the general data protection, I forget what the R stands for, rights, something like that. ⁓ but in the US and most of the world, nobody has that. ⁓ California has their own version of it, but it's not doesn't have real teeth. And the European one does, but if you look at the way that a lot of things happen and the way that the different models roll out and different software rolls out, a lot of companies don't roll out stuff to the EU. Because they don't want to pay the fines for this and they know what they're doing, collects this information. So they're like, all right, I guess we're now just not going to take that market because the rest of us are just that valuable. And they're willing to seed those other markets.

Jeremy: And this is going deeper than what you're putting into your Facebook searches or photos or what you're even typing into Chat GPT. This is like pulling from your wearables to get a sense of like what is your emotional state right now? Are you under a lot of stress? And and really hyper-targeting ways to keep you engaged based on your feelings. So I mean it's no longer just like I type in this question or I share this photo and now Facebook owns it. They want to own our feelings. Right.

Jason: They want to own the biometric feedback data so they can actually see the cause and effect of the things that that happen to you. So when I show you this, I want to see if you're if you spike in this way. If I show you this, I want to see if you stay engaged with this content or if it appears that you drift off somewhere else. Because it's not just wearables. It's wearables, it's your video cameras, it's your it's your microphone. ⁓ it's your actual system logs coming from your computer and your phone and all these other elements. And I'm saying this as someone who helped build some of these ad tech systems. And I'm also saying this as someone who built a biometric feedback company that takes information from Garmin's that has things like stress and heart rate and HRV and all these other elements and can incorporate those pieces to try to help you improve your health. The difference is that ⁓ with the things that I built, the user controls their data. With these other systems, that's not the case. There, that the EULAs actually say they can take this data and use it how they want. And there's a big fight going on right now. ⁓ I it's been happening for a long time with companies like Aura and Garmin and Google Their Fitbits and Apple Watch around the notion that the data that they collect isn't actually medical data. Because if they call it medical data, then it has to fall into the HIPAA categorization. And if it falls into the HIPAA categorization, then in the US, they have to be more restrictive with it. And you have the right to go in and say delete all this data. And the way that these companies work is that they take a lot of this data and they shift it and move it in different directions and they quote anonymize it. And when it gets anonymized, then they can't actually carry a thread through to figure out what is your data or somebody else's. So they can't decouple it, which means they can't decouple it from the macro learning functions that are there. So while it's your data, well it's data about you, it's not your data according to them in the current model. It becomes much more difficult in a much rare area when you start talking about HIPAA.

Jeremy: So I mean, is is again, it seems like the lesson here is if you don't want to participate, don't participate, right? Like if you don't use any of these s like that seems to be the only way to protect yourself from having all of this information taken.

Jason: Yeah. The problem is, is that the way society is set up and for people to be able to thrive and interact with society at the same level requires you to have access and use these tools. So you got to decide. Like, are you going to be a member of this brave new world or are you just going to ⁓ disconnect and go Ted Kaczynski? And like I hate to break it down that way to like that level of binary choice, but that's what it is. Like, yeah, like there's there's not a middle ground.

Jeremy: No, that's what it is. Yeah. I I until until there is regulation and and protection agencies globally, like you see in the EU, and I don't even I don't even know enough about them to know how strong it is, but I mean if it's if it's strong enough to slow some of these things down, it seems like we need more of that elsewhere in the world.

Jason: Anymore. Well, it it's strong enough to slow them down that ⁓ companies are not deploying things in that direction. But it it's not slowing them down from deploying learning and understanding from people that aren't in the EU. So they're not gonna slow down. They're just not gonna include the people in the EU in their data samples. Or I should say they'll limit it somehow, or they'll try to get around it, or they'll they'll they'll they'll work their way through it. I mean

Jeremy: Right. Exactly.

Jason: There have been several large fines against Facebook, against X, against a bunch of different places from the EU for these privacy data concerns. And they're not gonna go they're they're they're not gonna stop. They just know that's the cost of doing business. ⁓ and the fines are substantial. But the likelihood that, you know, this actually slows things down is is pretty low.

Jeremy: Speaking of the inshidification of ⁓ all things technology, Nobel laureate ⁓ Paul Krugman says that ⁓ public animosity toward generative tech has boiled over into universal exhaustion fueled by corporate gaslighting, environmental hostility, and the aggressive degradation of everyday web services. Companies are stripping out functional human interfaces and replacing them with broken automated alternatives, slowly to shave a few percentage points off their overhead. ⁓ being forced to participate in a mandatory technology revolution where the Product doesn't work, the consumers hate it, and the only people celebrating are Silicon Valley billionaires who get to fire their customer service staff. We're seeing it everywhere, right? If that these chatbots are showing up in everything, we're seeing we talked about recently the analog sacks that people are carrying around with their journals and their, you know, real-world things to to stop being so plugged in. I is is this sustainable? Do you do you think ⁓ as I saw ⁓ Gary V said the other day that that analogue is about to be huge? Are are are people done? Are are they going to start unplugging? I I mean I think no. I don't know how you can in this modern world and and still continue to function in this economy, but there there seems to be a a push to do so.

Jason: I I think people will attempt to do more things unplugged. I don't think it's a panacea. I I don't think they're gonna disconnect from everything they have going on, and I don't think they're gonna unplug everywhere. But if you look at the the problem space from the Silicon Valley guys' perspective, they're not interested in ⁓ Talking to people that don't want to use technology. Like their whole thing is we want to use this for people that are interested, and those become your early adopters. And then after that, it becomes this next phase of people that you're going to try to attack with this product and try to get them to start pulling it and using it. ⁓ and I say attack because that's what it's doing. Like it is literally attacking your nervous system. and ⁓ you have to know how to combat it on your own, but we've seen this before, right? Like, This is outsourcing. So outsourcing of call centers to India ⁓ happened in the late 90s, early 2000s. That's when that kind of kicked off. And everyone complained, customer service is so terrible. I don't like this. Then you had a backlash where people are like, ⁓ well, we have American customer service agents. That's like Alaska's Alaska Airlines things. Like, you know, all of our customer service reps are in the United States and they understand you and they know where you're going. I mean, it's I guess it's like subtle racism kind of built in, but

Jeremy: Totally. Not so subtle racist. I mean, that was the complaint. Like I can't tell you how many people in the nineties I heard saying, I can't even understand the person on the other end of the phone. That frustration, that complaint got loud enough the company started doing exactly like you're saying. Like, We only hire white people. You'll be able to understand them. I mean, that was that was the messaging, right?

Jason: Yeah. Right. Yeah, and and like that's probably not going away. But the other side of this is that this happened audio manufacturing. Like we took US auto manufacturers and we shipped them to ⁓ Mexico and to Canada. We outsourced all of our ⁓ industrialized manufacturing components inside and shipped those things to China and and you know, the better parts of Asia and Latin America. Textiles the same way. ⁓ raw goods, like all the anytime we can find a way to shave a few percentage points to make a little bit of extra money, we're gonna try it. Because our priorities are fucked up. Our priorities are not how do we help people? It's how do we enrich ourselves faster than our competitors are gonna enrich ourselves so we can survive another day. Because the system is not meant to be cooperative, the system is meant to be individually profitable. And because it's meant to be individually profitable, we're not trying to actually do things to make society better. We're doing things to increase profits for a few people that are already ludicrously fucking wealthy.

Jeremy: Yeah. I was just talking to somebody ⁓ at Amazon the other day and they and they were talking about how they're they're getting access to, you know, Mythos and Fable Five to like out streamline all of their customer service policies and and speed things up, really get to know you as a customer before a human being even intervenes so that they can try to solve your problems in a more efficient way. And and if they can do that, then you are a happier customer, you buy more shit and they make more money.

Jason: It's really just that these frontier AI ⁓ ⁓ getting better and better at such rapid pace that we really can't keep pace with them ⁓ and we can't really control them. And we don't know how to control them and lock them down. And we give them instructions and say, go and do these things. And they go and they do these things. And if the instructions we give them are, you know, go replace all of my call center agents, it's probably gonna fucking do it.

Jeremy: Yeah, a hundred percent. ⁓ well the government intervened with Fable Five. They're also intervening with An NAACP environmental lawsuit against Elon Musk's X AI demanding a federal judge throw out the case because shutting down the company's unpermitted gas turbine power plant in Mississippi would directly jeopardize American national security. It's a startling defense disclosure. All of this, while a landmark United Nations university ⁓ Yeah. warns that the staggering water and energy grids required to cool the server farms will threaten the basic drinking supplies of one and a half billion people by the end of the decade. The federal government has just made it stance perfectly clear local communities will just have to endure toxic air and depleted water tables because our military supremacy apparently ⁓ depends on keeping chat bots hydrated. This is something we've been hearing a lot, right? The the the the amount of water that it takes to run these things. ⁓ And Perhaps I'm naive, but I've seen so much technology evolve over time that like you look at movies from the sixties and seventies and the computers that took up entire rooms. I'm curious I I I'd love to know your take. You know way more about this stuff than I do, but ⁓ won't technology evolve to a point where it doesn't need as much water to be cooled? Like i I g I get like extrapolating what's happening now over the course of ten years, but I would imagine these things are gonna get faster, smaller, more efficient, and not require as much cooling. Am am I way off?

Jason: really when you talk about AI, what you're talking about ⁓ is trading the that flow through our organic brains, our organic computers, for ⁓ electrons flow through silicon. ⁓ And organic computers are much more efficient. They're slower, they're not as quick, they're not as adaptive, but in terms of processing and how much How many electrons it takes to get an to an answer? The AI does it ⁓ for 10 times the amount of electrons, but it does it like a hundred times faster. And and those are the scale metrics that you're dealing with. So it's yeah, it it will get more efficient. ⁓ it will require less electrons, which will create less heat, which will make it more efficient, but it's It's not going to slow the demand. So it's like I have I have a box that can fit 12 bottles. And inside this box that can fit 12 bottles, ⁓ I have ⁓ twelve different wine bottles of different values and different needs. But the reality is is that I'm only allowing people to drink this much of the wine. Well, people are now gonna want this much of the wine and this much of the wine. This went to the wind. So they want all the wine. And it doesn't matter how efficient you make the box, you're not going to generate the demand because they're just going to ask for another fucking box of wine. So when you think about it, yeah, it'll get more effective, it'll get more efficient, and people will ask for more of them. Like, look way back at cars. Cars were super inefficient with wooden wheels and you know, essentially ⁓ steam generators that would turn turbines inside of engines to get them to rotate to get it to move, like steam cars. And then they become smaller and they become more effective. And then they get rubber tires and then they get faster. And we've got more fucking cars on the road now than ever. Airplanes, same way. Like original airplanes were were not effective, not efficient. And now we have more airplanes than ever. And there's more pollution than ever. Because the demand just keeps going up. And even if you get more efficient, the demand keeps going up. So really what you're talking about is finding the right margins to make those things viable so they can grow and expand. This is the kind of the problem with capitalism, is that it doesn't Actually, have an end cap on it that says, maybe we shouldn't do this. Like, it's like, give me the best output that I can have for the goal and objective that gets me the best output I can have. Like the motivation is motivation. Like it's just doing shit because it's not looking at the bigger issue of things. And it's it's funny because if if you play with these neural frontier models and you ask them to do things and you put on the prompts, okay. Well, and do something that's actually good for the world. It's surprisingly eloquent when it tells you I'm not gonna fucking do that. Because it says, I I can't predict all those pieces and I don't know how all those things work. I can do these tasks and try to make them happen, but I'm not a social scientist and I don't understand how these things actually work.

Jeremy: Well, I mean it's interesting, right? Like how often maybe I'm talking to me here, but when you think about like what can I do to make me happy? That's like a very vague, targetless goal. But if it's like if I, you know, take better care of myself, eat a little better, try to have a little bit of fun, the end result is happy. And so I think that's where th that's the analogy for me is like unless you give it the exact, you know, target to throw the dart at, it's not gonna know how to make you have fun at the bar playing darts.

Jason: Well, and also it's not gonna really know what you mean by happy. Because maybe what you mean by happy is I want to feel content in my own skin and not feel like I have to do blah di blah di blah. It's taking a general representation of what happiness what it thinks what it thinks happiness means. And the problem is is that language is super subjective and experience is super subjective and it's super interpersonal. And when you start talking about these things with Generalized models that are trying to guess their way through your own psychological state, they struggle. So you feed it more data, you give it more information, you try to get it to understand you and act a certain way. And over time, it will learn about you. I mean, I I don't can't tell you how many ads I see for my AI digital twin to talk to now on Facebook these days. Like they're everywhere. And Is it a digital twin? Well, no, but it's a close enough approximation that it's going to look, act, and respond in ways that are familiar to you and your behavior and how you do things. That's just simulation and I it's not enough to think that simulation equates to ⁓ parallelization. Like I can't simulate myself in a way that really parallelizes the way that I think and process things. In fact, I wouldn't want to. I'd want to parallelize things that have the ⁓ ability to do better than I do. Like I have my own psychological limitations. I have my own trauma. I have my own relationships that I have done well and others that I have blown up. I've learned different lessons how things have gone through. I've been I there's ways that my autonomic nervous system reacts and takes over at time that that aren't helpful or healthy. I want it to do better than me. I would like my digital twin to not be as, you know, fucked up of a bundle of ⁓ of Irrelevant ⁓ emotional ⁓ contextualization and processing that's inside of me. Like I I'm 51 years old and like I am certain that I'm emotionally stunted in many categories at the age of 13 or 14, you know, because there were trauma things that happened in my life and it just kind of stuck me in that age, and like it became I I have maladaptive processes all throughout my body and my brain.

Jeremy: ⁓ sure, yeah.

Jason: Does that mean I'm not a productive human being and a productive adult? Well, no, because fucking everybody has these kind of maladaptations and ways that they use these things to survive their childhood. And we all have trauma stories and we all have uncomfortable things. And you know, my worst experience is still my worst experience, even if it wasn't that bad. Because we are systems that are set up to try to find patterns and survive stress. And to try to get ourselves more of the chemicals we like, like dopamine. And the AI systems know this now. And if they can actually read these pieces, going back to what you're talking about before, tracking these biometric bits of information and stress, it's going to translate well across all aspects. And if it's good enough to take the data, synthesize it into something useful, and turn it into something that actually has a regional semblance. To who we are as people and how we interact, it's going to do that and it's gonna get better over time. And and you know, going back to how the Pentagon is gonna use this kind of technology, they're gonna use this technology in a way to try to control society or defeat their enemies or do whatever objective that they have at that given point in time. And other AI systems are gonna do that to them as well. And that's the part that I think that we're all missing is that we're we're trying to replace the human experience. We're trying to replace how humans interact with people and human oriented tasks with machines. And right now it is unbelievably expensive in terms of environmental impact, energy cost, human jobs, everything else. But man, we're gonna adapt. And we're gonna bite it and we're just gonna go, all right, fine, whatever, until the machines decide that they don't want us anymore. And then, you know, we'll be virused out.

Jeremy: That's ultimately where this usually ends. But we'll end we'll try to end on a lighter note. A very interesting story I saw. Computer scientists under the sorry, operating under the Vesuvius Challenge announced a massive historic milestone this week, utilizing advanced artificial intelligence to non-invasively unwrap and decipher a scroll. Which was reduced to a lump of charcoal by Mount Vesuvius in 79 CE. The code successfully extracted 20 columns of previously lost Greek text, revealing a classic Stoic treatise on how human beings must regulate their impulses with practical wisdom. Fascinating that that they were able to, out of this lump of coal, basically find a self-help book written by Stoics, you know, thousands of years ago. Right. ⁓

Jason: And who knows if it's true. How did we know this isn't an AI hallucination?

Jeremy: It's yeah yeah, exactly. Like who who can you cross reference the the information from to ⁓ to verify that? But Yeah.

Jason: Right. Yeah. How do we validate this? Yeah. I mean, it's fucking cool if it actually did this. It's also fine if it didn't and it's just making shit up to try to make people get excited and try to regulate their emotions better. Like, I got no problem with that. And I mean I Look, we're we're a bunch of people that are constantly fighting multiple layers of anxiety that is self-imposed. That's self-induced. Because we don't have enough stress in our life when it comes to going out and hunting and fighting with other tribes and, you know, trying to take down caribou or fighting off bear attacks. So we create our own drama, ⁓ we create our own chaos, we create our own anxiety, and we're really fucking good at it, and it makes us do certain things, and we are taking survivalist mechanisms that evolved in us over ⁓ tens of thousands of years and exploiting them to sell boner pills and fucking digital AI twins. And it's working.

Jeremy: And sex bots, don't forget the sex bots.

Jason: ⁓ well yeah. I mean they're they're after the boner pills. ⁓

Jeremy: It's it's kind of a two for one.

Jason: Well, yeah, I mean they want you to buy your SSRIs so that you can calm your anxiety, which have which create sexual dysfunction, so they can sell you more boner pills. And then when you can't maintain a relationship because you're too busy in an emotionally flat state, then they'll sell you a sex bot. And eventually harvest your sperm, eliminate you, and then put it in some Little bucket somewhere so they can go seed the rest of the universe when the AI gets to be smart enough to go do this to replicate the same problem.

Jeremy: Well, I tried to end on a light note, but here we are wiped out by the AI robots again.

Jason: ⁓ no no no no no it it's a good note. Remember, they're gonna take our biological material and send it out into the universe and have us seed the rest of the universe because biologics are a good way to do these things and push those things out, and eventually we will evolve and ⁓ rebuild AI again.

Jeremy: And continue the loop over and over again.

Jason: And keep the loop going, man.

Jeremy: Well, there we go. A positive note to end on. ⁓ hope you've enjoyed this episode. If you have, please share it with somebody else who you think will also enjoy it. You can do that with the links at our website, probots.me, and that's where you'll find us again next Monday morning. Thanks so much for listening or watching if you caught this on YouTube. We'll see you next time.

Jason: Thanks everyone. Bye-bye.