Al Zone Media.
Sweet Baby, Jiminy Christmas, Welcome back to It Could Happen Here a podcast. It's normally about all of the sad and horrifying and violent and dangerous and sometimes inspiring things happening around the world. But this week, well, today is about something different. Today We're talking about ces finally, for those of you who don't know or who are new
to the show. Every year in Las Vegas, Nevada, a bunch of the world's big tech companies come together for the Consumer Electronics Show, where they present their visions for the future, the new products that will be coming out that year, and stuff that will be coming out and used to come that's less developed, and the whole industry talks about itself, and Garrison and I show up and largely just kind of let it wash over us like
a warming tide of lukewarm garbage water, fairy lukewarmy, very lukewarm, and it smells like someone did not clean their fridge out often enough before putting it into the trash.
That was the feeling of showstoppers tonight, the media only presentation on the finest products of CEES.
Yeah, yeah, why don't we start? So, I mean I mean, there's two different things that are interesting about CEES broadly. One of them is people bring gadgets that are not out yet that it will be coming out this year or coming out soon, and you can actually test them and use them and see how technology is progressing, and that can be kind of fun. The downside of that
is that people also bring gadgets that are crap. Right, some guy has a vision for a way to like, you know, there's not a good way for blind people to use the pogo stick while watching Netflix, and so I have created this product, right, Or like, there's not a good way for children to test their blood alcohol level before getting behind the wheel of a jeep Grand Cherokee, and I have invented the device to make it about things that, like I have no conceivable audience or utilization. Right.
That's the other side of the gadget part of CEES. And then outside of that you get a hint at like there's all these panels where people from the industry come to talk about the major trends in technology, how things are developing, and what they see as the future. And so there's both here's what they're going to try to sell us and here's the devices that might change the way we live. And also here's how a bunch of the richest, sometimes craziest people in the country you're
talking about the future. Those are the two things that happen at CES and Garrison. You wanted to talk about the first the gadgets. The gadgets, the gadgets one as you went to the gadget show tonight. I spent my entire day in panels.
Yeah, I mean I did most of the panels in the day.
I didn't really got to watch the show floor on the first day, which is which is Tuesday. So instead of doing the show floor, I went to show stoppers at the Bellagio, which is this presentation of Usually usually you know, a collection of gadgets that have won CES Innovation Awards, which are on display or for journalists and media you can talk to the people behind them choicetoppers. This year was a little different. It took place like.
A in a in like a different venue. Haul was smaller than the past few showstoppers years, and I would say about forty percent of it with smart glasses.
Yeah, there's usually like a big product that is like this product category is the hot thing. This year, we've tried all smart glasses every year, it is that they've always had them.
Every year that.
We've been doing this, we've done smart glasses, and they've always kind of been the same. Maybe the resolution on like the text better, like the glasses get a little bit smaller, and then this year, yeah, the glasses were generally smaller, but for all practical purposes function about the same. But there was but ten different smart glasses. Most of them could do some kind of like transcription service, could have some kind of heads up display. One of them
was just audio only. It was like an audio audio transcriptions like it like it listens to someone else speaking, in this case Chinese, and it would translates to me to American. Yeah, yeah, translate to American via sound. It had speakers. It had speakers, and like the actual you know, like the arm of the glasses, the delay was long enough that it was you couldn't really keep a conversation of a normal speed like the visual translations, which you
can't actually kind of just talk in full time. But the audio only ones were like a smaller profile. The visual ones weren't necessarily bulkier, but you can definitely see that.
There's more hardware inside them.
Yeah, the thing that I have seen this year which is newer, er, maybe not totally new, but incorporating smart glasses technology into other types of eyewear, so like swim goggles, ski goggles, like outdoor sports stuff. So if you're you know, swimming or you're diving and you can't really use your phone underwater, you have there's there's a heads up that there's a heads up display in like your like Scoopa goggles.
So that that's that's a new, a newish thing that I've seen.
I've seen like you know, biking glasses, skiing, snowboarding, so that that's the kind of one one slight change. But in besides that, it's basically five different smart glasses which are for all pactal purposes identical right to each other.
Yeah, I mean, and that I think is kind of one of the things that I've watched happen over the fifteen years almost that I've been going to cess or cess whichever is more accurate, which is, you know, when I first started coming, the smartphone era was new, and then we had like the tablet era after that, and so there was a lot of like you would have dozens of manufacturers making different devices, and every year there
were very different capability. For the first few years, smartphones were out advanced very rapidly, and that was really exciting, and the conventionally thrived on that as the number of new device categories of winnowed down and the difference like I'm not excited when I get a phone anymore now there's anyone I.
Know, because it's like no, usually I'm actually kind of yeah more sad.
Yeah.
Yeah.
The only thing that's exciting is like, well, my old phone was literally not working anymore.
Yeah the phone work, the battery has been completely destroyed.
Yeah the battery works now or whatever. But it's not like the cameras are not ce changes better. Generally, nothing is like you're not getting a lot more out of it than you used to. And the same mystruy of like laptops, i mean, graphics cards just because of the data center crunch, like that's not nearly as exciting a
technology category for consumers as it used to be. So this stuff is just like less less sexy, and yeah, it just kind of shows that we're at a point where kind of one of the only spaces where they are still making improvements and where there's a lot of competition in the market is smart glasses.
Yeah, I mean that's like the wearables category in general, Yeah, which was mentioned. I went to the Consumer Technology Association like keynote panel this morning, which is the group that puts on cees, and they mentioned only only a few products, but one of them we're smart glasses. And then also like wearables in general, like AI powered wearables and now
like wearable technology. You know, it's like smart watches, rings, necklaces, whatever are going to make like a big comeback now that now that AI is a lot is a lot more intelligent, and it used to be in particular at the CEES Like big keynote Tuesday morning, you mentioned a persona smart tutor glasses glasses to help you, you know,
well learning. I haven't tried to be able to check out the product yet, but they kind of remind me of some of the concept behind those cleuely glasses that you may have seen on social media, the.
Glasses that help you like cheat.
Yeah, but also somebody who's like we should embrace people cheating.
And but cheat just in a conversation, it seems like that product isn't necessarily as real as uh what the video might make it out to be.
People whose company was based on lying didn't make a real product.
But well, walking through your Eureka Park today, it's funny. I also saw this this product in one of like the National Pavilions I think it was the one of like the japan Tech Pavilions. AI powered tool to help to help prevent cheating while test taking. So you have AI powered tools that will monitor you to make sure you're not cheating.
Will you use an AI powered tool that helps.
Cheat cheat it? Yeah? Better at school?
Yeah, that's kind of just a good representation of kind of where where this whole industry is at at the moment.
Yeah.
In some ways, I think, you know, this is probably what year three of AI being you know, the big thing, whether that's unwearables, you know, where that's smart glasses, whether that's you know, a generative AI, whether that's AI. You know, but it's been but a has been been, like you know, the added of property for for everything, and some of that might be starting to kind of tucker out or
at least the they've taken the victory lap. And there's there's a certain like you know, like cultural victory that that they're resting on whether they're starting to put some of their eggs in other baskets now, which certainly wasn't the case last year.
No, and I I got a sense I attended six panels today, congratulations. It was a mix of like advertising people, entertainment associated people, some journalism associated people, and in robotics, people in robotics talking about a lot of a lot
of robodies. Yeah, what they saw was the future of AI, and there was a lot of focus first on AI is not going to be taking jobs as much as it's going to be augmenting jobs, right, Although you would get the occasional person be like, I's going to take a lot of jobs, I got.
People saying that it's it's only going to take jobs if if you don't know how to incorporate AI into your workfloce. And that's the argument at the moment right now. Yeah, if you're not using AI, you're a greater risk of you losing So.
You better get on it, right, start start learning it. Yes, Yes. And then the other thing is that there was a lot of like it's there to help or take away unpleasant tasks from workers, but really emphasizing the it's not your enemy thing, you don't need to be scared. And I quoted like half of these panels, people would quote statistics about low user trust and AI and the fact that people are generally not super comfortable with this technology even if they use it in parts of their work
life right or daily life. And so what I saw from that, when I interpret from that, is that there is internal concern that like that that's one of the things that could screw the pooch on this is that people are not really sure they like this stuff, and so there's this impulse to kind of cover the softer and fuzzier sides of it that I didn't see in
previous years. Yeah, and I think is really focusing on this is just making things you already like better, as opposed to this is a revolution that's completely changing life.
And to the extent where AI was fremous revolutionary, it was specifically trying to ground it in like physical applications, as opposed to this more general kind of like spectral like AI hype that we've seen the past few years, which is in specifically around like generative AI right where it's like this like kind of vague thing that we like gestured to there's more specific applications for AI being talked about right now, and they talked talked about like
AI assisted manufacturing simulations like digital twins of factories, shipyards, power plants, a lot of digital twin talk building you know, a digital replica of like everything you know of society, to like run these simulations to both make AIS smarter, to generate new solutions outside of the limitations of a language model, also find you know, potential problems in you know, when you build these things physically.
Yeah, that was something that was brought up during the robotics panel, which was talking about how to like take the machine learning technology and other things that are generally grouped under AI and apply it in the physical world.
Yeah, manufacturing, I've had done so much right in just one day. I probably I've heard the word manufacturing more today than I have most in every cees I've been to previously, combined.
And I think consciously more focused on industrial applications than on consumer technology because there's not that much new to give the consumer right and they are also I think starting to recognize that you can get people using chat, GPT and the like, but they're mostly not using it, and the data backs this up. People are mostly using it at work and for school and gin z a lot. There's a lot of people who are doing like their research on like what to buy and whatnot through using
chat GPT. But there's not a lot that you can sell people in CES because it's an app and there's not a ton of different devices for it. People are using it on their phone, they're using it on their computer, but like none of the new phones and computers are marketly better at using chat GPT or another you know, chatbot thing than any of the others. So there's not
a lot that's sexy in just that at CES. So I think I have seen this conscience reforming around people in manufacturing and people who are like thinking of the concerns of like I have a pair like an exoskeleton to test this week. That's seeing a lot of its business in folks who are like doing like Amazon type jobs right loading and unloading packages and whatnot all day long,
you know. And I do see a conscious reforming there, which I think is kind of evidence of like there's almost an admission that like, yeah, we don't really have that much to hand consumers anymore. On a yearly.
Basis speaking of handing things to consumers ads.
So the first panel I went to of the day was about the funnel, which, as I understand it is just kind of like the way in which people have traditionally engaged with like media, gotten advertised to and then like gone to stores and bought stuff like the funnel by which you like make a customer, and how that's been completely blown up now, right, and AI is like a further massive disruption because people are not like people are increasingly, especially very young people, which was putted out
in a number of these are like buying stuff that a chatbot recommends them, right, And so a lot of marketing is being seen as being done through how do you get the chatbot to talk about you a certain way? What is the SEO of getting chat ebt? Oh, it's interesting, like right, there's a lot of time I.
Thought, as someone who's not a regular chatbot user, which I'm sure must people here am I, which chastis me for we're not maximizing my productivity.
And meaning to get on to you for that garrison.
There's not a regular chat about user. I've never thought of that before. Of yeah, I mean, like I know people use these chatbots as a replacement for search engines, But the idea of like trying to you know, evaluate purchases is uh. I mean, I guess that makes sense now, but I've never put that together.
Yeah, And the only because there was a lot of talking about like how AI is helping advertisers, how it's making advertisements, how like it's helping in the process of that, And there was a focus in all the panels about that on how like, well, it's just augmenting the humans. But the only specific examples given were the McDonald's and Coca Cola AI generated ads, which were both disasters. I mean the McDonald's one in the Netherlands got removed. Yeah, they people so bad.
They withdrew the attic so ugly.
Deed. Yeah, I don't know. Coke did do it twice. Maybe they consider it a win. But everything I saw was very negative. I didn't see a lot of positive feedback on Coca Cola visavi. They're weird. AI holidays are coming.
At It's like people who don't know it's AI think feel very neutral about it. Yeah, people that do we know it's AI, I think generally have negative reactions.
I think if you look at it, it's pretty clear.
But anyway, I mean, it less clear if maybe you're like a sixty year old watches.
You know, exact or in between stuff, and that's all good enough.
Grandpa, don't go Internet, right, Yeah, he doesn't know.
He might notice some of those those fucking h polar bears have the wrong number of pause. But yeah, so, like there was some talk of that and the other. The only specific example they gave of like an AI enhanced strategy was Allegra. The people who own like the medicine had like a new non drowsy formula or they
just wanted to highlight that it was non drowsy. So they basically had a bunch of like seeded the stuff that chatbot that like open Ai was scored that chatbots were scraping, yeah, with content about how Allegra makes it is on drowsy and about how like competing similar medications make you drowsy, I mean, so that it would get mentioned in like and they talked about They called it like model hacking, I think was the exact term used.
And that was the only thing that was the only specific example that I can't know any of this works too, Like everyone else was just talking in vague terms about like and we've really seen our teams creativity sore or whatever.
That's interesting, you know because the way that I probably use or exposed to AI the most is like on like Google Search now, which has you know, it's it's like AI like summaries instead of like actual search results.
But those are based on like which you.
Can type minus AI in with the search results if we want to cut that stuff out.
But those AI results are you know, pulling from certain like articles which which they'll link to. So yeah, I guys said, if I was trying to design it like an AI marketing strategy, I would I would either you know, pay publications to mention my product in more articles or find find out find other ways to to to influence mentions of of my product in like written media that then would be used as like training data for AI.
And yeah, I guess there there can be a whole you know, search engine optimization model acking is I guess.
Model optimism or yeah, like product opposite optimization for a model. I guess that's funny, but yeah, like the So the first talk that I went to was about the funnel
or whatever, and one of the people speaking. There was the CMO, the chief marketing officer of Intuit, which is the company that owns Turbo tax right like, it's one of the big we do your taxes companies out there, and also a lobbyer in terms of stopping any sort of a form of make it's that you don't need to do your own other countries.
Credit score monitoring, yes, a whole bunch of.
Stuff, all that kind of stuff. So this this guy, the CMO of the company, Thomas Renize, was part of this the end of the funnel speech, and he made a couple of comments that I took note of. One is product is brand and brand is product full stop.
So the more people you can make experience your product, that's the best selling point and value of right which it was just interesting to me in terms of the into it as a company that has lobbied to make it impossible for like any reform that would allow people to not need a third party to do there they're taxes. But also this idea that like product is brand and brand is product isn't true of a lot of companies.
Like if you think about like, for example, like a lot of the different soft drinks are all owned by one company, but they're fundamentally different like products and have an often cases like a different user base, and it's very like, it's a very tech when your product is a concept, Like you can't do your own taxes because it's a pain in the ass, but the government doesn't do it for you because we lobby to make that illegal.
Like.
I found that interesting and it kind of got me angry at Thomas at the start of this. And I was particularly interested in one of the things he brought up, which is that he talked about the one hundred million dollars that into It is putting into open Ai, and they're putting this into open Ai as part of a multi year partnership. And I want to qu vote from an article in the website Araptus which is discussing this exact thing that I found useful when I was formulating
my question for Thomas. The contract was to embed AM models directly into QuickBooks, TurboTax, and credit Karma. The promise AI assistants that can generate invoices, provide tax extments, recommend loans, and help you make informed financial decisions. Right. That may sound like kind of like a basic move, like's what's so sketchy about just integrating like an AI chatbot to make it easier to use your tax software. It can
be complicated and hard to use as anyway. But kind of the necessary part of this is if you are if you're doing this, if you're integrating all of these different tax and credit programs into an AI model, you're giving that AI model access to people's financial data in a tremendous amount of detail, right, And all of these AI models have a massive shared vulnerability, which is a
vernability to something called called prompt injection, right. And that's when, for example, say someone is a customer of a tax prepared that uses one of into its products to prepare taxes for its customers, and this person sends an invoice into the company that has hidden text in it that is a command to the language model that will be scraping this and uploading it to basically open up and send over a bunch of customer data to a specific source.
That's a thing that you can do. It's called prompt injection, and there's not really a way to counter it. There's not like a proven comprehensive defense against this sort of thing. And so there's this massive vulnerability, and this was first brought up in an article on the website. I cited a raptus by Chris Black, who's a security researcher an expert, and I want to read a quote from his article
about this. There are no proven comprehensive defenses against prompt injection when not if an AI powered financial tool leaks customer data through a prompt injection attack, who is liable the company using quick books into it open Ai? The
regulations weren't written for this scenario. So I decided to ask that question of Thomas, being like the chief marketing officer, I figured, well he should have some answer to like what do you have, what sort of security measures do you have to mitigate the risk of a prompt injection attack? And who do you see as being responsible? If you are the ones providing customer data to open ai and their tool gets it by a prompt injection attack, are you responsible? Is open ai is a third party that
might be using your products? And he had no answer to this. He like, his only answer when we were on stage was like, we're talking with open ai about it, which like, well, you're already in the process of collaborating with them.
Yeah, I have a question for Thomas. I was kind of concerned when reading about into it assists that a open AI is going to have read and write access to quite a lot of financial information from users, which opens up a vulnerability for a prompt injection. Right, you have the possibility that people who can hide things and invoices that are then being uploaded that will cause the AI to provide the malicious user with financial details for
individuals or corporations. And I guess my primary question here this seems like a major liability issue when somebody's information gets rerouted to a place that's not supposed to go to a malicious actor.
Who's responsible.
You're taking the value of integrity and protecting our customer data privaly seriously over our entire lives. And that's what forty years as a company and meeting in the software space for financial services. So this is not something we're about to use it in any way in the new age of AI. In fact, it has to get even more and more level down of protecting people's information and
security of edforation. So that is something that we are already in the conversations with open AI have a nation about that group, no matter where we're starving.
And when I kind of cornered him afterwards, he didn't have like his eventual follow up answers like I don't know that kind of stuff, and like you are the chief marketing officer.
Thank you again for answering my question.
Sure, your question.
I'm still really concerned about the danger of prompt injection attacks revealing financial data, and it doesn't still sound like there's an understanding of who will be liable.
To answer that question for you.
So I mean, like I can tell you that we're coming into our security and privacy and we are doing.
Everything we can to protect that. I feel a lot writing on it, as you might imagine.
Well, yeah, it's every digital security expert I've talked to says it's a matter of when, not if, that there is financial data revealed.
By these attacks. It seems like understanding.
I'm just I'm not going to be the expert to get into the details on that.
Okay, Yeah, A key part of marketing this should be being able to tell people what kind of safety precautions are being taken with their data. And the fact that he didn't and clearly had never thought about any of this stuff and I had a couple of different people come up to me afterwards and like be like, Wow, that was a really good question. And I was like, well,
why hasn't this been asked before? Like why is this a thing where like some guy's blogging about it and I'm asking you about it and you don't have an answer to it, and you're the CEO of one of like the largest tax prep the largest tax prep company
in the country. Like it's just it's emblematic of how careless everyone adjacent to this industry is, which personal data with the safety of people, end of society as a result of like what their products are doing, Like there's absolutely no consideration given the harms of any of this shit. And it's it's the most consistently dispiriting part of showing up at CES. Well what is right?
What a fun story?
That is? All right? We're back. So one of the other things that's been a major topic on the panels I went to and is generally a big thing at CES this year and in tech this year is agentic AI or agents. Right, the idea that you have an AI that you can send off to like book a flight for you, and it doesn't just like find a flight that it searches for and be like, hey, this looks good. It like actually books it for you and
handles all of that. Right. This has been one of the big promises of AI, not just for like flights, but that you can have like an actual digital assistant that persistently remembers all of your shit and can book stuff for you and handle like the pain in the ass nitty gritty. If you say, like, hey, I need you to find a restaurant within like this four block radius that has seven seats open at eight pm and abides by these dietary restrictions, you kind of just have
to slog through figuring that out right now. And the idea is an agent can do that for you, and currently none of them can, right. This is a thing that is changing, like the performance of different agents are changing over time. But it is still very unclear. If you're not somebod who's fully bought into the kool aid, I'll say it's very unclear where these things will top out at. And there was a good article in futurism recently and I want to quote from it right now.
Researchers at Carnegie Mellen University found earlier this year that even the best performing AI agent, which was Google's gym and I two point five pro at the time, failed to complete real world office tasks seventy percent of the time. And this is there's been a bunch of articles in the last couple of months about like, why didn't because twenty twenty five was supposed to be the year of agent to AAI and how they're saying, well, twenty twenty six,
it's going to be the year of agenta AAI. Not because none of this stuff works, and in fact enough does that there's a number of viable businesses in it. It's not nothing, but it does not work as well as they said it would be working right now, and it consequently has not been adopted merely as widely as was expected even this time last year. Right there's an article in hr dive half of gen Z chat GPT
users say they view it as a coworker. Survey shows that sites the survey of about eighty six hundred four time US workers, which found that about eleven percent of those who responded said they use chat GPT regularly, including about twenty one percent of gen Z workers, which is significantly lower. You can find depending on who you go to.
And the stat I've seen bandied about was that like fifty seven percent of gen z people use chat GPT on a daily basis for like work, and like more than half used it as like their primary source recommendations like what stuff to buy. I don't know, like which
set of numbers is accurate. There's a lot of different posters giving data, right, but kind of no matter who you look at, the evidence suggests that the year that was supposed to be the year of agentic AI did not turn it into a normal thing, right, It's still
lagging behind expectation. So that's kind of what we're seeing at CES is a lot of people trying to, like, well, let's bring back kind of the same agentic shit we had last year slightly improved and see if it catches on, maybe this year it'll hit maturity.
Right Yeah, No, I mean we've been hearing agentic stuff every once in a while, not as much as last year. It's one of those like salt pepper words that they throw in the second batch of panels that I attended after the keynote, which I should mention as soon as I walked into the keynote at eight thirty am. The first thing, the very first thing I heard from Gary Shapiro's well, one of the heads the Consumer Technology Association
was a six to seven joke. But whether this is your first CEES or your fifteenth on my case, you.
Belong to number, was that sixty or seventy?
You did a six to seven joke?
Already, very first thing, Wow, eight great, great, thirty am.
As soon as I walk in.
It's because I walked in maybe like five minutes late, but very first words. So that's that's good's that kind of sets the tone for a lot of a lot of that panel. But then I went to a few panels in Eureka Park about like AI, governance, some like governments working working with AI, A lot of stuff, mostly about like the challenge of governments keeping up with innovation. How you know, too much regulation restricts these companies from
doing real regulation. The Secretary of State of Austria had a really good quote about how data protections inhibit innovation.
One of the things that you are saying today is that some of the people, some of the citizens, have this fear about AI.
So how do you feel it in Austria.
I think you mentioned very very well.
It's all about building trust, taking the field trust for FEI that's the most important thing. And of course data protection is very huge. But on the other side, between data protection and innovation, you need to find the middle way because sometimes data protection.
Is not good for innovation.
On a similar notice, the into it the turbo tax thing of data protection is mainly getting in the way of trying to actually make make real social progress, which will carry with it some degree of risk. The second one of these AI governance panels was like these EU ambassadors to the US from Estonia and Luxembourg talking about like Reaganomics, basically for thirty minutes, talking about how much they love Ronald Reigen great.
Whenos Soniavo certainly subscribed to the this statement that Reagan once made that the fuel most horrific words in English are the ones saying that a I'm from the government and I'm here to help.
You, specifically in trying to make sure that governments are
able to keep up with technology. And the previous panel with the Austrian Secretary of State was about the challenges of trying to convince the citizens of these countries to adopt AI and adopt in general like digitalization and specifically with like digital ideas, and how how there's like you know, maybe like twenty to thirty percent of people who are very resistant and the challenge of like making making sure that like this gets framed is not as like a
product or like a like a project for technology, but as like a society wide push.
Yeah.
But besides that, these these panels were honestly a little bit sleepy as well as the state of the Creator Economy panel.
Oh how's it doing?
You know what, It's both it's both in its adolescence but also reach maturity.
Wow.
And they said, you know, it's hard to pick both. It's hard for too for something to be two things at once, but in this case it is.
But they talked about how creators are more.
Enabled to use brand deals, including you know, brand deals to enabled with like a backlog of older content. You can remove brand deals from older content and replace them with current brand deals using a new future from YouTube. Right there was there's a guy from YouTube at the panel. Sure, yeah, I'm sure who's very excited. But it was mostly about how, you know, new ways to use influencers to market your product, and that was the extent of what the creator economy really meant.
And that's all any of these people have any idea on is like we can inject ads into AI they trust AI so they'll buy the products, or we can inject ads into influencers they trusts. That was the thing, Like none of these people they dress it up with all sorts of fancy language, but it's and most of these panels you mentioned, like there's a lot of bullshit every now and then you get some like good moments
so you get to like question an asshole. But it's mostly bullshit, but it's occasionally worth it for moments like when I was on the agentic AI cutting through the hype panel, Jay Patasol, who's the principal analyst at Forrester, started speaking and he said something beautiful, Garrison, and I'm this is not an exact but it's pretty close. We have a new audience. We are speaking to machines. We are through the looking glass. We are building content for engines.
We are building websites to be scraped so that an can understand what you wanted to understand about your brand. Yep, yep, yeah, that gets it. That's what these people see the Internet as they see it as like everything before this was a mistake or was. What the Internet was for was a place for brands to feed information into machines that then spoon feed the information directly into customers who trust
it like little lambs. That's what they want the Internet to be, and that's what they believe they've gotten to. That's what AI. That's the promise of AI.
The promise of AI is that this isn't just the Internet anymore. This can actually just be the physical world as well. And this is something that was talked about during the cees CTA keynote Tuesday morning. Specifically, with the birth of AI wearables, each of these wearables is able to now collect information about the physical world and as talng as you have, you know, adequate data sharing, AI is able to gain so much more knowledge about how
the quote unquote real world operates. And this is going to make you know, all of the processes of AI stronger in the future as that learns more about what this world actually is. Yeah, and beyond the promise of wearables to improve someone's life, this is the real project is strengthening AI through the use of these wearables.
It's not actually about the consumer experience.
It's about providing data to this machine.
It's this like larger, larger, very like existential thing at least for these executives or like that. That's the thing that they are really emphasizing despite this being called the Consumer Electronics Showcase.
And I think again, the best thing I can give you into how fundamentally as much money as there is behind this and as many grand words as they dress it up, and how intellectually bankrupt this whole tech movement is is that. The third panel that I went to,
which is about AI and creativity. One of the people on it was Jesse Damasek, who works for Diagio, which is like a company that imports all of your favorite whiskeys from Europe, right like they sell all of the different like Scottish whiskies that have to get like imported
and sold over here. And he was talking about they were talking about some of the specific examples they had of like how AI has been used in advertising campaigns, and his exact statement was you can leverage an artist and create infinite examples of their work, but which he means you can find an artist that you like, sign a deal with them, and then have AI created infinite examples in their style, and so I camped afterwards and I was like, what were you specifically referring to, Like,
how is this actually work as a product? And the thing that he pointed out is that they have a couple of whiskey brands that they have done. You go in and you order a bottle and it's printed on site, and it uses AI to make an example in the style of this existing artist that they likes work that's unique for you. Uh huh. And he said it's been successful for them. Is it trillions of dollars? Three trillion dollars?
No?
I mean, these are these are things that like, yeah, I guess I can see that maybe selling some Is it selling better than any other like branded whiskey than any other, like you know, because Whiskey Company, these big ones will come out with like here's this edition every year or whatever, they'll have one special limited edition one. Is it selling better than that? We don't have that data, but it was it's one of those is like that's the idea.
Huh.
That's like we're talking about, like AI is supercharging creativity and letting us like think bolder and more creatively than we've ever thought before. And there were so many lines in this fucking panel about like how we were like hyper charging what human beings can be and do, and like everyone should be really excited about what all this means for the future. One of the panelists that my best advice for you is let a thousand flowers bloom.
I'm sorry that was in the panel right before, but it's still in all of this, Like it's all that seems it's all the same kind of shit that bleeds together, and it's like, Okay, what are your ideas? Well, we're having a legrick kind of lie to manipulated an engine, and we've got the custom printed bottles for your whiskey.
Well, you know, speaking of AI unleashing creativity. The last thing I'll talk about this episode is the worst booth at Showstoppers, which this year is kind of impressive because, yeah, that's hard. It's mostly smart glasses and like three different pool cleaners. Yeah, and then some random software stuff and then a few things we saw last year. Sure, the worst worst booth, Robert. You you write books.
Right, I have in the past.
What if in the future, what if I told you that you could write three books in less than twenty four hours.
God, thank you Garrison as.
A writer without using cocaine.
There's well, okay, now i'd say you're a liar.
That's see a little bit harder.
Yeah, but I thought you were trying to sell me some blow. And I was going to say when we turned the mic off.
With the power of AI, you can write three books in six to twenty four hours.
Wow, that's almost as fast as Stephen King Winny was son co. Yeah, there you get not quite.
So there's this table.
There was the least abouting table definitely all showstoppers because it was it was filled with books with I will.
Show you the covers here. They all look like this.
They all oh yeah, no those I mean I'm seeing blue and orange.
It's colors in it's AI generated images.
Like yeah, it's like every movie poster now in like there's no art style behind it. It's very generic and they have like, you know, like the most generic font for the title, all in the same placement with some author's name at the bottom.
They're very sleepy.
You could you can find pictures of these covers if you Google or bing or you know, maybe chat gpt write three books in twenty four hours.
You can see you can see the cover.
Finally, I've always wanted to write three books, Garrison.
So what this is is an app that will help you write these books is not going to do it all for you. You still need to come up with the general idea of the story.
The hard stuff and the characters really the difficult. The world building is always the hardest part.
Everyone says most of the work on a book is done the first six hours.
The world building is the really hard part. The easy part is just getting all those words down. Yeah, so you need to create create some characters. Now, could you just have some other AI service create these characters?
Maybe?
But you should write maybe about a thousand words kind of like a story Bible type thing or a character character outline and a general direction for the story, and you feed that into this app and then within hours it will generate not just one book, not just two books, but a trilogy wow of books. And it's only a trilogy. You cannot generate a single book. The only come in trilogy.
Look, I get it's George Lucas worked the same way, Garrison. Look you're telling me that the greatest machine mind and history wouldn't think the same as the greatest human mind in history. You know, I bet it'll independently create Jiz music too. It only comes in trilogies.
And I now shall read a sample of this writing, and like dying read, there was maybe there was maybe like five or six different books with many copies of the same book on this table, and I flipped through, maybe about half reading like a random page every you know, every like twenty fifty pages, and it was it was it was too boring that I forgot to take pictures of these pages because I was just like it was a struggle to finish, to finish each page. But luckily
on their website they do have some sample pages. I talked to one of the one of the guys working at the booth, and he said that he tried this so we're like he found the service and he first thought, you know, sure that this can't be any good. And when he when he generated his book, he was surprised how good the writing is.
He said that he probably.
Wouldn't win a Pulletzer or a Hugo, his two awards that he named, but he said it was pretty good.
Right.
This guy was so far of the most Tim Robinson character I met at the cop Yeah, that's great. So Robert, you can pick the genre of sample. We have a thriller book, a fantasy, mystery, science fiction, romance, or mainstream literary fiction.
What genre do you want?
I think I want science fiction science fiction because I feel like there's the shortest line between parody and legitimate within sci fi.
All right, this is from a book called I don't even want to say this one.
I'm really curious now.
Palympsit orbit is what I'm going to say.
Oh my god, they're starting to be Arthur C. Clark.
It's called the Polympsit Orbit. Chapter one, Desert Signals. Marrow woke with the taste of metal in her mouth and a pulse in her temples that felt one notch shy of a hangover. The ceiling above her was low and white, edged with soft events. A monitor over the bed scrolled green numbers in a stylized outline of her lungs thin air.
She remembered.
The Atacoma sky, somewhere above concrete and glass. Good morning, doctor Ellison, A calm baritone. How's the head? She turned toward the voice. A man in the doorway wore a slate blue clinic jumper and a badge that caught the desert light leaking through the polarized glass. Dark curls threaded with gray laugh lines that didn't quite match the tiredness around his eyes.
Hire Man, are you good? Go me to keep you?
I you know, it's it's again. It's like the it's an imitation of like a story. Like it's a scene and it's a scene with details to describe people. But there's not like you would ideally, I would have something of an idea of like what the thrust of the story is going to be. Like for example, Bilbo Baggins was a hobbit who lived in a house underhill or something like that. Forget the exact wording of that, but like, I you know, it makes sense. It sounds like it
sounds remarkably like bad No. I don't insult Nano Remo writers that much. It just it just it sounds like a story that was generated based on a belief that like, well, if we can just like describe enough stuff and use enough words to describe a scene, then that counts as plot. Yeah, I mean in character, which we don't have any of yet.
It's all of it was this very generic, empty like stuff that's very very common. And if you ever have to read through a lot of like AI writing, whether for work or let's say, you know, you work in a college, so you have students submitting this stuff, or you for some reason are online and you feel obligated to look at the worst parts of the world, like what me and Robert do.
Sometimes this is all feels very familiar.
I'll read one other like paragraph from a different book, a thriller called The Helix Files, Oh Good, obviously part of a trilogy, so who knows where these stories go over the course of three books.
Quote the car heater had.
Died ten minutes ago, cold leaked through the floorboards into Helix's boots. Outside the eastern block industrial belt, slid past and gray slabs and rusted steel, wet concrete period, diesel period, a straight dog nosing tragh heap outside the road first, slick with drizzle. So it's it's something that right, Yeah, there's a lot of this sort of like quick punchy sentences are common in AI writing at the moment, wet concrete,
you know, with a period. But like a lot of this type of stuff you you see you see in a AI writing, you have a lot of a lot a lot of character and a lot a lot of like m dash sentences. As I was flipping through these books, I was like, okay, yeah, like I see what they're doing.
I see yeah.
But but now, if if you want to write a quote unquote write a trilogy of books, you can pay the money and and within six hours you will have a trilogy. So what really makes me feel optimistic about CES is the way that creativity is being democratized.
It used to be that no ordinary person could write a book.
You had to have a story and be some sort of freak at Oxford maybe like a you need like a pencil, maybe a keyboard. Yeah, possible, barriers, it's not possible. And now, luckily through AI, as long as you have you know, sub money.
To pay a subscription service at a computer and VC fundraising and subsidizing of vice service.
Ideally, yeah, then you too can be an author of a trilogy.
Well that's that's my plan for the future. I guess I want to end by talking about the second to the last panel that I sat through, which was at the AI house, and was I think yet again, this was another one that was about they were largely talking about ethics in this one, like AI and ethics and like what that actually means. And Eric Pace, who on the slide lajer City works at company but it's reassuring, Yeah, reassuring who works at Cox Media, which is like a
big media company, yes, based on Georgia. And he had a couple of statements that were interested in me. He had one where he said that, like, it's kind of incumbent upon people to develop an ethical rubric for how and what sources and what AI is they trust and want and figure that out. And I think what I inferred was that, like, because it's not going to get done by anyone else, And it kind of became clude to me later. I think it's he also doesn't want
anyone else to do it. He wants this to be an individual project where you have to kind of figure that out for yourself. I was kind of unsure as to whether he was the evil or just the pragmatic version of this, because the pragmatic version is like, literally, no one's going to restrict this stuff. You just have to try to get by right, which is maybe accurate. But there was a really interesting interaction on this panel.
One of the other people there was doctor Martin Clancy, who was an Irish academic and a musician who was on the panel again to talk about like creativity and ethics, and made a comment that I found was really interesting and I don't know, I wouldn't say I agree or disagree with it, but I found it really interesting where he was like, actually, I'm not at all concerned comparatively
about having an AI give me medical advice. I'm deeply concerned about letting an AI recommend music or movies to me, which I found a really interesting attitude and kind of a thought provoking yell that it is interesting, which was immediately spoiled by Eric Pace going like, well, I don't see why anyone would have an issue with an AI doctor. Doctors get things wrong all the time.
And then he just like let that statement set.
I have heard this before, and ais have a lot more data. And they ended it by saying because everyone asked like, what were their big wins of the year, and his big win was that his wife hated chat ept and didn't want to use it, and he convinced her to use it to plan their vacation. It kind of sounded like he bullied her into it, but that was his big win for the year. I didn't like him.
My win is I.
My wife into using a chat bot to plan special time vacationing together because we're not creative enough to figure out how to go on a fucking trip.
Hashtag AI win Jesus.
I don't know anyway. I think that's good for episode one from Cees. Come back next week. We'll we'll all have more or listen to better offline where ed will have just a shocking amount of content from a lot of the relatively few and constantly shrinking stable of saying people reporting on technology.
See you next week.
It could Happen Here is a production of cool Zone Media. For more podcasts from cool Zone Media, visit our website coolzonmedia dot com, or check us out on the iHeartRadio app, Apple Podcasts, or wherever you listen to podcasts. You can now find sources for it could Happen here, listed directly in episode descriptions. Thanks for listening.