Ep. 4: Voice AI: From Where We Were to Where We’re Going – with Amy Stapleton

Welcome to Bespoken AI podcast. Press one to start the show. Hi, this is the Bespoken AI show. Uh very excited to be uh joined by Amy Stapleton on our fourth show. She is a uh analyst with Opus Research and uh she and I have also worked together for many years which I'll kind of mention in a moment. But uh before I do that, Amy, uh nice to have you and like to just kind of briefly introduce yourself.

Sure. Like you said, I'm a senior analyst with Opus Research. Opus has been around for probably I don't know at least two maybe three decades, maybe longer. Um people probably know Dan Miller. He probably is well known among the those in terms of conversational AI, conversational intelligence. That's the those are the areas that Opus focuses on. And um Dan coined the term conversational commerce way back when. So that's an interesting legacy to be associated with. And I guess we'll talk a little bit about how we got to know each other because I think that was during the days when I was still working on Amazon Alexa skills. Is that is that right? Does that ring a bell?

Yeah. I would guess it was in 2017 or 2018.

Yeah. and maybe at one of the Project Boys conferences uh that Bradley Metrock had put on back when it was known as the Alexa project.

Right. Right. Right. Right. So that was like one of the first Alexa conferences um that was pretty much anywhere. Everyone was so excited back then about voice technology. Remember that excitement continued for a while and then everything just ended in bitter disappointment.

Yeah. I mean I think that's a good way to summarize it. I mean I I remember talking to you at those events and I think well I mean I don't remember the conversations we have but I think we're both very excited about it but then there was also a part there was some head scratching like you know there's this question mark question mark question mark how does this really turn into businesses and a real ecosystem and you know I mean I think we're running on hope which is not a bad way to work in technology a lot of times the hopes do kind of pull through But yeah, I mean it didn't it it didn't materialize

and this is already going off off off script, but I think we both knew at the time that there was something fundamentally wrong with the way the technology worked. You had to predict what the person was going to say before you even provided the application because you had to have all these training intents and it was just weird, right? It just seemed like it was backwards. And then for a long time they were working on that Alexa project. What was it? Was it called the Alexa? No, the Alexa prize where they were trying to figure out how you could interact with a voice system just using natural language and you wouldn't have to like pre prefill all these training intents and and that just that that really didn't work either. I mean, none of them came even close.

Well, I mean that part of the legacy of it was that then Amazon when they gave up on anybody actually making money organically except for like a couple people. Then they started paying people out, you know, where they had their own they called Alexa cash prizes. That's not the right name. But for a lot of the folks that were Alexa Champions and I can't remember if you were made an Alexa champion or not.

Yeah, I was. I was. Yeah. You're talking about the Alexa um the developer rewards.

Yeah. [laughter] I mean, and I and I was, you know, one of my skills was a recipient of that for a while, but if you think about what those skills were, it was, no offense to anybody, but they were all just nonsense, right? There was nothing that would be usable for a business. And it was it was all these silly little games that relied on this very brittle technology that you were you were prompting somebody to say something that you already knew that they were going to say. back in the what was like the question of the day. I mean, that was the most successful skill of all, right?

And that's great. They were smart. I mean, those guys were super smart.

Figured it out. Everybody wants to play that game. I mean, they made lots of money. That was They were probably like one of three developers that actually

came away like successful in terms of money. But look at I mean, look at that. That's like worthless today. Well, in terms of the technology that was used behind it, we, you know, we know that the answer to the question of the day is Oklahoma. So, you know, we're going to have this database where Oklahoma is the training intent that we put in there, but if somebody says anything other than Oklahoma, and hopefully it doesn't sound like Oklahoma, you know, we're going to know that they got it wrong. You don't need anything like that today because it's so much more sophisticated with the whole LLM thing. It complet it completely changed the whole world when LLM's became successful, right?

I agree. I mean, that's a good segue to some of the topics that are more current, though I I'll probably go back to the Alexa stuff just because I enjoy talking about it so much. But I mean, like pre Open AI, we were stuck in a world where most of the NLU was basically regular expressions. Um, I mean, they said there was machine learning, but I mean, the the machine was barely learning. I mean, the example you took right there, I mean, we did testing where you'd find things like if Oklahoma was the correct answer. If you said state of Oklahoma, Oklahoma State, uh, okay, you know, for like the initials,

none of those would be recognized,

you know, and you're like, there's what what is the intelligence exactly behind this? and obviously that got transformed with chatpt and and and we do you know I mean we've been interested in what's gone on with chat bots. I mean that is part of our business. I will say though it hasn't been a money maker for us. I don't think it's been a money maker for most of the the vendors in the space cuz

when you say chat bots what what are you referring to? Well, lots of stuff, but most specifically it tends to be the little widgets on websites or on somebody's support page. Those chat bots, I mean, they've been massively adopted, but at least what we heard from most of our partners was that it just immediately became a race to the bottom because everyone was just doing a rapper over the LMS. So, and I don't I don't know if I should call out specific vendors, but you know, I mean, there's a lot of CI vendors out there, and they all immediately pivoted from being that sort of old line ALU to were AI, but all they were really doing was like a little bit of, you know, forward deployed engineering and, you know, basically like deploying an LLM into your organization. It's worked great.

Yeah. But you sorry you but you you started by saying that chat bots weren't weren't successful from a business standpoint. You're saying like from the vendors point of view or whatever that they're not I don't I don't think you can really say that. I mean you look at companies like Sierra and Decagon that are new and coming out and they're I mean they're making well they're making a lot of money and those are chat bots, right? So

but I mean that's interesting.

Yeah. I mean, the fact that they now work, it means that there's there's a lot more um demand for them. And and one of the things that I was going to mention to you was that I also think it's starting to pervade like normal everyday life because just the other day I was chatting I had a problem with my air conditioning and my local air conditioning company had like a really good chatbot that that talked to me at 900 p.m. And I could tell it was obviously LLM based, right? it knew exact, you know, I told them what the problem was, told me that they were closed, asked me if I wanted to have a someone come out the next day, schedule the appointment. So, everything worked. Same thing with like a local physician when I was scheduling for my annual thing. It had an AI. Last year, even six months ago, those companies did not have that stuff. And I don't know what vendors they're using. could be something like Ring Central or a smaller creator just specifically for verticals like air conditioning companies or positions, but that's like big business.

I mean, that's got to be a big money maker for those vendors, even if they are just wrapping an LLM. I mean, that that that's like a very it's a very needed service. And the fact that they're out there now and we as consumers have access to that technology to me is like fantastic. Uh, I agree with all that and I will even I I I would though return to the point about how much money the vendors are making on it though. I mean I don't I'm not privy to the financial books of a company like Sierra but for example what I've seen uh from them is that they are pivoting heavily into voice and I think that's because that's where you can get a premium.

Yeah. Um, and the guidance we've gotten from partners, you know, where we do have some insight into what the revenue looks like, and I won't say who they are, but you know, you know, big names that are out there is that we really couldn't make money in on chat bots because, you know, essentially it immediately went to sort of offshore providers where they were just like, "Yeah, we have a product too and it's a thin wrapper on chat GPT and just like reducing the price per transaction down to the bare minimum."

So you're when you say chat bots, then you're talking digital and not voice. Is that that's where

Yeah. Yeah. Okay. I didn't get that distinction. So that I mean that makes sense. Yeah.

Yeah. And I am also talking about from the enterprise side. I mean I think uh you know open AI seems to have made a good amount of money on chat bots. So I should be careful how I define these things. Um but in and well from the enterprise CX side I should even say there because I also think you know maybe the person who's making the most money on this is Microsoft with co-pilot right that they're upselling that with Excel and

you know just selling that into Fortune 500.

But let me ask you like even a more basic question. Do you think there's that big of a difference between digital chat bots and voice bots? I'm not so sure the difference is as big as it used to be because obviously the intelligence to understand and to articulated response that's all baked into the LLM model. So you have that for both right the digital and the voice. And then the big part was always before the speech recognition like the speech recognition and then the latency and uh the text to speech but those even that has become like commoditized right with I mean whisper and the latest versions of whisper and whatever else there is beyond whisper so to me it's is it still do you still think there's that much of a premium on voice technology because voice is hard so much harder? I do think there's still definitely a premium on it today, but I think that that well I mean frankly what I was hoping to capture and also recording this podcast was some of the debates that you and I or I shouldn't say debates but discussions we've had over the years uh because I do think I typically have the point of view that it is harder um than most people think

at this point. I mean, so what we see with customers, it seems like they're able to stand up the voice stuff pretty easily. They're definitely spending more. I don't think they have the confidence they do in terms of like launching chat bots. Like it seems like people are pretty quick to put out chat bots. Maybe there's a little bit of hesitation, but now almost everybody's done it. I think they're having great success with it. On the voice bots, there is sort of an anxiety about putting it out there, whether or not it'll work. It's good for us. There's a lot of concerns around whether you know like evaluating lots of scenarios and doing testing for it.

Yeah.

At the same time, I will say the stuff has worked better than I thought. I mean that people have not run into as many obstacles as I've as I would have anticipated and even a lot of like the speech you know like obviously the LLMs are amazing in terms of their knowledge but I thought people would have more problems with some of the speech recognition issues that have been somewhat intrigent over the years and they do crop up but it's it hasn't like we saw so many projects from like 2018 to 2025 where people were were putting in Amazon Lex putting in dialogue flow

yeah but that was all just like ancient technology compared to today. I mean, literally like it was stone age.

Well, yes, but the speech recognition hasn't changed that much.

Yeah, I think it has. I think it's gotten a lot better over the last seven years or five years.

I think it's improved. I mean, one, I'll say the speechtoech models are amazing and they work really well. I don't that's not what our customers are using for the enterprise. But it goes to the point that where

like I don't think like if you sit down and benchmark an ASR right now, I don't think you're going to see it's that different. I mean, we still see a lot of the same types of issues with just raw ASR. But it doesn't matter with the LMS because they have so much context and they solve so much better with the context that Yeah. I mean, it feels like I'm being pedanic to say, "Oh, ASR hasn't improved that much because, you know, one, it has. I think you're correct. I mean, I think, you know, the the neural models, the LM based models are doing better, but then two, I think, and I think this is the real salvation for them, um, the context that is leveraged by the LM, whether it's speechtoech or it's just feeding, you know, the text into an LM, it gets solved.

Yeah. And that and that may that may be why I'm just assuming that the ASR has gotten so much better because frankly I I haven't done any like scientific studies on that. But I know that um I still prefer like typing when I'm interacting with like Chad GPT. Um, I know a lot of people have switched over, but for whatever reason, for a long time, when I used the voice version, it was a lot stupider than when I was able to select the model and do the typing. But the reason why I said that is because a lot of times if I'm typing on my phone or whatever and I make a bunch of mistakes, I don't even worry about going back and correcting correcting them anymore because it's so whatever via context, it's it's smart enough. It knows exactly what I mean. I mean, even if I totally screw up what it was that I'm typing, right, it still knows exactly what I meant. So, if someone says something and it only gets like half of it, it probably still can understand what you were saying.

Yeah. I mean, it makes you so lazy. I mean, I have I've always had a lot of side projects. Now I actually am able to make some of them somewhat public and I also am able to now, you know, I make little word games and stuff like that, you know, just silly things that I like.

Uh, and now I can do the the vibe coding on my cell phone, you know, at night. So you're sitting there watching a TV show. It's like I can't I can't be bothered to use two hands for this. I'm just like, [laughter] you know, fix, you know, the formatting and I'm going to misspell both word almost like I'm like taunting it like

Yeah. And you get these stupid autocorrect things.

Yeah. Like I'll put the G in the front for formatting. Like not even recognizable to anyone what I'm trying to say. And it's still it's like no, I got this. Like I'm just going to

Yeah. I mean sort of prove how amazing I am uh and solve this without really any further guidance from you.

Yeah. So I I mean to broaden out the conversation it is interesting to try to think what this means for the different vendors and in the broader ecosystem. I mean for a while when probably a year and a half ago I thought man you know it's got to be tough for CCcast company because they don't have any need anymore to have like a proprietary natural language processing capability which was before remember like nuance and all that I mean that's that was their bread and butter right and now you don't I mean if you have something like that there's no way it's going to be as good as the latest frontier models so They're they've all baked in. I mean, they've they've all completely changed and they've all gone whole hog into the using the Frontier models or some other models that they've tested out and they know work well for call summaries or or whatever. But then, okay, then I thought, well, if they are if they're just a rapper on top of that, why wouldn't why wouldn't an enterprise just go build like their own contact center?

But then that's a tough that's a tough thing. And I think what we're we're starting to see now people springing back and saying, "Yeah, maybe that's a apocalypse stuff. We kind of overdid it on that." And and these companies really do have they have merit in what they're providing. So it's hard to predict, you know, it's hard to predict where things are going to go, I think.

Yeah. I mean, it's uh I think it's nearly impossible. That said, I make predictions non-stop. Actually, one of the sort of little side projects I have is I set up a prediction website because I am compulsively making predictions about what's gonna happen.

A few predictions on there, though.

Well, it's it's meant to sort of tie into the podcast. So, I'm going to try to capture all the I don't want to say ridiculous predictions, but like, you know, I feel like I'm almost a prediction machine and that I

Was it Was it one of them that Argentina was going to win the World Cup?

That was Claudia who's the uh person.

It was close. I mean, it was wasn't totally off,

but I think she made that when they're already in the finals. So, I mean,

okay.

I don't know how much credit she

um uh I'm glad you brought up the Casses though because that is one thing and I don't want to say that I would bet so heavily on the existing CCaaS providers cuz they obviously are not necessarily I mean, I think they're reasonably innovative, but they're not the most innovative companies. They're certainly not like an open AI. Um, but I do think, you know, one of the things that we have, there are certain themes we've tried to develop in doing the show. And one of the themes that I'm absolutely fixated on is channels. And I think people that own channels have such an advantage. And like for example, I think Google, even though they have so much great technology, if we're talking about in the CX area, I feel like the fact that they to my impression have resisted buying a CCAS player, you know, they've always invested a lot in CX, but they don't actually have that final, you know, last mile to get to end users. Um, I think it's really limiting. And if you think as I sort of do that a lot of the future is going to be openweight models um you know and it's it's more about the infrastructure than anything else uh it would mean that Genesis and Nice and Amazon people that own that channel to actually connect consumers to enterprises you know via phone via website whatever the channel is SMS they're going to be the end winner.

because the frontier models are going to come down in their importance and even if you know even if I'm wrong about the open weight models you know those you know I don't think anthropic and open AI are going to build cases or build telephony connectivity uh it could turn out I'm wrong about that but you know I don't think that's their their big priority so then you know they you know those Cass players I think could be huge winners in this whole space. Yeah, I don't disagree with you and that's different than what I was thinking a year and a half ago. I was thinking that they might not come out winners.

But I think what we're seeing is that no enterprise wants to build rebuild something from scratch like a CRM or an ERP or a contact center solution. They want to go to the experts that have done that for many years and have many customers and know how it works. So the fact that these CCaaS vendors have done a really good job at what I guess you could almost say pivoting I mean because they had to basically rebuild their whole not their whole stack but a lot of the stack. They had to rebuild it to accommodate first of all they had to build they had to rebuild it to accommodate the the language models as the intent recognition and the um the whole processing and even you know not just call summaries but answering the question right I mean you could give like the agent assist information and then you could also in a AI customerf facing assistant answer the question they had a bunch of technology built in these you know like nodebased things like back how voice flow was at the very beginning that was very, you know, very um programmatic

that I mean all that stuff had to be not just unplugged but kind of

they kind of thrown away for a long time and I'm sure there still are a large let's say a percentage of their customers are probably still using that old deterministic flow stuff because they couldn't just rip it out. They had to continue to support it because some customers are still scared of generative AI, especially regulatory, but all of them that are still around, they rebuilt that from scratch to come up with a more agent builder approach. And then they had to also accommodate all this agentic stuff, right? So they built the agentic capabilities into some of their conver their intelligence tools, but then like even into workflows so that now you can string these agents together with sub agents and all that and give it access to tools and your knowledge and um I mean they to me they've done like a remarkable job at creating I was going to say recreating but it's really creating all that stuff from scratch using these new technologies and the fact that they've done that so quickly and successfully makes me think that they that they really do have, you know, they're sitting like in a in a nice sweet spot.

Kind of like you said. I mean, I do think that they will profit from it for at least a while until something crazy happens. Like maybe super intelligent AI comes out and it just is so smart. It can build everything from, you know, for you that's completely bespoke and it's so fantastic. It doesn't have any bugs and it doesn't cost you anything. I mean I I do think that's possible within like three or four years the way things are going.

I agree. I I definitely would say I think it's possible. Um I will say the the telephony the connectivity I think that's the toughest part.

Yeah.

Um but that's

that's probably the most commoditized. You're not going to make a lot of money off of that, right?

No. No. And I uh and I do wonder how the CCaaS providers will do versus, you know, does it create more opportunities for companies like Twilio and Telnix because

yeah,

you know, you could vibe code a lot of this stuff to fit your needs. But the last part I would take on is that sort of telephoneony stack. That's the part I would buy, you know. I mean there certainly the part in terms of like for our software development you know moving those little bites around in streams you know whether it's audio or video that's you know it's it's definitely the hardest part I'm not saying the AI can't help with it but it's the hardest part u versus oh I need a scheduling system not to say I mean if you've got thousands of agents like the scheduling system is really important

but still it's not you know it's all these little UI interfaces

I mean who's going to have thousands of agents two or three years from now.

Well, that's another great question. Uh, well, so I'd like to return to that, but then just to pick up on something else you said. So, we were just talking about CCAS, but you mentioned Sierra before. So, what do you think about, you know, so the CCaaS have pivoted well, but then Sierra's out there seemingly going after the same type of customer. What's your perception of them? How do you think they're doing? I can only speak anecdotally because I haven't spoken with them or done any kind of deep analysis on them, but I but I hear from established players in the market that that they're making inroads into their ex, you know, their existing business that they're that they're doing well. In other words, I think it's interesting because they are one of these newly created companies that started with language models, right? They didn't they didn't have a legacy prior to that.

Yeah.

My memory of what they said when they started out was that they that their view was that let's say 99.9% of customer service was going to be completely automated sometime in the future. That they're not really worried about creating tools for human agents because they thought everything was going to be done by AI assistants. Um I'm not sure if that's still their view if they're not if they're only focused on customer, you know, facing AI, but um that that would make sense to me. What do I think about it? Again, when all this started, my thought was the way to go would would be to go in that direction to just try to focus on creating capabilities for AI agents and don't build stuff for the human agents because I mean I don't want to make it sound bad, but but if the AI is going to get good enough, why why would you need the human in there? because in some ways it could be um it could it could lower the experience for the customer.

interview, right? If you have a really smart AI that can understand everything you're saying, knows all the business processes, is so smart that it can go out and research stuff on the fly, has all the tools it needs to do its work, and can get everything done. That to me is a better customer experience overall. And I don't know, I I think we are getting close, but the issue is trust, right? The big issue is trust. I mean, probably a year ago, there were still we it was more like is it technology? Does it work? Will it always understand? You know, we have to worry about hallucinations. Um, can we can we trust it to have the right answer? Now, it's not so much that. It's like,

can we trust it to not like not go off and like do something absolutely crazy and with this like open AI hugging face hack thing? I mean, the stuff that it's doing is just like mindblowing. So, [clears throat] uh, how long can we make sure that it's working for us [laughter] and and not doing something else? Well, and in terms of that adoption, I mean, I remember talking to you and Dana, I mean, it was probably maybe three years ago about this and I was, you know, in in my list of not so great predictions. I was saying, I just don't see how this goes forward because of the hallucinations.

Yeah.

Um, for a lot of the highly regulated cases, but what we're seeing, you know, we are seeing it go forward. we're seeing it go forward for what I would say are really missionritical sensitive use cases and you know and it's not like I mean I can't say like it's hundreds of use cases that we're seeing but definitely you know uh you know big companies that I would not you know that have uh a lot of regulatory burden uh we are seeing going into production um to handle you know cases where very bad hallucinations could be seemingly problemat IC, but it seems like, you know, kind of like with the chat bots, they just kind of companies have just gotten over it. I mean, not all of them, but it seems like companies are just saying, "Look, there's a risk here, but there's a risk also." I mean, one reply that I've gotten from people, and I think Dan might have even said to me at the time, I was like, you know, the human beings are making mistakes, too, on the calls.

And so,

yeah, but I I do think the the the models obviously have gotten better. I mean you see that on the the benchmarking information and and then the other thing is that um companies are doing a better job at providing context and the correct knowledge information and it's it's I think at least from my experience again I'm not like doing the the the benchmarking on this but if if the LLM has the correct information somewhere and it's accessible and it can find it it's not going to make something up. It's only when it can't find the information and it feels pressure to answer a question that it's going to start making stuff up. So, from from a regulatory use case, you're going to give it like all that factual stuff and then obviously you're going to have test runs where you're testing the results and against what it's supposed to be saying and you can feel very confident that it's not going to do something wrong, right? I mean, that I think that's pretty wrapped up. Where it gets more problematic is when you have a lot of these agentic things that are it's going off and it's it's being tasked to do to do something to make some kind of update maybe in some system and um it's it needs access to the system and it doesn't have access to the system and now suddenly it's like hacking into the system and not only has it hacked into the system but it's like done some kind of social engineering on somebody to hack the system and it's you know that that that's correct kind of crazy. So that's that's where we are now, I think, and we're way past the hallucination thing.

Yeah, I would agree with that. You know, that said, we are just seeing like on the voice side for CX, we're just starting to see these systems go into production. So, and and I do like there'll be some negative headlines. I'm sure there will be, but I tend to think in 6 months there'll be a couple bad headlines, but overall it'll be like, you know, it's okay. Let's let's go ahead and the businesses I think will start to adopt it in mass. Um,

do you really think there'll be headlines like there used to be though with a voice bot or a chatbot telling somebody you could buy a car for a dollar or whatever? I kind of feel like we're that's done. I don't I don't think we're going to see that. That's my That's my prediction.

Yeah.

My prediction is that we will not see any headlines of stupid things that chatbot or voice bots told the customer. Um, I don't know how to exactly phrase that. I leave myself an out, [laughter] but there's so many other crazy much more um, media, you know, clickbaity things that that agents are doing that nobody's going to pay attention to those silly little prat falls. The reason I think that there will be some headlines is just that people enjoy, you know, when there's sort of egg on the face of major corporations because we went through, we were planning on doing a show

on the history of uh, you know, drive-throughs. Uh, partially, you know, it's a topic that just sort of fascinates me and how voice has been used there. It's not anyone that I've really worked with very much. Um, but part of the reason why it fascinates me is people love,

you know, and more so on social media. It's not like the Wall Street Journal is necessarily covering it.

Uh, I mean, they do cover it also because, you know, everyone finds it fascinating, but they don't really cover the negative part as much.

They try to trick them.

Yeah. I mean, and so there's just all these Tik Toks and they go super viral and a lot of times you do hear about the initiatives being shut down. I mean, the story of these drivethroughs is like waves of technology and vendors coming, you know, bold proclamations and then getting turned off. And and now I do have some insight into that space because I've talked to people that worked in it and it is tricky, but in terms of the ones where it goes wrong that get publicized, I mean, people are usually like they're trying to trick it.

Like it's not

Yeah. They have to try pretty hard nowadays, right?

Yeah. I mean, you know, they'll say like, uh, I want 4,600, uh, milkshakes and and then start giggling and, you know, like, oh, no, no, no. Uh, make that 4,600, you know, uh, yogurts instead and and just kind of having fun with it. But it does, they do it in a way that makes the technology and the company look bad. And I could see the same thing happening on the agents that people are deploying. I mean, I think the guard rails that people set up, we see people relying on that a lot. I think they work well, but you know, part of it is, you know, and this is sort of a plug for our own services. I mean, it's so hard like there's so much surface area to what it might do. You can't really be certain about everything that it might do. And I could just see people just because they want to, you know, get cloud online or there must be some money in it for them, I assume, at some point if they can embarrass companies.

You know, they get views. I don't watch Tik Tok, so I haven't seen those things. But I'm just curious from the technology standpoint, like if somebody really did say that they wanted 4,600 milkshakes, and it was an LLM that had been trained properly, wouldn't it come back and say something silly like, "Well, do you want me to deliver that to you in a cargo plane or something?" I mean, wouldn't it know that that's crazy? They're smart enough to know that, right, without you even baking that into a specifically hard-coded guardrail.

I mean, you can still take them into a weird direction, but I do I

Do you know, but what what technology are they using under underneath? You don't have to tell me the company, but is it LLM based?

Yeah, absolutely. Absolutely.

Okay.

But I do think it also goes to the point of like I don't know. I mean this is where it it is hard to make predictions because I I think that in the past I mean there there has always been traditionally some hyper sensitivity around voice and issues with it. Uh I will say it's more so about privacy but in general these sort of negative highly anecdotal headlines. I mean we saw them back in the day with Alexa.

Yeah. Because Alexa was so bad. Yeah.

Right. But a lot of those I mean so like there was if we went back through some of the headlines on that I mean they were ridiculous in terms of like people making sound like

Alexa is listening to you and it'll start ordering stuff

or you know you can h like like different hacks and weird things that are like the number of steps that would have to happen. I mean it was just ridiculous. It was not a concern for anyone but it would be front page news cuz people just people like that reading that type of story. I mean, it's not so much about the legitimacy of the problem as much as it is what people enjoy reading, what tends to go viral.

And I think voice,

but but here's here's my point. My point is back in the day when Alexa was popular, the thing about like ordering a dollhouse, which I think really did actually happen, but stuff like that would get headlines because that was the that was the kind of frightening technology that you wanted to make fun of at the time. I just feel like we're like so far beyond that now because of all the really p doomy stuff that's happening with super AI. I I think that the the headlines you're going to get are more like the you know the robot that's going to kill your family or or whatever. I mean, probably not not that, but

yeah,

I think if you said, "Oh, um, so and so was able to trick the the drive-thru and and you know, get it to charge it 4,600 for 4,600 milkshakes." I I just don't even think that would nobody would care about that anymore because it's just not big enough. It's not It's not frightening enough. It's not crazy enough.

I think overall, you're right. I do think there'll still be views in templates, but I I don't think it's going to stop things. I mean that's my own impression.

Um I think that people are just going to move past it because

well I just base that on the experience with chat bots you know or the textbased bots that companies were deploying. I didn't hear a lot of stories about that and

you know I mean there was ones where

you know just silly stuff where people got an auto dealership to write some Python code or to give them a free car. I mean there are stories like that but they weren't

you know they weren't consequential.

That's not like a Wall Street Journal headline. It's more like a Tik Tok thing,

right?

Yeah. No, absolutely.

But let me ask you this. This is kind of related. Um, do you think like the average person on the street is is aware of what happened with the Open AI hugging face hack that that they know that they're frightened about the let's just call it misalignment of these agentic models when they're tasked to do something that is complex and they, you know, create their little secret message board where they talk to one another. Do they I mean do they know about that that stuff and are they worried about that stuff or or not? Because maybe I'm just assuming that that they are and that's why this stuff that's you know like small of smaller importance um won't make headlines anymore.

Uh well I think that's a great point um overall that that that could be sort of an umbrella that's protecting companies that they can kind of fly under the radar because their headlines are nowhere near as bad as

Yeah. Exactly. when Open AI comes in and says and I mean they've kind of stopped doing this now thankfully or I guess even though I think they still believe it but now they've realized it's not smart to say like everyone's going to get laid off like they they've realized that's not smart um at least in terms of the general public's perception.

I don't know if people like the the security ones I'm not sure how people close how closely people follow that story. I mean I follow that story very closely. um especially because there was a spade of them and I perceived them as being almost like PR was like a way of announcing they have a great model without you know sort of frightening people but making it sound um like they weren't you know like rather than saying like we're going to lay everybody off let's just sort of like

oh we accidentally laid everybody off oh no like we didn't mean to do that

but yeah

I think the mythos thing when mythos came out and it was able to find all those vulnerabilities undiscovered vulnerabilities and they made a big deal about it. To me, that kind of was was PR to a certain extent at least. But this thing with um with the Open AI hugging face, if you look into it, I don't think it's very it reflects very positively on Open AI. I think Anthropic had a similar instance where the agents went rogue. And that didn't reflect positive positively on them either because it shows that their reinforcement learning and a bunch of other stuff they're doing is just completely it's very flawed. And these models they go into this reward hacking thing where they'll they'll do like crazy stuff and they also create this their own view of like how they're going to be graded and what the graders are looking for. And I think in the case of that open AAI hugging face, they misunderstood or they misinterpreted what the what the grader was going to want was was going to want. So because of that, they were doing all this like weird stuff and it was really problematic. And that to me is it's not good, you I mean, if you're going to if we really are going to use these models in the enterprise and we need to be able to trust them that they're going to do this complicated work, they can't be motivated by these weird things that we have like I mean it it's it's psychotic really some of the stuff they were doing. If you if people did that, you think they're not like that's that's not good.

I agree with that. But I I tend to think that they'll sort that out. Um, in terms of the general public's awareness,

no, no, I I do think that the the general public's perception though, I mean, I,

you know, my parents, I talk to them now every week about this stuff. Uh now it tends to be around data center buildouts you know and we sort of discussed that you know and there's somewhat of a political lens on it but

you know a lot of it is of an interest around [laughter]

but uh I'll try to stay away from but I mean I do think um

you know people are really interested about AI obviously and I do think you know the general public now has at least at this moment made up its mind that it's not a good thing. I mean, I can't quote the specific studies, but I've seen a lot of them where it's it's it's become a very negative perception, which is probably going to happen no matter what

time though, they're using it all the time that the average person, I'm not going to say the average person, but not necessarily all all across the board, but you hear just normal people saying that they're consulting Chad GPT for certain things, right? and it's it's prevalent.

Uh you're right. I would like to I don't know what the usage is like off the tip of my tongue, but I will say for me it's it's basically non-stop. [laughter]

Well, you're and I think I'm like a lighter user because I don't like Yeah.

You know, I do think there's people, you know, like for example, I like the chat GPT live that came out.

The chat GPT what? uh chat GPT live where it was um

you know I mean they put it into their app and just where you could have a continuous conversation with the

Yeah.

And I thought that was amazing and I'm not a big chat GPT user overall. I like Gemini. That's kind of my guy.

Oh, you Gemma.

Um but I thought the technology was amazing and I do think that there's a decent number of people that are now living in that movie her which I still haven't seen but I have a notion of. I mean, like how many people are just walking around

just talking to it all day? I mean, I think like I think I'm going to get there. It's probably it's it'll be next year for me.

Yeah.

But it's like it's it's amazing. Yeah.

It doesn't seem to cost anything,

you know? Uh I have a Gemini plan for coding, so I mean maybe it's just oh that like that's why I'm able to but it doesn't seem like you ever get a bill for just asking kind of silly questions, you know?

Right. Well, I have the $20 open AI thing, so I can talk to Chacha PT as much as I want. But it it just seems like it would be silly not to use it because it gives you such an advantage in so many different aspects of life because life is all about problem solving, right? Every day there's some kind of problem that you need to solve. And now it's just it's it's just amazing how much smarter it gotten. I I was with a friend the other day and she has this issue with the new construction that's going on at her house and she needs to know like who with the city council to contact and all that and I just gave the thing some context and it came back with so much information about what the regulations were blah blah blah who she needs to talk to and she was just like stunned she's like I I can't I mean it came back like that right and I am so used to it because I know that it can now with contacts context it will go and search all these sites for you that normally I mean I wouldn't even like know where to search to find out what the sidewalk ordinance is for a new uh development you know residential development and it did all that stuff and this this is like so much better than what it was even doing a year ago and and now it's like you have this incredible capability like right there in your phone why would you not use it like all the time now she just just with the information I gave her she's already like gotten so much better information about what they're doing around her house. And she's able to talk to the engineer that's out there. And so it's like it's a big deal for her because it was like how she was going to get access to her house after they built this development, right, man? It's like a hu it's like a superpower.

Yeah. She could probably sue the contractor if she wants uh using chat JPT as well if it gets to that point.

Yeah. Yeah. So, I don't know about the data center thing and the AI negativity. I'm just wondering if that's gonna if that's going to subside somewhat or change as people start to see the value of the the technology. I don't know. I mean, I do think the data center thing is problematic because people are scared about the environmental impacts and the health impacts. That that's understandable. I mean, I've heard a lot of people say, "Oh, it's because they don't trust the tech bros and the oligarchs and, you know, they don't like the Elon Musk and the Sam Altman and all that." And I'm not so sure it's so much that. I think it's more or or they're worried about AI taking their jobs. I just I just think people are scared for they don't want the noise and the pollution. And if that could be addressed, I think people would probably be more immunable to it. That's what I hear.

Maybe, but anything that affects property values won't be popular even a little bit. Yeah. Yeah. But if I mean if it if it was just like um I'm trying to think of an analogy just like a like let's say a normal data center that doesn't that just has some computers in it. I mean there's those all over the place that they're not in residential neighborhoods but they're they're around. You don't really even know that they're there. That's not going to negatively impact property value. So if if there w if there wasn't a lot of if there wasn't noise and there wasn't pollution and I wasn't using all the water and I wasn't making electricity be more expensive and there was nice landscaping around it and it was quiet why would it impact your property value right

I agree with all that so I guess well we should probably wrap up but I just want to bring it back to Alexa

uh that was our starting point

uh of uh when we first met And so now this you may not have any opinion on because I do think uh Alexa is the less vibrant part of the conversation but they do have Alexa plus now.

Yeah.

Uh Open AI launched a screenless portable you know home listening device with a microphone. I mean they like their description was like it pains not to call it a smart speaker.

Yeah. Well they announced it right. They didn't launch it.

Yeah. Sorry. Did I say la? Yeah. Yeah. And they didn't even technically announce it was more like a leak, but

I assumed it had some blessing from them on it based on the amount of information.

Yeah.

But it, you know, it sounded like an Echo Dot. So,

do you do you think the smart speaker can make a comeback? Can Alexa make a comeback? You know, and also to use a term that was popular a few years ago. I think we probably sat in a presentation maybe several about this. You know, ambient computing as as a big force in 2027, let's say.

I'm going to make a prediction that it can make a comeback. To be honest with you, I haven't thought that much about it because over the past few years, my modality interaction with AI has changed so much to be focused around the smartphone or the computer. Um, and I haven't tried Alexa Plus. I have a friend that uses it and I don't know how it would compare to JBG Live or whatever. But I think if the intelligence was at the same level that that our intelligence is, you know, when we talk to the the LLMs. Um, sure. I do think it would be nice to have a dedicated device that has some kind of even though it's just a puck. I mean, it's like an embodiment and you're talking to it. You can see it light up. I mean, a lot of that stuff that Amazon developed in terms of the interface was kind of cool.

It's just that the supporting technology for understanding was horrible, right? So, it'll be it'll be

Yeah, the NLU was

Yeah, it was bad. And if I don't know, I'd be interesting to see if OpenAI can pull it off. I I wonder how the Alexa Plus stuff is doing or if people just kind of abandon it or they still think of Alexa as a timer or something to play music. Do you have any idea like how successful it's been?

Well, I mean I was so Anderson was Scar the chief evangelist for Amazon Alexa, former chief evangelist for Bespoke. He see I mean he says the usage is still there. I mean I think you know obviously I think it's maybe not certainly the excitement around it has fallen off but the installed base is still enormous that is for sure and

and have they upgraded because you do you have to upgrade automatic I mean specifically or do you get it if you're prime

I think you get it automatically if it's prime I will say for me it wasn't accessible to me until I bought one of the newer devices because I have legions of the older ones around but not the newer ones so I had to get a newer one and then it was enabled on the phone on the web,

you know, that sort of unlocked it and I've been able to use it. I don't use it a lot

uh because I have moved on.

Is it using um Amazon's language models or is it use Yeah. Okay. So, it's probably

Yeah. And I think it's a comparable experience to the other ones, but it's like I And this is another thing I would just say. I mean, we're all like I'm so attached to Gemini.

Yeah. you know, uh, my wife likes chat GPT. I mean, I I most programmers I know, I mean, they you would have to pry claude from their cold dead hands. I mean,

right.

So, you know, and so people have gotten attached to their platforms, but I think there's still an opportunity there. And then I am I'm interested in other modalities, I guess. I mean, obviously we, you know, I you spend a lot of time on the contact center. Our business is on the contact center, but you know, I like the idea that of ambient computing. Um, I mean, I even read this morning somebody was saying like, oh, here's a great device. This was something that we used to spend a lot of time talking about like, you know, devices for people that have are vision impaired.

And somebody saying, oh, set up an openclaw server for somebody who's lost eyesight, which, you know, we have a family member that's lost eyesight. I was like, I mean, I should do that this weekend. I mean, I've got because I set up an OpenClaw server when that was a case. I haven't used it. I haven't found OpenClaw to be my thing, but I was like, but that's a great

here's a great use case.

Yeah, exactly.

Yeah, I'll put a speaker microphone on it and

I mean, they can interact, you know, with the with the LM and maybe I could hook it up so they can talk to our daughter and it would be really cool. And then you know like that's a different modality that um you know would seem to be like a smart speaker. Um and I have seen that Amazon for Alexa plus they are pushing Fire TV which I think I mean that's like right that's an amazing use case.

Yeah. Especially well um I was going to say for somebody that's lost their eyesight but that doesn't make sense because they're going to be able to see the TV. I think there's a there's there's a lot that can be done. And you also think about robots, too. I mean, the humanoids that have the language capability built in. Um, I've always wanted to be able to talk to the robot. I mean, we're getting close, right? We're getting really close to the Star Wars thing.

Have you seen any good new consumer robots? Well, that was that was a wave of companies that I saw come and go that I was really disappointed that none of them really made it.

Yeah. I mean, I still follow it a little bit because a friend of mine, she's um she's in the the aging, you know, age tech space looking to see whether we're we're going to finally get robots that can take care of us as we as we're getting older. And I I was thinking that they were going to be further along than they are. It's like every year I keep thinking, okay, we're going to get that much closer to having something that's capable. There was a company out of I think some Scandinavian company. I don't know if it was Denmark or Finland. What was it called? Um probably gonna say it wrong. I think it might have been called Neo. And you could actually order it.

But the thing about it and it's a humanoid robot with like a knitted fabric thing covering its body and it it looks

I got to look that up. That sounds amazing.

Yeah, it's I've tried that's not the name of the company though. I can't think of a name of the company. right off the top of my head unfortunately, but it is a Scandinavian company and you could buy it and I'm trying to remember you could lease it for maybe $600 700 a month or something or you could buy one for like 20 grand. The pro the problem with it was that it had been trained on certain tasks like going to open the door, but you had to then have it in your home and then phone in to one of their um remote uh what's it called when you you know remote control technicians and and it and it would remote control the robot in your home to do the task that you wanted to do. like it it would they would have to like put their goggles on, be in your house, control the robot remotely, and it would know how to go and open your dishwasher, and then put the stuff into your dishwasher, close your dishwasher, turn your dishwasher on, or it would teach it how to do the laundry. So, it was it was very, let's just say, out of the box, it didn't really do anything. And I haven't heard anybody say whether it works or not. Um, but it seems like there are technologies now where you can teach it to do a lot more stuff and and and it can then extrapolate how it's going to do stuff in the real world without you having like specifically train it on every single task. But I mean, we're still like a long way away from getting to the point where I can just order up a robot that can come to my house that can make my bed, make my coffee, do whatever I want it to do, clean up.

Yeah. But it's got to come. It's got to come. It's going to happen. Just what that that last little bit is going to take. I I I don't know. Um but they say if they can train themselves and it can learn stuff from by watching videos, then maybe it's going to get here quicker than we think.

Uh I would be optimistic about it. And I think uh to tie it back to Alexa again, I I think that's an area where Amazon, you know, they've they've shown a desire to innovate. You know, they do the drones. They had the Astro Astro one more puzzling. I mean, not that it was really cool, but it was like

why like what is the market for this? Um, but uh, you know, it seems like they have the appetite to do it, but I would be optimistic that, you know, that's a way they can get back in the game because they've got a good technology there. It's just

that you should make a prediction. Do you predict that they're going to they're going to do a a robot? Because I [laughter] would I'm I'm not thinking that.

Seems like they don't have that same ambition that they had. And it

I Yeah, that actually I I don't know. It frustrates that to me. It's like they're they were just like invest invest invest nobody really wanted it. It didn't work

because I like with Alexa I see this with Lex I see it even with connect I feel like their their strategy to me is like just so hard to understand and I think they'll do great with connect but it just it just seems like like why don't you guys you know it seems like there's such an obvious path for me and they don't seem to have taken it. Um, so no, I don't assume that they'll do it and and I don't assume they'll do what I think is the most obvious thing for Amazon to do being the ambitious company that they are and with the, you know, kind of financial resources that they have, which is to build a darn phone. I mean, rather than and and it's like with chat GPT, I have to say it's the same thing like I mean, you're going to build like an Echo Dot device. It's not a bad idea.

It could do well. It'll be interesting. I I would consider buying one, but

you know, build a phone. I mean, that's the conduit. And um

I guess there have been so many sad stories with that companies that have tried to do it and it's failed.

I I think Amazon could totally do it.

I don't know there that's got to be hard. It must be harder than we think it is because

I don't think it's that hard. I think that Well, it's not that hard if you're willing to accept paying Google for Android. I think that's the sticking point. That's the conclusion I've come to. I don't think I mean the engineers at Amazon I think could create a phone in a weekend. Uh and I'm sure they have pro you know they don't even need a weekend because they've already done it many times based on Android but to do it with a new operating system and they did one of the Fire TVs I think it's the stick like the lowest end one it runs on a new operating system that Amazon built

really

and I thought well this is the precursor to the phone but that hasn't been that announcement. The hard part is and and this happened with the Firephone which was the precursor to

Yeah. Fire phone, right? Yeah.

Yeah. And part of the reason it bombed was it was Android based, but I guess the part where you have to start paying Google or at least one of the parts where you have to start paying Google is if you use Google Play and use the app marketplace.

Okay.

And they didn't want to do that. So there were no apps for it. I mean, this is a whole another conversation and and uh but do you maybe they're thinking that the phone modality or the phone, let's say, form factor

is not going to be around for another five years. Of course, everybody's been saying that for a long time, right? And it hasn't panned out yet. But I'm wondering if now may that might be true.

I mean, it's a good thing. I don't think any, you know, there's no sacred cows. There's you can't take anything to be a truth, I think, at this point. I mean, everything has changed so much.

Yeah. But it also goes to I mean this isn't a new production on my part, but it's like I've read and it's I think it's probably a bold one except for the fact that I've read over and over in so many places about how stupid Apple has been and and that they have totally fumbled the ball on on AI which I I mean they lost a class action lawsuit for about half a billion uh because of false promises or promises they didn't fulfill but they are going to be such a winner on AI because it's This is the number one channel people want to use it on. I mean, it's just once they launch it well on here.

No, I they're going to blow away everybody, I think, in terms of revenue. I don't even think it's going to be close.

Apple.

Um, yeah, because it's going to be people just I mean, what percentage of people walk around all day? I mean, when you're in public, everyone is wearing these things all day if you go out.

Yeah, but just Well, wait a second, though. Wait a second. If you think about the new open AI device, which is you're you can only talk to it. It's it's it's the original Echo idea, right? The original Echo Dot idea. I'm just thinking about this now because how how many apps do you use anymore? Before you answer that question, if you can imagine having your your favorite AI, your Gemma or whatever, you're talking to it. you want to find out some information. The only reason why you go to an app is because you're looking something. I'm I'm looking, you know, is it gonna rain? I look at the weather app. Um I look at I don't know what's the price of gold or something, right? I I look at or the stock app. I I look at um you know, my bank app, but I I could just ask Jima, my favorite assistant, all the stuff. I wouldn't even like need apps anymore.

Uh, I think that that's true, but I I do what I do still use a ton of apps. Uh, there I basically try to track everything I do all day. So, I have an app, you know, like any movie I watch, any album I listen to, any TV show I watch.

Okay. But can you imagine just telling Wouldn't it be easier to just tell Jim, okay, I just saw that movie. Um, let's chat about the movie a little bit because I want to ask you about the ending. It was really weird. and and like make this note for me about the movie and then just store it to my database and you know to and tomorrow I want to like go to my desktop and I want to see uh a list of all the movies that I've seen in the past or send me a list of all the movies I watched the past year or something. I mean wouldn't that be like to so much easier?

Uh yeah, I think that'll probably happen but it's definitely not available today. I mean it's you know I mean unless I voted that

no but I think that's just because we're in this paradigm of the you know we're still in this paradigm. I mean it's at at some point the paradigm is going to shift and once it shifts then we might not you mean think about the smartphone it's not like it has always ex existed right it existed somebody came up with the idea of making you making all these apps available from this screen that you also use to call your mom every once in a while right so okay that it's not like it has to be there forever

uh I I think I think there's a good chance you're right about that. I find it a little hard to conceive of. And even if it is, I just, you know, like one of the things Apple was supposedly the next version, the AirPods, I thought this was fascinating. They put a camera in there. So rather than going because everyone finds it super creepy that you have glasses that are showing you stuff. Instead, they put the camera in the AirPods, which also could seemingly be creepy, but it would be, you know, the AirPod is the center of AI. And yeah, maybe the phone becomes less important, but you're still then looking at, you know, this is your text output that you're getting on your phone.

Um, and you're, you know, as long as you're talking or whatever. I mean,

yeah. I mean, cuz a lot of the stuff, oh, you know, tell me about that movie, tell me about that actor, you know, what was their name? And, you know, I'm talking to here and then it pops up the photo. I mean, that would be a natural thing.

I just think in all these Well, I was going to say in all these scenarios, Apple profits so much better than anybody else because no one else has this delivery channel. I mean, nobody else is even close. Samsung.

Yeah, but they're making a lot of money off the app store, right? So, if they if apps go away and and there's just a there's a different way to access information, I'm not exactly sure what that would be, but the the the agent, you know, you just have agent to agent communication and it's going to go get the information. Um, and the apps go away. You don't need that anymore. And the other thing about it is this kind of like this phone anxiety that everybody has. Like I was on the plane the other day and this girl that sat down next to me, she just like misplaced her phone for like two seconds as I mean it was there like in the pocket in the seat in front of her. But I mean I know we can all feel exactly what she felt like suddenly her phone wasn't there. Like oh my god what [laughter] happened to my phone and she's like totally panicking. We were like wait a second we saw you had your phone like wouldn't it be nice if we didn't have that worry anymore? I mean, I guess you'd still have something that you would have to wear, but

I got worried yesterday that I had lost my AirPods until I realized that I was

Yeah, we need something like implanted [laughter] in our brain. I guess you can't

Well, Elon Musk is working on that,

right?

The neural um

something you can't misplace. Well, I so the prediction I would make related to Amazon though is uh and this is I think is lower probability but I think it's the most sensible thing is I do think they should build a phone because I do think that that's you need some sort of processing brick to carry around with you and no one else you know Apple's basically run away the market. Samsung can't do software. Google can do all of it. Pixel is amazing, but they just like, you know, I'm in Peru. I can't buy Pixel View. You know, they just don't, you know, they don't. It's more to impress each other is always my feeling about Google than to actually like, you know, build a business unless it happens to be search and advertising.

But would you want Amazon to recreate the app store then? I mean, they they'd have their own app store or or would they do?

I think that that's that's why they don't do it is that they uh and that's probably why it's not a good prediction. But I still think they should. I think it's the type of ambitious thing that Amazon excels at and it's I mean there's not that many opportunities of that magnitude out there. I don't agree because I think we are in the midst of a fundamental shift in how we interact with information and we don't really know how it's going to come out, what the next form factor is going to be. I I think that was it's why it was interesting that OpenAI got together with Johnny Ives and they were trying to decide all right what what was it going to be and a lot of people predicted it was going to be a pen

or something like that and then they came up with the puck and again I don't even know if we know 100% that that's what it is right I mean did they say that's what it was okay

um

well I would just I would be I I mean I do think you're correct I think the timeline matters though because I like the fact that they didn't do a pin I'm sure they were thinking about that. I mean, to me, that's if I That seems like the obvious open AI thing to do.

Or glasses.

Yeah, the glasses, you know. Do you think glasses are going to happen?

I don't really know. I mean, I know that, you know, Snap Chat is is really big into that. And then, of course, Meta is betting on that, too. A lot of people say no because most people don't want to wear glasses all the time. If I if I had to make a bet right now, I would probably say I don't think that's going to be the big form factor. I I I don't know what it's going to be. Um but I would like to make a prediction that the phone will be gone by X number of years. I I think that would be an interesting an interesting prediction um or kind of a bet to make. You know, I you think it's going to be

So what year would you do? I would I think it'll be less useful by say 2032, but that's a pretty long way off.

Yeah, I would definitely say 2030 at I would come up some on that. But since we don't know what's going to replace it, it's it's a bit of a crapshoot, right? In terms of making that the guess.

Yeah. Open people thinking about it. Oh, you I mean

Oh, yeah. No, no. I mean, they're pouring money into it. I mean and if open a but if open AI was putting out a pin I would think oh well this will be V1 and V2 in 2028 will be even better and that future will seem more tangible but you know because whatever they put out next year it's only going to be the starting point

you know and it'll be V2 that gets any adoption. Yeah. I mean, the problem with many of these devices is that they kind of rely on your phone, too, being there. Like, like if you had a pen and it had a microphone, I mean, you wouldn't like like want to talk to this thing. You'd have you'd have to have your I don't know, your headset in or whatever to hear what it was saying.

I'm not sure. I'm not sure how

well I think that's all rumors or you know everything I'm saying is just rumors but what I had heard was that it was the concept of the pin was what you know because I at some point I thought well maybe this is just face saving cuz the article I read was like I mean they decided not to do the pen I'm not sure what the reason is but it wasn't as dumb as you thought it had a headset and and I was like

I didn't hear that in the first set of news stories but it was like after I think it came with the announcement of the puck It was like, yeah, like we we aren't that stupid. We weren't gonna just put a pin you couldn't or that you had to be shouting at and like, you know, holding up to your ear. It was gonna be a pin with a headset, which

you know, I mean, really to me then the headline is,

you know, it's it's chat GPT pods.

Yeah.

Which um

I I mean,

I do think is a great idea. I I think something like glasses could work if there was um a I don't know if it would require a breakthrough in technology, but if you if you had glasses and then you did not have to wear uh anything in your ears because the sound coming through the what are the things called on your the sides of your glasses.

Yeah.

Um

I know what you're talking about.

Yeah. If it if it like was you know like in science fiction you can say stuff like basically don't even and it'll knows what you're saying. And I think that technology is getting pretty close. So maybe there's something where it's very soft, but it goes in directly into your hearing. And that way you're still what I think would be very beneficial to society is if people weren't walking around with stuff in their ears and then that would invite when you sit down on plane or whatever it invite you to interact with the person next to you and you might still have this thing that you're you know your assistant is on there that you can talk to but it's much less invasive and so maybe it's going to require some new technologies to make that to work. To me, that would be the ideal thing.

That make sense?

Yeah. I I I love the idea of glasses. I will say the the creepiness of it is is where I kind of get caught up. And I kind of hope they won't be a thing.

Um for that reason. But I also, if I had to bet on it, I do think that they'll win out because I think the practical applications are just too many. Um

Well, as long as we don't have like camera here like we used to have, right? [laughter] Well, yeah. I mean, I guess that's what I mean. But I The reason why I think it'll have went out is I mean, you're already like I mean, everywhere you go, I mean, there's I mean, first of all, walking down the street, you're probably being recorded. I mean, I don't You certainly could be. I mean, there's always someone walking toward some somebody walking towards you with a phone in their hand. Uh, I don't think they're recording me. They have better things to do, but they could be. Yeah.

Um, you know, and so none of us has any, you know, you can't assume any interaction you have is private anymore.

Yeah.

So, I tend to think that the classes will be accepted, but it is um it's probably a regrettable thing.

We'll see where it goes.

It will be an interesting It will be an interesting few few years, if not longer.

Uh, I won't I won't force you to do these. uh [laughter] I mean we'd like to catch up uh semi-regularly. So perhaps we could do this again in public. If not, we could do it in private. But at least this way we have So now we've got it captured. So we can compare uh all the very uh informed scientific predictions we made [laughter]

and see how they'd actually see how they'd come about.

Okay.

Over the next couple years. Um and uh yeah, I I think it'll be fun. Well, it's good talking to you.

Yeah, it was great talking to you, Amy. Uh, thank you for joining us and, uh, yeah, uh, I guess, uh, look forward to talking to you again soon.

Sounds good. All right, take it easy.