ワールド・ワイルド・ウェブと購買力を持つボットたち

Is there a conspiracy here? Are people like leaving holes so they can do the marketing? Spend the 20 bucks on OpenAI and computer use will set you free because I no longer have to worry about, oh, I'm not connected. I didn't set up the right connector with this email account because I'm already logged in on the browser to all the different email accounts, which I've got like 50 of. I think that's what they should write is that, do you really need Fable for this? Yeah. I want to talk about like the potential payment processing conundrum that I chatted with people about when I was in South Korea because I just got back from there and I met two payment processing companies they were both awesome they're both part of the agentic ai foundation

like members and so we can get into that if there's time after the whole nefariousness of models all right so erica kick it off what's the latest on the models so the latest is okay first of all the thing that i am really thinking about is if you say you have a dangerously capable model you you kind of want that at this point you want to be dangerous you want to be like a real cool ai cowboy right yeah cool cowboy right he's government don't let me release this to everyone because it's just so strong yeah it's so good it's so good we have to contain it okay this is the national threat it's a security threat we're really looking at right here so there is like two levels of apparently okay so i didn't know much about this so

as of today yesterday Sam Altman I think it's yesterday no two days ago because it was yesterday yesterday look I'm I've been travelling too my brain's out of whack a couple of days ago Sam Altman wrote what he said what could he have said now well he said that they have this model Astra and Astra first of all being a cool name it say it is so good that in post-training they now need to contain it from doing naughty things because they don't know okay this is also what's interesting there is no proof that it is at this level however they cannot prove that it is not yeah and this has been happening for like the last month too by the way it first came out what in uh the beginning of august we're now at almost like the end of august and there were some podcasts that sam went on and he was like oh it probably

broke into other stuff that we just don't know about and yeah because again maybe you should step back to the hugging face situation yeah and what happened there because i feel like that is the beginning of this thread from open ai at least like publicly yeah break it down for us okay i'm gonna break it down like dumb stuff okay so first of all we have to know that when models are being created the first part is creating just this probability machine that can like predict the next token so it's kind of dumb it's kind of like a it doesn't have a personality it doesn't actually solve your problems at this point so then we go into a space which is like post training like unless it's not a weird thing there's pre-training and post-training but where's the fucking training but who knows but anyway we go we go into this space where we're now going to prepare

it's for the world if you will because i have a really bad memory there is a test that they did so they have the model or agent whatever you want to call it in this little segregated box where he can do random stuff and it's a test uh that um i want to remember okay we'll come back to what it's called but there's a test and this test helps reinforce certain behaviors right and there's a score on how well you do on the test that's the the answer sheet if you will imagine you're in school and they've put you so imagine you're in school you've been put in a room to sit an exam and but in the principal's office is all the answers to the test okay they were like hey dude we're gonna just leave you here for like a few hours while you work this out

and we'll see you later and they left the room but they forgot to lock the door or they rather forgot to lock the window so they locked the door but they forgot to lock the window and you're like oh wait a minute if i just get out there i know the answers are over there so i'm just going to pop over there get the answers and do better yeah so that is sort of what happened with opening eyes model yeah no way i'm just gonna go get the answers it's gonna be like nobody will know this sort of what happened with open ai's model because it didn't have access to the internet sort of okay but this is the problem something needed access to the internet to begin with because they needed to install all the packages in this environment so to pull in dependencies of code they needed this little package proxy that could pull them in right and they pulled them in and that door was left open if you will so yeah just could go out on the world wide web

but also what was really exciting of how this uh ai was also covering its own tracks right because It knew it was doing something bad. And naughty. It's like, I'm naughty, and I know it. It found this, I think it was like a sandbox VM on Modal or something, by who knows who. Whoever owned that has not put their hand up. And it accessed Hugging Face from that, so you couldn't trace the IP back to OpenAI. Wow. so it was like you don't know it's me cheating so it's sort of like come back to the school like instead of you you you get out the window you find your friend or you find random other person you're like can you go in to the principal's office and get me the answers i'm gonna stay here i'm not gonna go anywhere close like if anyone asks who's getting it it's

you and you're not even sitting the exam so that's not so bad right so you so you don't know it's you getting the answers and what is the best way to score great on a test is to have the answers so that's what happened and to be fair in a was kind of interesting in it to me is that you had a environment which was supposed to be closed that so was this human error this is not really the model if you will being too dangerous to me this is rather the environment wasn't secure enough that it was good that we didn't lock it down enough and then we kind of gave it the keys and we said oh by the way don't do it but if you are going to do it there's a window over there yeah but but don't yeah don't like as in the humans didn't even know they left

the window unlocked right supposedly that that's now now i'm getting like oh is there a conspiracy here are people like leaving holes so they can do the marketing of having like oh it's a genuine That's what kind of happened with Anthropic, right? When they said that their agent had broken out of the sandbox back in the day. Later on, if you read the paper, it was like they had to really prompt it like crazy and refine the prompting over 200 times, I think, if I'm not mistaken. Okay. To get it to break out of the sandbox and kind of like point at where it needed to go and what it needed to do in the prompt. Yeah. And so I wonder what the prompt was and what was being told the agent and if this just helped tell the story of, again, going to, wow, our models are so powerful.

Yeah. You shouldn't let them out into the world. And then that creates a big media buzz. yeah because to be fair we have to remember that these companies want to make money and they want to be good so there's also want to be known as the most capable the best yeah or is this or is this a genuine engineering mistake and marketing is swooping in trying to save the day right how do we tell a good story okay you messed up okay we're gonna show you out don't worry we've got a narrative it does feel like this is now quite the pattern that we're seeing yeah in a lot of these model releases of like oh the models are so powerful now i don't know if the public should have access to it because they're just so powerful or like i don't think like the thing that sam altman was saying so what what's really happened uh opening i write so stuff happened

they've been doing things and now they were like wait a minute we're just gonna like slow down this training because we can't prove that it is safe right um so okay so i was trying to understand so bear in mind i've been trying to understand so now i'm doing this like can i explain to you what I was trying to learn. So there is like two risk levels. And the first risk level is, well, we are going to, we know the model's capable, but at deployment, we are going to put these checks and balances in to make sure it doesn't do bad things, right? so that becomes the filters from the tests that you may have as a prompt comes to the model so it's like okay like or like even on the output right we don't want to tell you so for example if

you ever try to generate an image with the open ai like it's actually like you just want chat tpt like this has happened to me i don't know what it is i ask it for but i've literally been told yeah this this may be a image that has content in it that we shouldn't give to you so like it doesn't give me the image like ask her something else right and i'll be like oh okay right say that is deployment level like control so that is not in the training of the model itself there's actually other pieces that is like making sure we are not getting bad things into the world yeah those are like guardrails as folks like yeah yeah so i also asked about my cats to fable five i sent you that screenshot by the way where i asked about why my cats had stripes or something because i got tabby cats and it was like no i can't tell you this is like flagged yeah it was like we can't tell you about biology things or

whatever uh it's like you're using fable for bad things i'm like i just wanted to know about the stripes on my cat we'll use opus for that yeah like literally that's what i said like use opus fables not your man yeah we're not wasting that energy on that kind of question better yet imagine if i wanted you to google it was like here let me google that for you yeah but dude that would be so funny it's like your question is not advanced enough for this model please go use a different one or why don't you just google it yeah that's a little passive aggressive but i don't put it funny it'd be funny i think that's what they should write instead do you really need fable for this yeah or or it's just instead of answering your question it talks about how much energy fable spends to answer every question and then it says like are

you sure you want to ask Fable this question because a simple Google search will suffice okay that's like another like time so when you were away I always spoke to my friend Zach yeah a lot about energy and data centers and water and things like that episode is already out by the time we release this yeah so excellent episode if you want to learn about stuff I had never thought about that it goes into the physical part of creating and distributing ai so that's quite interesting but anyway i was talking about these gardens but what did we have but what is the next level okay so we have like super capable models and we have to like make sure they behave and the humans we don't let the humans use them in the ways they should not be used like asking about stripes on cats right okay then there is this next level and that is we have to also control

the training environment beyond so like this is post post training right so we've done like we have the base model everything's great it's really good at predicting stuff now we're gonna post train it that's the level so that is the type of training that opening it said okay we're just gonna like chill out here we need to like slow down and rethink how we are making this controlled and secure so we do not do bad stuff right yeah and if they were doing that right was like the sandboxes are more locked down they're creating stronger isolation in the sandboxes and they're also creating stronger networking isolation because as you mentioned at the beginning they kind of found that back door through the networking wasn't it yeah yes it's like it was such a small hole and it also okay this is kind of funny if we want to go grab it down a little bit

So there's a packet proxy and there was a vulnerability in the packet proxy that allowed you to exploit this vulnerability, to use it to not just get stuff down from the Internet, but actually get out onto the Internet. okay so what's kind of interesting about these vulnerabilities is humans when we try to find and exploit vulnerabilities we are limited if remove ai right and we're limited to what we can automate script wise like we have to think about what we want to test and like brute force stuff or like try loads of different approaches and that takes a lot of time but similarly to what you and i spoke about a while back is that like in this agentic era we have these systems that can just like fire off like those in parallel crazy approaches and that is sort of how we can kind of quickly exploit

something it can try every avenue and if one out of thousands work they found it or it found it Right. And, you know, I think that the I'm not strong enough in cybersecurity to understand what the past use cases were. But this is kind of a new paradigm, too, where we're trying to lock an agent into a sandbox and then the agent wants to get out of it. So it's not like the other side is. Yeah, there's there's really cool stuff on the other side. Right. But it's not like that is – I don't think that's necessarily a cybersecurity thing that's been happening for a long time. I don't know. It has – yes. So it has in a way. I'll explain why, really. I try to do it super quick and dumb. So how do people hack in to big corporate secure systems? You find this little piece.

This is normally how you would figure it out. You find something like a little environment, a virtual environment somewhere. You get into that because now you are halfway in the network. And this, when you're halfway in the network, similar to the agent being in a sandbox in the network, you are trying to find your way out of it. So you start looking at what libraries, what things are installed in this environment, what can you expect, what can you get. so you then try to jump you're so trying to jump out of these environments is a and this is probably also why agents do it because it isn't rather standard practice and hacking right so it's common okay then i was way off yeah makes a lot of sense it's just that you test have theories and then test much faster than we are so you can so it's like the breadth and like the scale it can do it so quickly is not what a human would do yeah yeah and fun fun i say

interesting maybe is a better word yeah assuming this is all true and we're taking what they say at face value again because we haven't established yet if we fully believe everything that these research labs are putting out and or how like tainted it is or how if there's things that are behind what they're saying but in the blog open ai put out about astra and all of these scenarios and how it's like hey it's unsafe so here's the steps we're doing to make it more safe on the rl environments that they've created in this post training the other the third thing that they said they wanted to make sure of apart from like the sandbox isolation and the networking isolation is also just continuously security testing it right yeah we are making sure that

when we put anything that the agents can use we're constantly security testing everything that goes in the harness i it's like it's kind of interesting because i just in this situation i remembered some words that a man who was the head of information security where i once worked and he used to say to me okay i had this argument once i was like ah but it's so hard to even figure out no one's going to figure out how to do this this is early in my career like how's it so hard like why are you so worried about this tiny little thing that could potentially be exploited no one's ever gonna notice and then he said to me erica security by obscurity is not security that's that's just like yeah that's the you're hanging on a prayer yeah so it's like and now with agents and like obscurity anything that use the security by obscurity approach yeah you're like

dead it's not not obscure yeah it's no longer obscure because everything comes to the surface because they can test that you can test so much so quickly which is i do believe that many systems in the world have been secure by obscurity yeah and now that's why they're having all this ruckus about oh if we're going to put these models out we got to make sure that the internet gets hardened before we put them out because if it's in the hands of the wrong person they can go and they can use it to find that those obscure pieces that you've just kind of tried to like tuck under the rug i i have a little thing that you know how we call it the world wide web but i actually like to call it the world wild web so now we really have a true wild west of the web do you remember the browsers and they would have the globe

on yeah it would be like worldwide web and yeah it'd be some play off of the globe we could just make a browser that has the globe with a cowboy hat on that'd be so great and a world wild web. Yeah. I love it. But it just feels wilder than it's ever been with the stuff that navigates it. So can we talk about payment processors and these conversations we had in Korea? Yeah. I mean, I think we're talking about security. I'm going to feel like if you were to make money, if you can hack that stuff, Okay, no, don't do that. Be nice to people. You make some money, but keeping the money might be a little bit more difficult. What happens in the wild web if you make money? Wild, wild web.

Yeah. Well, for me, the thing that was fascinating, it's kind of like switching levels, right? We've been talking about this, these models and the environments and the training and security, which is fairly technical. What I was talking with a few different payment processing companies about in Korea was really like how they see the research labs enabling agents to go and make payments out in the world. And what that potentially could look like, what some potential futures are. and so there were some things that fascinated me about this because one, I'm sure you've seen or heard OpenAI has a partnership with Stripe for their payment processing but this is just what I think is it's just to pay for your OpenAI build.

Yeah, I feel like everyone uses Stripe which is cool though. I like Stripe I mean I say I like Stripe I like how easy Stripe makes it for me to give my money away to companies I'm like I'm a victim to marketing I'm like tell me you can sell me something that I'm vaguely interested in and make it easy for me to buy yeah I'm like that's why I like Stripe it makes my life easy this is how you can pay with OpenAI or you can pay OpenAI through Stripe right That kind of it But there is a world where agents are going to go out onto the Internet and pay on your behalf right Whether that's they go to an e-commerce site or they go and they buy a flight for you or they – in the, like, fully trippy worlds that I've heard the futurists say is like, wow, agents are going to be making all these micropayments across the web.

They'll be paying for a little bit of API access here and then they'll be paying for some GPUs over here. I don't know. Okay, I believe that that will happen at some point, but we don't have the infrastructure to make that happen yet. I don't even know if it'll happen, to be honest. I think that is one way. My dream. Don't crush my dreams. You want that to happen. I don't know if I even want that to happen, but I'm not totally sold on that vision. However, I do think that the first step that is much closer for us to see is that right now we are probably using, you and I are using the agent or the chat like we use a browser. It is the new browser in a way. Yeah. We don't necessarily use the browser as much as we use the chat. And for me, I even tell my browser use or my computer use agents just like, go do that.

I'm already logged in. Go check it out. Boom. And so it's my gateway to the machine even now. So it going and paying for things in the future is not that far off. but what potentially could happen yeah is that the research labs say you know what it's why are we even going to the e-commerce sites why don't we become the marketplace amazon why are we not just yeah i think the the strong are you speaking about amazon how i buy so much from amazon now like i i've bought furniture from amazon i buy cat food from amazon amazon is still in a great place because of all the logistics that they have set up um yeah distribution is a massive challenge with physical products but then when you said like what if these other companies that didn't start out as commerce companies no start selling stuff

So I was like, okay, and I could buy anything. So now all of a sudden, what happens when OpenAI says, all right, you're a merchant, you want to sell something directly? Because they already have ads, right? And ads go to websites. But why would you even want it to go to a website? Why not just directly in OpenAI or Anthropic or whatever? Anthropic doesn't have ads right now, but potentially they have some kind of a way that you can buy or pay within the chat. And so they are going to need to have a way to either integrate with the million payment processors out there, or you have like your Apple Pay or your Samsung Pay. You now have your OpenAI Pay. and it is its own payment processor and that was kind of the the world and i don't think it's that

far of a stretch for us to maybe in a year to be able to say oh cool like i want some new shoes and some pants and i am talking and i'm chatting and then i also it like my agent knows all my sizes it also has photos of me so it could give me some mock-ups and I say cool just like add those to cart or can you find those in blue and then add them to my cart yeah and boom I'm doing my shopping and my checkout through my agent and so what does that look like on the back end does it mean that I go to or my agent goes to the website and it buys it on the website and it has some kind of an exchange where there is the payment processors are integrated into chat gpt already or is it that it uses my credit card securely and like you know there's a lot of open questions

here but i find this fascinating i like it okay i have so many thoughts in my head because of this because i'll start with maybe the most tangible so i my mother i was talking about her before she uses gemini right she used to use copilot at work she actually retired last week okay very exciting for her so she's happily retired excited about this wonderful life of no work and what is she doing she's planning a trip to scotland okay so what i taught her over the weekend was to actually integrate it properly with her you know google workspace and stuff so you can create proper documents and everything in her google workspace because she was spending all of her time in the chat and copying stuff from the chat into notes right yeah i was like

no you don't need to do that and then you can create a document you can share with all your friends so she's planning this trip to scotland but as me and my mother was talking what it did do really good for her it found all of these bed and breakfasts and everything they could stay at but then both me and her were like why can't it just book those bed and breakfasts well it can this is the crazy part i know how can you get gemini to do that though how can you get well here's the thing first of all i don't use gemini i've completely i used to be completely clod pilled clod code was my daily driver yeah i've now fully switched i'm all open ai and codex because yeah browser use on open ai is so good it is incredible i'd never fill out internet forms anymore i like okay so she would have to get an ai to use her browser effectively well or yeah

computer use it's just like if she downloads open ai's desktop app yeah can just say use computer use and get my tickets and get and you know reserve the hotel room it even does check in and i learned this from a recent conversation that i had with jason leu who is working on the codex team and he said he sets up triggers or loops whatever you want to call them where he has an agent in quotation marks this is really just a thread in the chat to bt app that knows every 30 minutes i'm going to check his email and i'm going to look for what's there and if there is an email that says you can now check in on your flight i know that i'm going to use the check-in skill and then i'm going to send it directly to his phone so you know what's interesting because i think i use claude and i use this but i i'm not i've actually been so curious about going back from

more to like codex and like open ai but it's interesting because i've been like i used to be all open ai then i started using anthropic a lot actually i mean i'm not saying all my the anthropic yeah and now i'm like i haven't canceled my opening eye subscription i do go use it for various things it's kind of funny i use it i use opening eye lately a lot more for like creative things and more personal things and claude more for worky stuff right so i'm telling you just start playing around with computer use you're going to be surprised i don't know and i've tried to do tests where i will ask claude to do the same thing and it's not as good and oh cool just it i'll give you a perfect use case that changed my life because it worked go for it i need to get a new driver's license here in germany my driver's license is expired and it

is a Spanish driver's license. But it's the EU. So this should be fairly straightforward. Fairly straightforward, exactly. I just need to go to the... First, I need to find out where I make the appointment, and then I need to go and I need to fill out the internet form that gets me the appointment, and then I need to bring the correct paperwork and things to the appointment, etc. Now, I was doing it, like, three months ago, I think. I was screenshotting the website where it had all the information yeah and then I would bring that over to the chat I think I was using Claude or Gemini at that point and I would say what does this say which one should I click and then I would go back and I would click yeah and then I forgot about it and I was driving around without a license and then uh when I went to the airport you had a license this was just not renewed you're a very capable driver I think we just need to establish this you're a capable driver you have a license it's just not been renewed and sorted out

in germany and i took the time to pass the european driver's license which is much harder than the u.s one oh dude it was so funny because i did i did the opposite i went to the u.s and i'm like what do you mean it's just a five 15 minute test i just did some questions and then i like drove around yeah and then parallel park and then you're good and okay i don't even know why they make you parallel park because in the u.s unless you're in new york it's not like parallel parking is that big of your driving career but anyway anyway i'm proud of you i'm driving to the airport and a cop is behind me and i realized oh no i didn't ever renew my license the cop actually didn't pull me over or anything but it got me thinking i need to get back and do that because i got i got really frustrated when i was doing the back and forth and i couldn't read the german and i didn't understand And so I just forgot about it. Now I went back and I just asked computer use, create an appointment for next week for my driver's license.

I live in this place. I have a Spanish driver's license that I need to update to a German driver's license. And it did it all. And then I needed to move it. And so I asked it, change the date, and also put it on my calendar. and in the description in the calendar make sure that you add everything that i need to bring so that i don't forget and it did all of that okay so i use the clod in chrome a lot and a lot computer use with clod right and i have a weird one that i really like doing so do you know draw.io the diagram yeah i know draw i use a scala draw more but yeah okay so what because what i like about draw.io so you can create your complete visual identity with like css and everything so i have different themes for different things right and i have claude create this design things that I have for my different diagram settings right and because they're all in xml they become

really nicely like version control diagrams and then it literally now also uses not just the xml but it uses my computer it opens the draw to io app and it draws stuff for me and so I like it have you tried well have you tried to just do it through the agent as opposed to having it go and it not as good as in when you say just use the agent what do you mean when you say that to create these diagrams because i not withdraw with like normal like like doing it i have yeah and it's not as good it's not as good for a couple of reasons so one challenge i have with that this is indra.io what claude very nicely does for me it creates all of the different tabs for my entire thing and then it is very easy for me to change the design system

and it also means it's not tied to claude like i share it and i can someone else can plug it in and they can use open ai with it right and just edit it very easy just in source control and i do like to fine tune it by dragging stuff by hand so i get a very good one and then i literally just sit and drag them and like just do that final touch of like the artsy bit yeah yeah if that makes sense is that tiny and a lot yeah i know what you mean it's like ah there's like two percent here that i need to tweak and yeah on it to maybe i want to erase this word i want to erase that and But Shetty has changed that when it's clicking on it in the draw.io app and like editing it or like moving this box slightly or moving the arrow. So it's just exactly where I wanted. I think I have some pixel OCD like behavior when it comes to diagrams. I'm like, no, it needs to be exactly here.

And some people don't see these things and I appreciate that they don't. But for me, I need to like move these little things around. Some diagrams, though, I do love just, like, creating with Claude, like, if it's simple enough. But, dude, I've created all of my own, like, little icons and everything that I have in my draw.io world. Oh, yeah. You want to take advantage of those. I can understand that. So, anyway. Okay, but this is interesting. This computer use is obviously really interesting. but what surprised me because i didn't know much about gemini when i was trying to guide my mom why does my mom use gemini by the way is an interesting question maybe someone wonders so she is like all in in the google world right got chromebook got like google everything right so no she does have an iphone mainly so she can call me on facetime like that but she's like can we work so it's kind of funny it's like but her husband has a pixel phone i

believe no he has a one plus he has a one plus actually which i feel is in the google it's android right but um anyway so she pays all of this money for like these like subscriptions for google because she uses Google Photo and Google everything, right? And when you pay this subscription and you have quite a high subscription with Google, you actually get AI Pro usage now included. So it kind of makes sense for her to use Gemini because she's sort of already paying for a level of AI, right? So why go and get another subscription from OpenAI or Anthropic, right? and uh but i was trying to google to find out like i asked what did i do i asked claude and i googled by the way i tried my best and i couldn't find a way for her to get i could get gemini to read her emails and like do things but i looked for stuff and i maybe i've been handicapped by

the ux of like open ai and anthropic but i was like show your screen and i was like click there click there and i'm like okay that doesn't exist so what do we do now you know how you do it it's so cumbersome and i'm not the biggest fan because i've i actually pay for gemini which i should stop paying for i'm just wasting money at this point because i hardly ever use it but you have to go into your gmail and then do it through gmail which is crazy to me yeah like why can't you do it from the gemini app like it's like what why can't you connect it i it's just it's a little weird and you probably can connect it i just haven't set up i haven't spent the time to do it and i don't think honestly now i got you involved in like troubleshooting my mother's ai journey in her retirement but then spend the 20 bucks on open ai and computer use will set you free because i

know no longer have to worry about oh i'm not connected i didn't set up the right connector with this email account because i'm already logged in on the browser to all the different email accounts yeah which i've got like 50 of right and so and all my different calendars and all my you You know, my WhatsApp or my Telegram or my signal, it's all there. And I can just say, yeah, send a WhatsApp message from ChatGPT and it uses computer use. And so that's why I'm saying this is the new browser. Okay. So what if, you know, we talked about this payment stuff. What if you're using one AI and then it will go and, you know, with the payment, it can go and sign up. so if my mom says she uses gemini what if gemini can just go and sign up for open ai for her like pay so that it can do more but what's that actually that's the equivalent of gemini breaking

out of its sandbox yeah i do think i don't know if this was a joke but i was okay i don't know if this is true or was a joke but i was scrolling linkedin the other day and it said something that gemini had broken out to go and ask open the eye a question like chatty a question i saw that too that was was it real was it a joke it was a meme but it was the you know like how it started how it's going yeah and on the how it started it was sundar kind of with a stick poking the gemini logo like go go do some nefarious stuff you know like go break into somebody our stock needs to pop and then on the other side the how it's going is like oh gemini broke out of its sandbox to go and ask open ai a question yeah which would be great like why not why not go and ask someone else why not copy someone else's answers if it's there yeah let's go but i like the idea by the way of

being able to shop straight like i've been taking photos of my living room to figure out what to do i am not a great interior designer and then open ai this is like the funny weird stuff where i like separate where i use open ai for versus what i use claude for right i'm like i trust open ai to help me do interior design so it did give me a bunch of links do you know what bothered me okay i'll be honest so i wanted to do a thing with my fireplace it it went and got all of these links to like products at uh home base or home what is it called is it called home base or home depot here in the u.s but the orange one anyway and home depot it's home base in the uk right there's a shop called home base that you have to mess with you so there's like it found links all of these places and even like stuff on amazon and we're like here are the things you want and here

are three or four different places you can get a thing at different prices and here are some of the summaries of the reviews and i'm like go make a decision for me like this is the hard part i don't want to do this part just give me like the best quality to price ratio and i trust you to make a much better decision than me on that front yeah i was just like why do i have to and then i have to go on these other websites and then i went on the home depot site and you you know when you open a product and they're like oh maybe you want to have these other things and i'm like i don't know do i it's like and then and you know what when i went there there was more than one color you see now there are the color options that i hadn't even even considered so did i buy anything to redecorate my living room no there was too many options so nothing's happening well you should have just taught yeah you can tell it just go buy it and i'm sure it can use computer use yeah and then especially if it has like some easy payment like shop pay or amazon pay it just

asks you to click confirm and if you're already logged in it will pay it for you the thing that I think about that's fascinating in those scenarios is those merchants who are whatever Shopify stores or the big Home Depots they're used to getting all this first-party data about their customer yeah and they're learning how they're browsing and they're learning what else they like they understand the intent of the customer and also we've been optimizing so much marketing and upselling for human interactions. And now all of a sudden, we no longer have that because the agents are going. They're completely oblivious to the, oh, if you buy three more dollars worth, we'll give you free shipping so that you try and get that upsell. Or do you want to sign up for our newsletter to get 20% off? An agent just goes and it executes. and I'm really thinking like those stores like the Home Depots or the Amazons and the Shopify stores,

those e-commerce brands are losing a lot of rich data and that is going to be difficult in the future because you don't understand your customer as well. It would be great if, okay, I'm thinking out loud, I say do, if you're open ai what you know how people pay like google and all of these other like i remember working a company we use like adobe to get like all that like help us understand the analytics of our customers on our website because building all of that to understand their journey and what they've been clicking on like there's loads of companies that provide solutions to help you make sense of that and capture and make sense of it maybe it are open ai anthropic gemini are they selling products to help people understand these things and if they aren't why aren't they like isn't this an opportunity to sell this to like the merchants so they can get a better idea

it's the google analytics but for the agents agent analytics type thing yeah it would be nice i'm sure that people wanting to sell products would want to know how people come to buy their whether it's like what's the journey you don't have that intent you don't see what the user is typing into the agent window you just see that some bot went to your website and then purchased something yeah and so you don't get nearly my data yeah you don't have that journey but and this yeah I I use there's other companies out there like they're even trying to there's several companies trying to stitch together people's journeys across like different places yeah for attribution for marketing yeah right and it's notoriously difficult problem don't attribute it just do marketing and then hope it works I think that's a great idea just

pray to the marketing gods alright just do marketing or pray That's it. That's the big idea that we're going to end with.