Google's AI Opt-Out for Publishers and Higher-Spending AI Shoppers

0:01Hello everyone. Welcome to the third episode of Marketing to Agents. And as always, I have Lalit with me who is of course actively looking at what's happening in the industry. And today we have some exciting things which happened over this week to discuss and share our thoughts on. Welcome, Lalit.

0:23Hey Ashwani, happy to be here. It has been quite an eventful week. so happy to get started.

0:30Fantastic, Lalit. All right. As Lalit, yep. This has been a great week for updates and things happening in in the agent space. So why don't Lalit let me just get started immediately with the first one which is we heard that Google Search Console AI opt-out toggle will go live this week. And this has been enforced by UK regulators. Now again, of course, as with everything, there are both pros and cons of this. To my mind, Lalit, the first thing which comes out is that you know, having publishers or having websites or having brands voluntarily be out of AI searches or AI overview AI mode, I don't know if because I don't know why it would be it would make sense. Of course, it is in a way to protect publishers, but at the same time also, if you're not getting visibility there, then how do people discover you if they are searching? So it's a mixed bag, but I would love to hear your thoughts around this.

1:35Yeah. I think this is also very important because it is the first legally enforceable like a first-party opt-out by a publisher, right? Because nowhere ever in the world such a thing has ever been, let's say, brought out. So we'll we'll have to wait and watch how how actually, you know, the what are the impacts of this? But if I were a publisher and if I have, let's say, something very let's something very important, let's say, there is a lot of human effort goes into coming up with, let's say, an analysis or let's say, a research report, right? And I've already established enough distribution and brand for myself. And this could be a like a traditional publisher like an economist or a Harvard or like a a very well-known let's say, a business review you know, magazine. They anyway have a lot of pull and the fact that they might want to opt out from this is that when their audience is, let's say, querying on Google, right? And if they know that their trusted publication is not a part of the AI answers, they would have to, you know, come back to the publisher's website and then maybe use AI on that website, right? So, I think it makes sense for like already established publisher

2:55Brands. new ones definitely should be thinking twice about it. but I can see, you know, there there is a case for publishers to enforce you know, that their audience can have a much better experience on their own portal in their own, let's say, walled garden experience. So, that's it. but I think along with this, I think Google Search Console also has started to give the AI search performance you know, insights. That is again a very like a welcoming update because prior to this, I think most of the marketers were sort of looking at, you know, visibility tools which are giving like a some sort of an approximate visibility and an and a very empirical search volume analysis on various prompts. So, it was a a very very, let's say, approximate sense of their performance on AI searches. So, this change I think will give marketers like a real view into how is AI search actually happening, what searches have high volume, what kind of prompts are people actually, you know, providing in their domain. So, that's a that's I would say a welcoming update.

4:09Perfect. So, let's let's break it down into two parts and would love to dive deeper into the first part that you said. So, let's talk about the publishers now. So, for example, like couple of questions here from what you explained. number one is why did this never happen in the organic search era? So, let's say in the world of SEO, why did this never happen? Where why were there no why did regulators never enforce that, okay, I don't want to be ranked in organic search or I don't want Google to, you know, even even crawl my content. So, just block the crawlers. Number one. And number two question related to this is that why do they need to opt out? Like I'm I'm under the assumption that if there is a paywall to their content, like Economist or New York Times, most of their content is behind a paywall. So, doesn't that automatically stop their content being shown in organic or in AI overviews?

5:09Yeah, so great two points there. So, first is like in the erstwhile era of search, the the decision of information consumption was still left to the user, right? So, the question in the user's mind was very in a very lossy fashion, you know, expressed in terms of keywords. and the user would have to read the the whole content on the publisher's website, which although would have been listed in the search results page, but the answer would not be given, right? So, answer would still be obtained by the user when they're consuming the content. So, that that would that would only happen when there is a click and then the the user actually gets into the publications website whether they are paid or fully monetized through ads. That's a separate question, but the publisher could still rank as well as expect that that the user will the user relationship is controlled by them. The user experience and the relationship is controlled by them. So, that has changed with AI. And your second point being that they could just paywall or limit the content. But, I think what is also happening is that there is a lot of cross cross-referencing that keeps on

6:25Happening. Right? Where let's say a a tertiary article is basically referring a primary source. And that and you know, some of it some of the content is already out there. And the primary source, which is the publisher the entire content is not paywalled. Some of the content is still visible and then maybe AI can hallucinate the rest of it. Right? So, there are things which you know, which can still result into the answer being synthesized and the publisher being cited. And the you know, and the user still getting a sense that this answer is synthesized with this publisher citation. So, this is something that they can take at the the face value of the publisher's trust. Right? So, people who have you know, a very high regard of their hard work of you know, journalism and let's say you know, editorial effort. And they already have established brands. They they if they opt out, then I think the AI overview answers or AI mode answers will not be able to cite that website. Right? So, if my most preferred publisher is not even cited. Right? Then I still have this fear of missing out. Right? As a user. Right? So, I would although I might be a little bit more aware after I read the AI mode answer,

7:43But the fear of missing out would still make me go to the preferred publisher's website and then use the AI there.

7:51Understood. But but just building on top of exactly what you said. So, let's talk about a slightly more now a future state or in slightly extreme scenario. Now, let's say it has been 2 years since this and since this regulation came out and let's say top 20 publishers or top 50 publishers have actually, you know, decided to opt out of this. So, does that in some way then lower the credibility of AI answers in general or AI mode in general? Because if AI cannot access if the search engine if the LLMs cannot even access main publishers or key publisher content like authority which are let's say real authority on that matter. So, for example, let's say Harvard, you know, medical school or Stanford these take away their content from AI mode, then does it really then does it reduce the overall, you know, trustability of the AI mode or the, you know, AI published answers? You see that happening let's say after 5 years if there are tons of publishers who take away and they wait for only exclusive deals or licensing deals like we discussed in our last episode, then what's the future of an AI mode in in general? If this really becomes a trend and it's really adopted aggressively by

9:09Publishers.

9:10Yeah. I think see there there would be certain type of information which is highly, let's say, gate gate kept. like extreme cutting-edge research and something which is not not, let's say, accessible in public domain as something like let's say an event which is not happening for public to which is of public interest and which is not being covered by popular media. And if it is a very niche domain and a certain publication maybe funds a lot of research and gets a lot of information, right? So, and if they end up keeping the information, you know, to only private to them by opting in by opting out, then I think for those domains, the user will it will be hard for the user to trust the AI review, right? Or the AI AI mode answer, right? Because there might be tertiary, let's say people who consume that primary source and then they maybe, you know, o- opinionated publish their opinions and publish their maybe like a their own synthesis, right? So, but still people can make mistakes, AI can also make mistakes, right? So, for those domains, I think the I'm assuming that the publication also has enough an ability to distribute and maybe build

10:27Their own audience, right? So, those people would rather prefer a similar experience on that publication website or their their internal thing. But, obviously, it has to be AI-driven answers. but they would not go for let's say a an AI mode, right?

10:45Generic, yeah.

10:46In in in our own, let's say day-to-day usage, we can we can see glimpses of it, like glimpses of this behavior, right? there are many popular car review websites, right? Where any new car there are there is an enthusiastic community. They they have like rich reviews about new cars, old cars. And this is an important decision, right? So, people want to go deep into those reviews and then make a purchase decision, right? So, and compared compare someone, let's say user A goes to the the the the car review destination portal, and then they come back they come out, let's say emerge and fully aware about what to buy, all aspects covered. And compare that user with, let's say, another user who directly goes to an AI mode and just, you know, gives car A versus car B as a comparison. And those sources, they probably do not cover the entire thing or maybe hallucinate or it might have also, you know, some sources which are not trustworthy, right? So, the user experience might be different, right? And as more and more people get to know about this high-quality, you know, information source, they might they might actually then directly start preferring going to that website and using AI there, right? That's what I'm thinking.

12:03That the FOMO, the fear of missing out, will will still work in favor of the publications provided the publications have enough brand awareness.

12:13Understood. Yep. I agree with I think my view also resonates, here where it's the I think this regulation is a welcome change. It gives power back some power some control back to publishers. And for really high-quality publishers and who are working in a niche or who have very, who control a niche really well, I think they'll have some more control with this. And then, of course, like we discussed in our last episode, they'll always have an opportunity to go into licensing partnerships with these LLMs so that at least their content is used and they get, compensated for it if, LLMs are using, are building off their off their content.

12:54Yep. Absolutely.

12:55Yep.

12:56So, let's move on to our next item in our docket today. so Adobe Analytics published some data, very interesting for, some of our listeners and the brands that we track. the data was that AI referred, shoppers now generate 53% more revenue per visit than other channels. So, if you're a brand, if you're an e-commerce brand, or you're a D2C brand, and you have a mix of traffic coming onto your website, the traffic which is from AI engines or LLMs or traffic coming from those that channel actually earns 53% more revenue for you rather than, you know, organic, direct, or any other channel that you were on. So, yeah, this finding seems to be number one is, of course, significantly shifted from what the behavior used to be last year. So, now people are relying more on AI search and then consuming once they have zeroed down on a brand or on a product, I think they're spending more welcome to spend more there. So, that's what we are observing. This directly, of course, has an impact on what we have been talking about around GEO, AEO investments, and brands being able to

14:13Being taking that lot more seriously in showcasing their product catalogs and presenting their catalogs to AI searches.

14:24Yep. So, Ashwani, can you hear me?

14:29Yeah. Yeah, yeah.

14:31I think I think this is like what we have been observing with many of our customers, right, at Airpals. This insight from Adobe actually just confirms that, right? Now, what we are witnessing here is that AI is emerging out as a trusted intermediary for non like internet users, right? So, they no longer want to directly go to information sources or websites, and you know, have the patience to browse through the information that they need. This intermediary is earning their trust increasingly with every passing day with its with the increase in its intelligence, with the increase in its capabilities. And this intermediary is now actually in some sense, the the primary source of, like, you know, what products people buy, what you know, like how do they how do they even, you know, let's say evaluate products and you know, ultimately make an opinion on something, right? The the whole demand generation and branding piece, that is still separate. I think there are different mediums where users go with a more browse intent like or just maybe like a feed intent where they want to consume what has been sort of offered to them in a newsfeed

15:50Kind of an experience like Twitter and various other destinations. That is where I think the demand generation can happen, but once let's say someone is seriously looking for something, right? I think this trusted intermediary has a very important role to play, right? And what we also observed is that the the memory with which the you know, the your longitudinal interactions with these AI engines tell a lot about you to these AI engines, which is then utilized to make you know, recommendations and your search experience or rather your AI experience much better, right? So, it already knows so much about you that, you know, the recommendation that that it ends up giving, it is of such high relevance to you that when you visit the website, right? You you have very very you know, a very rare chance you might, you know, not like that recommendation, right? So, that's what we are observing, you know, when this this kind of stats come out.

16:52Absolutely. So, I think I think you mentioned it at the at the very start. I think thing which something which changed in my mind after listening to this was that hearing about this data point was that AI is no longer just a search like people are not just searching or using it to validate something or know an answer, but now actually they are starting and even to some to a large extent influencing their buyer journey through AI. Like even even in e-commerce buying journey. I think earlier we were thinking about businesses, B2B buyers journey, but I think now AI is playing a key role and of course discovery of your brand, but even convincing a buyer to, you know, how your brand is maybe superior over others or how some other brand is superior. And then when the buyers finally coming convinced to a particular brand's website and they're a lot more confident in spending much more money. So I think AI search has become a critical piece in actually a buyer experience also now. And I think this shift will will probably continue more because just that with so many products out there, with so much so many web search out there, I think it will automatically this is right

18:09Workload for AI to take up where they actually reference the answers, where they showcase more things and they're able to build better comparisons, on-the-fly comparisons. Like you said, personalized to me or personalized to a particular user and thereby direct them to a right right query.

18:27Yeah.

18:27So follow up on this, Lalit, is so if this is the behavior again, follow up on this would be does that change how an e-commerce buying experience happens? So does that change how people let's say biggest e-commerce site, Amazon. So if today people go on Amazon and there are hundreds of their ads of certain brands. If I'm searching for let's say a a category, let's say shoes or mattresses. There are of course Amazon has its own ad inventory there and of course there are number of different brands. Could be some could be private labels, some could be their some could be different smaller, bigger brands. Now if that is getting done by AI or some of that search is happening, some of that buying decision is being made by AI that means and that is happening outside Amazon, let's say. How does that then impact couple of things for a large e-commerce large e-commerce portal, for example? Number one is their ad inventories, how useful are those? Number two is discovery of different brands then on that website. Now, because the discovery is happening outside Amazon, does that change how I as a small brand promote myself? So, I think it does it would change certain paradigms for e-commerce. Like, what do you like where do you see that going?

19:44Yeah, 100% I think so all all intermediary digital players, right? Or aggregators or, you know, websites which are actually made a living by just aggregating supply, right? Whether it is e-commerce and Amazon or it could be, let's say, a travel category like booking.com, etc. So, they would have to look for more, let's say, adjacent value proposition. And if you if you if you re-look at what Amazon did very recently, they actually opened up their entire logistics infrastructure as a service to inde- to independent sellers. That was I would say it was a step done very much with an outlook of their reducing relevance in the discovery, you know, in in the changing discovery paradigm, right? So, if consumers can discover what to buy or even directly when when they are in you know, in chat with our GPT about, let's say, the couple of alternatives of a category that they are evaluating. Could be shoes, could be, let's say, water bottles, could be televisions, right? AI can directly hit the information source. It could be a Shopify website or it could be a direct the seller website, right? And it no longer needs the Amazon portal or because Amazon

21:01When compared to the original the seller's website always had inferior quality information, right? Amazon was always like a like it was like a like an a checkout in you know infrastructure where if I already know what I want to buy, searching for it on Amazon and then buying it and the trust that comes along with, you know, the buying process in fact the checkout process, the refunds guarantee, the on-time delivery, I think those were the core promises. And as the discovery gets disintermediated, right? And when individual sellers are directly products and sellers are getting directly recommended by AI engines and and AI engines are also doing a great job in terms of onboarding this e-commerce not e-commerce, let's say e-commerce sellers by allowing them to upload their catalog, integrate deeper partnership with players like Shopify. So, Amazon obviously has a very high risk of getting you know, reduced relevance or or I would say diminishing relevance in the future and that's why they had to sort of open up their logistics infrastructure so that they can continue to take a like a pie of the e-commerce value chain, right? So, that that we're going to see in various other, let's say, domains as well, right? So,

22:18That's what I would say.

22:22Perfect. And let's you haven't or like I haven't come across anything which happens today with AI on Amazon's website, right? None of these large e-commerce players I've seen they have in they have brought in AI in helping buyers make a decision, right? Like I have not come across anything. Have you seen anything which let's say Amazon or you know, any big Best Buy or any of these big retailers have they have you seen anything they've done in their core online shopping experience which is enabling AI for buyers journey?

22:58I think they have they have they have been attempts. Like Amazon has Rufus, which which obviously promises a better discovery, but I've never had a very satisfactory experience with Rufus, right? There was a time when I was comparing vacuum cleaners. Very lately I was also comparing air purifiers, and didn't have a very satisfactory experience. I had the best experience when I went to, you know, these AI engines like ChatGPT, right? The you know, two reasons why I think it's not happening is one is that they obviously had inferior product information. Sellers were anyway considering them as, let's say, checkout infrastructure. And and had, you know, their product listings are also bare minimum information, right? Obviously reviews are there. Reviews are helpful. Which but reviews are also available on the seller website, right? And that is the first thing where the information quality itself different than Amazon. The second thing is, I think, the UX, right? The UX is not very optimized for, let's say, a discovery let's say, you know, kind of an experience, right? The And so far these products haven't never made to the

24:16The discovery, let's say, part of the journey, right? They've always been Discovery used to happen somewhere else, and then when when you come to Amazon, you should be clear about what you buy, right? So, that muscle memory still sort of makes it more frictionful to do discovery on Amazon.

24:36Understood. Yeah, I think that the That's a good point, and I think we'll see probably a lot of work happening in this space, and especially I think for brands, I think this is a great opportunity where earlier they had to fight for user attention in an Amazon world. Now, I think with the search blowing it up and AI search bringing more more discovery there. I think if newer brands, smaller brands can also attempt to invest higher in GEO AO and thereby driving more user behavior towards and user convincing towards their towards their products. Perfect. Like the next item on our docket is around this whole open-source shift to open source. So we've we discussed in one of our earlier episodes about how there are companies which are now moving to open source. And they are instead they're not using cloud anymore or open AI just because of the of the of how expensive it has become now. The token cost and therefore they are now moving some of their workload to open source. Some companies have completely shifted to open source. Now of course this is not a consumer level change. This is not me going and searching on Deep Sea or

25:55Or some other platform. This is This is a business let's say a business that I use is using now is now using an open-source model for their AI workload. So do you see that having any impact on how how me as a user or how me as a brand will have to adapt to that shift? Let's say if a lot of this you know AI traffic moves over to open source. Then do you think that shift making any difference for a consumer or a smaller brand or a B2B company?

26:33Yeah. So let's unpack this a little bit more. so you're absolutely right. I think as AI gains more maturity in terms of their workloads, right? some workloads will will will be very satisfactorily delivered by open-source level intelligence, right? And there will be some workload which requires frontier intelligence, right? And as open source models continue to progress their intelligence, we will see that a good chunk of even consumer use cases are being delivered through open source, not maybe directly consumers interacting with an open source platform, but they might be interacting with let's say a Manos agent or a Hermes agent or let's say, you know, a Siri, right? So, and could be any other any So, that that surface area would be a lot more fragmented than it is today. Today it is, you know, like top two top two three players or largely, you know, the the top consumer companies. But as we see inference getting localized, local models getting better, more and more, let's say, you know, experiments happening on, let's say,

27:49Wearables. So, we'll see that the the the first touch point with AI or what we call as AI assistant, that surface area will also have lot more lot of fragmentation, right? And open source models might have a very strong, let's say, part to play in terms of delivering intelligence on that, you know, fragmented surface area. So, when it comes to AI search and discovery of, let's say, products and vendors and services, I think what this brings the attention back to is the quality of the website optimization, right? Because there's no one search index that you might you know, you can you can rely on, right? some of the search engines have made it very difficult for developers to access it, like Google Google Google search index is no no longer available for developers. Bing is also putting restrictions, right? So, newer and newer search players will come up, right? we have seen companies like Parallel, Xbar, Firecrawl, and then Brave. There might be few more. There might be a few Chinese ones also, right? So, so I think people would optimize behind the scenes for their, let's say, cost. and we will see a very fragmented consumer touchpoint as well as the

29:07Search index that powers it. So, the only control that a brand or a business has is on their own website, right? And making sure that the website is in general very well optimized, has a lot of information, and is available for, like, all search index builders. Anybody and everybody who's trying to build a search index or any type of directory, the website should be welcoming for all of it, right? So, I think these are the things that, you know, are going to be more important when it comes to AI search discovery in a in a more fragmented, fractured consumer surface area.

29:48I really like I think that's a that's a great viewpoint. I as I was researching more about this whole shift to open source, one more thing stuck to me that with the recent with the recent announcement from the US government where they are banning the use of Fable 5 cloud model to to other entities, non-US entities and governments. Now, if that is if that ban persists or if there is so-called weaponization of AI which starts happening where the latest models or the most advanced models are only made accessible to a certain country or certain set of companies, what does that in that sense then? How does that change the demographic then? Then do every country starts building their own models or relying on their own infrastructure? Because tomorrow no country would want to rely on something which can be stopped by another another entity, right? So, what are your thoughts around that? Again, I know this is slightly away from open source, but then just because of the nature of fragmentation, does this also add another layer of fragmentation into how the AI agents or AI searches will behave eventually on the internet? What do you think about it?

31:07Yeah, I think the that one event that you're talking about, right? Like the US government, you know, enforcing a ban on Fable 5 or Claude Fable, I think that is going to accelerate the the you know, a lot of these sovereign AI efforts, right? Because what people will start to realize is that they they have to really invest in their own, let's say, native AI infrastructure, because in AI the way things are drastically shifting towards AI, right? All workload, knowledge work, information access, right? It is becoming, as I said, a very crucial part of all, let's say, information pipeline, right? Whether generation of information or consumption of information, which is the lifeline for, let's say, modern economies, right? So, you don't want you you obviously would want to de-risk or as much as possible control every piece of the pipeline, right? So, if somehow that one piece stops working, the entire economy can come to a halt, right? So, I would say it is you know, it's a it's a positive event with respect to the acceleration

32:24That is going to provide to the sovereign AI efforts. it also makes it very urgent for, let's say, independent AI labs to join forces and maybe create, let's say, a more let's say, a global level you know, like a research lab where no one country owns it and it sort of belongs to the people of the world, right? and so, you know, that that frontier intelligence should be should belong to humanity, right? So, there we will also see, let's say, some labs which are just like there is a security group called the Anonymous, right? I think so there there also might be efforts where some really talented researchers in a very anonymized way way, right? Might maybe pull in their resources and, you know, like the decentralized intelligence effort, right? Which which actually has been not getting too much attention, too much oxygen, so to say. That actually might actually get a lot of push, right? So, decentralized intelligence which is basically a paradigm where people can contribute their compute, com- contribute their data sets, and researchers in this eco- in this decentralized ecosystem can work on frontier, you know, can work to build frontier intelligence. So, so I think

33:43That that will also get some impetus.

33:47Fantastic, yeah. I think you're right. This is a net positive for the world where where at least there'll be more homegrown talent, there'll be more sovereign AI investments which will happen. but for the users, like I'm still not 100% sure. So, if if I as a user will Do you Do you see in future since like countries mandating that if I'm a user in any the UK or in Europe somewhere, they mandate that a European that my search query or my AI search query should only go to a model which is like locally deployed where the search query does not go out, where the data is kept in the U- in in UK or in the Europe. So, do you see even those kind of controls coming in which might really limit you know, limit again the the ability of AI search to, you know, actually traverse or do a lot of things. Do you see that also paradigm this getting involving into that as well? Or it won't have a material impact on how things are how the behavior will be the users.

34:56Yeah, I I can definitely see more niche domains like medical medical research sorry, medical AI medical queries or let's say queries related to financial markets, right? So, those are domains where which are different from general search, right? Where there might be a better case to have more regulations in terms of privacy retaining the data. yeah, that's what you know, I would say This is a very new domain and I think people are quite paranoid about their privacy and access to data. So, somewhere I think progressive countries would have to subsidize or incentivize local AI usage, right? So, you know, closer home in India as well what we are what we what we should be seeing so seeing soon is that a lot of these government contracts, right? So, if someone is building solutions for the government, right? So, government would probably give a portion of the you know, the money for the contract in credits for the sovereign AI, right? For example, let's say if there is a model that the government is supporting, right? So, it some of that would be in

36:13In in kind, right? in in terms of credits so that so that the local AI consumption development is given more impetus.

36:27Perfect. I agree with and I think like I think you you said it right here. I think this topic is the the the topic is evolving so fast and there is one time one side where there is so much opportunity, there's also so much fear around it and there's also so much protectionism which might come around it. So it'll be it'll be interesting to see how this molds over over next few years, next few months. But yeah, very you're seeing every kind of behavior happening around this technology. Like and again, very similar to how any new technology is treated, whether it's nuclear, whether it was computers back in our time. So could be multiple things, but yeah, great to see that at least there are there are more progressive mindset today around it versus yeah, versus more delimiting mindset. But yeah, we'll see how this distributes. But yeah, this was these were our key items in the docket for this week. Anything that you think you've come across, anything which is which might be interesting and which might be worth you giving a thought here.

37:35I think I'm So one other recent development was like some of the companies like actually launched their variables like snap launching a like a mixed reality headset, mixed reality glasses, right? And I think we have we have we are also probably not too far from let's say the the open AI consumer product, right? So I think that would again be a little I would say a very important inflection point in terms of again AI search and AI information consumption, right? So AI will become like at least I'm I'm of the opinion that AI in a positive sense will become even more integral part of our day-to-day information diet and information let's say pipeline, right? And especially with you know like from the topic of search, the lens that we have here, I think personalization will become extremely important, right? So, so the only vector that people would have to ensure that they're getting the right kind of discovery is to is to be is to be thoroughly detailed about about their own products, right? So, you know, yeah, should have more case studies,

38:52Should have more reviews, should have more details about the customer, your you know, product use cases, etc., right? So, more and more details about the product, the better it would be to get discovered.

39:06Yeah, and actually that's a great segue. so, you're absolutely right. I think any new revolutionary technology will also enable a new form factor. So, we saw mobile coming in with the internet with apps. We saw again watches coming out. We saw of course, AirPods coming out, things like that. So, it'll always be a new form factor which comes in into picture. Now, with AI being so revolutionary, there are ideas around you. You're absolutely right that Open AI device which is like which is like a phone, but it sees everything around you. Then there are glasses already in the market. Then there was thoughts around a pendant coming in, some company launching a pendant. And then there is also there were again again about AirPods with cameras where they're more specially aware around what's happening and they can give you add that context to you as you are as you are going through the world.

40:01Okay.

40:02So, absolutely right this and because AI inherently as a technology can consume a lot more data and then, you know, give you a very crisp understanding of what's happening around you, I think they'll be the products are right now more focusing on enhancing other sensory perceptive. So, it could be what they hear, what they are seeing, what they are spatially aware of, where they are going. So, I think some of that those domains will get a lot more attention now.

40:31Yeah.

40:32So, it'll be Yeah.

40:33So, just just a last bit on I think I think you you alluded to a very interesting way to look at information and form factors, right? So, if you look at the last 30 30 35 years, right? So, with every new, let's say, device innovation or hardware innovation, right? So, there always has been addition of a new information dimension, right? Or a new information aspect, which was not earlier available, right? For example, from newspapers to websites, right? It was on-demand consumption, right? newspaper is only available, you know, at the time when it is published, and the next next comes the next edition comes out the next day, right? So, anytime availability is the new dimension with the form factor of the websites and the and the internet. With mobile, I think what the new information that came about in the information consumption equation was the location, right? Where you are, right? And then the information around it, right? That would that would change what you consume, how you consume, right? now with AI, I think there are a bunch of dimensions coming together at the same time. So, you know, AI first of all, it knows a lot more about you. Second, it actually can not just it's multimodal, right? It can see, right? so

41:50When I say AI, let's say form factors associated with AI. For example, let's say a glass, right? Or a AI application on your phone, right? It can see, it can it knows you more, it is multimodal, and also it can judge on your behalf, right? So, I think these three dimensions coming together is are going to just fully revolutionize how we consume information and what new applications are getting built.

42:18Absolutely. And then just one more access around this to to keep in mind. I think any new form factor the the new form factor will succeed if they help user like disproportionately more in some new paradigm. So, for example, just building on your example of how mobile enabled location and thereby Uber became a great use case, right? So, it was not just that whatever you were doing on your laptop, you were able to do on mobile. No, mobile enabled absolutely new use cases and you could do them 10x better. So, similarly, I think a success of a form factor, so today if it is glasses, right? There has to be something some use cases which you can do 10x better with a with let's say AI glasses or with let's say something else, an AI pendant. It should not just be an incremental improvement on your process because otherwise it will just be information overload, right? If you are seeing your feed on your glasses, like it doesn't really maybe that's that's my thought that some use cases have to be disproportionately solved better. And then the you then that form factor will be adopted significantly faster and more aggressively.

43:32Totally. Totally. I think if the if AI can see what you see, right? With glasses, right? And it can it can communicate to you in in form of a whisper, right? I think that that angle is going to unlock new use cases.

43:48Yeah. And let us just take that one more step forward. For what use case will it make a lot of sense? Let's say maybe you're going to a conference. Like just a small use case. And then it just tells you, "Hey, exactly that person or XYZ person which is highlighted on my glass is your exact persona. Go speak to them." Again, it's a there it's a blurred line between privacy and and information. But I'm saying exactly that technology, but for what use case? That it makes it 10x better. And that's where the adoption will come from.

44:21Yeah, absolutely. Yeah, there are there could also be many other use cases like, you know, like a like an extra pair of eyes and brain, right? Just imagine if you if someone just had their own twins 24/7 with them, right? So, and that twin is documenting whatever is going in their life. Like what am what I am seeing, what I I may not be actively looking at something, but it is in my camera view, right? And imagine a situation when you are searching for that one thing which you don't know where you do keep, and your your your glasses can remind you that this is on, you know, the right-hand side of your desk, right? Because you may you might have missed giving attention to that piece, but since the camera captured it, it indexed it, and it it knows what it is, and then it helps you all of in in all of that, right? I think it's going to make people more dexterous. It's going to make people more make people more aware you know, about the the ambience or scenario that they are in, whether it is your conference, whether it is you know, if you're traveling to a new city, right? So, those form factors are I would say like just I would say, you know, they're about to erupt with new use cases.

45:35Could be. Absolutely. I think that'll be exciting to see what use cases and how they open up those form factors as platforms. I think those will create a lot of application-level opportunities. 100%. Fantastic. Lalit, thank you so much. Such an amazing conversation. Especially loved your thoughts around open source and how things are changing with with new regulations coming in. Thank you again for an amazing episode today.

46:03Likewise, Ashwani. Always great chatting with you.

46:05Thank you.

46:06Thanks.

46:06All right. Thank you.