Google's squeeze, GPT Live voice and Cloudflare's AI bot rules

0:01Hello everyone. Welcome to our sixth episode of Marketing to Agents podcast. hope you're getting great insights as we are covering latest news. This week again ton of updates from Google, Meta, Anthropic, and of course all the platforms in the space. We heard good news from coming across the board. So, as usual, I have Lalit with me as well. Hey Lalit.

0:28Hey Ashwani. Good to be here.

0:31Great. All right, Lalit. So, we had a ton of news coming out this week. I would like to start off first with news coming out of Google. So, there were couple of different news pieces coming out of Google which were relevant for the AI search and the visibility. So, would love to start off with first AI overviews. There was an implicit understanding that the AI overviews were cutting down on bad traffic. But, I think now there is a research published which is talking about that actually good traffic which was going on websites now is getting reduced because of the AI overview. so, Lalit, would you like to talk a bit more about it? What's happening here? How does that impact publishers specifically? And how does how do companies fight that how fight that battle?

1:25Yep. So, Ashwani, if you remember Google has started to give answers for let's say specific queries. for example, if you ask let's say what is the date of birth of Brad Pitt, right? So, it would it would have directly given you an answer. And this kind of a UX was there for a long time. I think I don't remember the exact date, but I think even before ChatGPT the search engine experience was in a in a way increasingly answer led, right? But, and you know, for that Google has always put up put out this defense that, you know, for such a very for such query a visitor would anyway bounce off because they would right away just look for the answer and then you would have bounced off. And Google is introducing this kind of a user experience only for, let's say, low-quality queries, right? And with in the last few years, like as we've seen, you know, the search behavior is now increasingly AI-driven. Google also has gone a little more aggressive on the Google Google AI overview and the AI mode. And it was still hiding under the same defense that, you know, maybe the it is only it is only basically fending off traffic which anyway would not be spending a lot much, you know, much time on your website. So, there was this independent field

2:42Study that was conducted where, you know, a portion of the traffic was you know, diverted. So, basically they took they took subjects and then like roughly 40% were basically shown AI overviews and and another like 60% were shown not shown AI overviews and they were directly shown the web results web results. And and it was surprising that there's this random distribution of these visitors, there was no significant change in, let's say, bounce rate. There's no significant change in, let's say, time spent on the website. And there's also like an equal ratio of high intent, let's say, lead generation also. So, it was kind of like an eye-opener that the defense right? The the low-quality click defense that Google has al- always put out in defense of the AI overview and introducing more answers in the search engine, which ultimately is reducing clicks on publisher websites and then you know, and for some publishers it is loss of revenue, some for some publishers it is loss of, let's say, pipeline, etc., right? So, this defense is now sort of you know, blown up. Like people this independent study shows that know the quality of traffic which the AI overview is sort of withholding itself with itself is not necessarily low quality.

4:02So that's actually pretty significant from a publisher point of view.

4:06Absolutely and if this is the case then what do you think as a publisher as a brand I should do about it like because Google never asked my permission to crawl my content or show answers from my content. It's not showing that content so now showing that answers in overview. What do I do to be able to to be able to ascertain that yes I am really the publisher of that. Does citation help there? Is Google actually providing that okay yeah my website was the source of this information which Google provided an answer for? At least some level of attribution of the credit or citations?

4:47Yeah right now Google hasn't put out any guidelines on like what what will get you a citation in Google AI overview. But looking at this study I think publishers who are sleeping on the AI engine optimization the AEO right? There are still a lot of publishers who are sleeping on it. Because they still believe that high quality visitors are not really being shown the you know AI answers. So this should be like a wake up call and every publisher should start being becoming very serious about optimizing for AI engines and like specifically AEO and ensure that their content is getting proper citations which would then result into let's say some bit of traffic converting into clicks clicks.

5:37Understood. Understood. So this is Lalit about organic. Now what about brands which are also doing advertising with Google or or like trying to do search advertisement. Now that search advertising is also below the overview section. So, do brands also faces face squeeze from that side where now the clicks would be lower on the advertising because the AI overview itself would give the right response. And then therefore people going lesser lesser journeys getting started.

6:09Yeah. So, so broadly so categories where so search categories basically search categories where someone is explicitly looking for a service provider, right? the the percentage of times you end up seeing a Google AI overview, it's actually much lower. Like if you right now also search for let's say best res- Japanese restaurants in Bangalore, right? It it might it might not show you a Google AI overview. It might actually show you the local listings, right? So, Google is smart that way in in terms of deciphering the intent and then then tailoring the search experience based on the intent. but why recently what Google has been doing which has also been reported by a lot of advertisers on social media that Google has started to sort of replace the advertiser generated description of the ads with AI generated descriptions. And as usual it also contains the disclaimer that it AI can make mistakes. Now this was like a surprise for a lot of advertisers who are advertising on Google and then they when they were just monitoring their own their own ad impressions and they were surprised to see that this AI generated description is now appearing

7:27And the the whole articulation of AI can make mistakes, it it's it's it's it doesn't really sit well with the brand, right? Because they were very editorial and very thoughtful in terms of putting out the description. And to my surprise advertisement were not taken in confidence for this kind of a change. So that is something that you know is not really a good I would say good thing that Google has done.

7:52Plus there is as you said there could be risk of misrepresentation. There could be there could be certain there could be certain promotions about the brand which they wanted to highlight. So all of that gets changed automatically then.

8:04Yeah. Maybe it was a temporary thing and maybe they might roll it back. But there should be an explicit opt-in.

8:10That's true. Yeah. And not just that I think there is another news where Google might change the bidding bidding methodology or budget limits on the campaigns thereby trying to monetize more from advertisements. So that means there would also be while the organic visibility is getting compressed there might be higher cost if you are trying to advertise also. So both these factors might impact the organizers by publishers directly and brands.

8:46Yeah. Absolutely Ashwani. Yeah. So So there are a lot of advertisers who are running let's say ad campaigns on a budget constraint and they typically go for setting up the budget and then letting the letting the AI engine the Google engine optimize for clicks, right? So this is typically called performance max, right? And what is what has been come to our notice is that Google Google Ads team has confirmed that mid-August they're going to go for a price change and it is likely going to be more expensive. Especially for this automatic bidding kind of an ad. So you know if you are a publisher I think Google is sort of sticking up I think you know it's kind of squeezing you from both ends. So on one side it is making it kind of like a black box in terms of discovery on organic side. And on the paid side also, it is sort of playing games where ad copy is getting changed without notice, without opt-in, and then the placement of ads are also sort of getting some sort of a you know, there's not a guaranteed display as always. And also this price rise.

9:56Interesting. A lot of so, businesses heavily reliant on organic, either organic on Google or paid traffic from Google, I think they need to start rethinking aggressively about how to get featured in those AI overviews or trying to optimize their presence on such search on such AI searches because I think if that's right now that's the most dominant positioning in any search-based experience today. So, yeah, that's a I think a good learning on the brands and the publisher side that focusing on that. moving forward with that, now that we're discussing Google, Chat GPT, there was a study done around Chat GPT where something came out which was that each of the different models showcases different brands for the same prompt. So, for example, if I'm searching for okay, give me a list of companies in this space, and if I'm using different modes of Chat GPT, it leads to different answers, thereby different brands are getting cited, different referrals are getting highlighted. So, very interesting to to hear that there are that even the AI visibility now has to be

11:14Planned for whether you want your answer to you want your brand to be shown on instant answer or in a thinking mode answer. You know, so prioritizing kind of content and reasoning.

11:27Yeah. So, basically, so what was what was observed that between the ChatGPT instant mode and the ChatGPT thinking mode only 25% roughly 25% of citations had an overlap. Like 25% of websites were cited in both modes. And there was a strong inclination in the thinking mode to go for more authoritative websites like more academic resources or like traditional sources with more domain authority. So, and in the instant mode and more leaning towards social media and Reddit. So, very interesting. So, as a publisher, I think it's it becomes equally like, you know, both modes become equally important. So, if you're looking at your visibility metric, you need to look at visibility not just across models, not just across engines, but also across modes. So, it becomes like a multi-dimensional grid now.

12:24Oh, yeah. And also it will then depends on the web the service of the product that you're offering as well. Because let's say if someone is in a thinking mode, they could be This could be more of a research topic and they might there be less propensity to go and click or, you know, go and engage with a service. Whereas instant mode, like you you were giving the earlier example where you're looking for something very very specific. There is a higher They could be a higher intent for you to then go and click a service or use a service, for example. So, from a brand, you also have to position yourself in understanding which mode is more relevant for your business. So, if you are a thought leader or you're trying to be a media side or you're a thought leader publisher. So, then probably you want to be sure that you're getting listed in thinking. Whereas if you are a specific very niche service that you're offering, then I think, yeah.

13:18Yeah. And another very recent like the the other interesting update about ChatGPT is that they recently launched the GPT live voice mode. Right? So, they're calling it as GPT live and it is their first full duplex voice model voice model, right? And that voice model can basically respond to you while you are speaking. It can also interrupt you. It's full duplex as I said. And it can also delegate thinking to another model for for let's say conversation points where thinking has to be done or let's say something has to be fetched. And that experience Ashwani is fantastic, right? And I have seen social media like you know opinions about this is so good that people are now spending lot more time on the GPT live experience. Where they are So some users actually on social media they they actually confirmed that they are keeping the chat GPT live always on while they're working and as a colleague they would just you know call out chat and then ask a question and then they would they would get an answer from chat GPT. So I think that is also very significant development where the shift in consumer behavior like if especially a lot of B2C searches, right?

14:37If the voice experience is is is at such a great user experience level, right? Then I think the voice search becomes all the more important and in voice search as we know I think citations are not really that important mentions become lot more important, right? So in my own recent like just in the afternoon I was basically engaging with chat GPT live in a voice conversations about let's say which agent framework is more important for let's say building a team of agents, right? So it automatically suggested LangChain, right? LangChain and LangGraph, right? And then it also listed a few others, but then it it prioritized based upon my memory that it has that you know for your project, I think this might be a better choice because they have a lot more case studies from enterprises. So, the other things that were mentioned, obviously, now I I have more recollection of those. It had mentioned Mastra and a few others. But, if you just look at just in case if this becomes a significant user user behavior, I think mentions will become all the more important. What do you think?

15:53This actually Yeah, is a very very interesting paradigm. So, voice as a mode, even even when ChatGPT launched in 2022, the voice mode became very popular very quickly. Just because it gives you a feeling that you are, you know, conversing with someone real, and then you can just bounce off thoughts. You don't have to When you're typing, you're already you have to be more structured. Usually, people are more structured because you have to think through something, and then you have to type it. Speaking is a lot more natural by default, right? So, I think if this mode, and given given OpenAI, now I don't know what's going to happen to that, but given what OpenAI was building with with Jony Ive and the and the hardware that they were building, which could be your companion and see the world around you, and tie that now with the voice experience. Now, this could be like a like a humanoid who's always with you, sees the world around you, and is available to converse all the time with you. So, then it becomes something really good. Now, I don't know if just with what what we've heard in last couple of days with you know, Apple suing OpenAI and all of those, so so we don't know where that would end. But, yeah, I think this is a great paradigm.

17:09And like again, I said that, right now the frontier for engaging with LLMs is mostly chat, and of course texting. This because opens that new frontier and then there is another I think there are a couple of startups I was observing in in the valley which are now bringing all of that engagement on your on your lenses. So on your glasses. So So yeah, I think this could be a great step to increase mass more mass and options where people could just talk to it and then be a companion all the time.

17:41Yeah. I think so just from my own personal experience of using ChatGPT over voice for over last 6 months. I think what the difference with GPT live is that it actually feels like you're talking to a human and the tonality and the the little nuances that it they have addressed it sort of feels like that it it's a friendly companion. And the earlier voice conversation mode was more of a turn-based you know, experience where it would be it would it would you just have to wait for your turn and then it would be a little more frictionful as compared to what we have today with GPT live. And with that reduced friction I think the usage will be you know, like we we will see a we will see a lot of usage shifting to voice. Earlier it was probably maybe like a voice query when when you were probably driving you would just simply double click and then you know, but now I think I've started to keep ChatGPT live as a always-on companion on my desk, right? And I like I I find myself you know, talking to ChatGPT a lot more since the release of GPT live.

18:54But let's still just just double clicking on that. So but the kind of queries or the kind of conversations those would be would be still extremely short and not very deep, right? Because for something which is very deep or research-based or where you have to get a service or where you have to get a specific you want to get something done, you would still go on the desktop version of it, right? Because voice would still be just for quick queries or quick questions that you have as you're going through things. Right? So, I I would think that the quality of engagement would still be more casual on voice versus lot more research or work driven or like getting something done driven when you are on your system. Right?

19:40Yeah. So, no doubt about it. I think that the intensity of work will be obviously lot more on desktop. But, if you look at what what we are seeing as a pattern is like even coding agents like Codex and Cloud Code, they all have shipped remote access or remote control kind of a mode where you can take your work with you. You just have to keep your laptop always on and you can manage your Codex sessions or Cloud sessions from your phone, right? So, what we are then moving towards is work which is always on, right? So, there might be an intensive version of that work which continues to happen on desktop. But, you can continue with the same conversation. you might maybe maybe start a conversation maybe when you are let's say morning doing going for a walk in the morning, you can talk to chat with GPT live. And then, just say that you know, create a create a schedule task for that. Or you know, just add a recurring task for this call. What what I do initiate a deep research and email me something. Right? So, I think that's a gateway. Voice GPT live is going to be a gateway for getting the work initiated. a lot of this thought thought level work and deep thinking that people are into, that happens in brainstorming sessions. That happens when people are

20:57Let's say in in contemplating mode maybe over a walk or while doing let's say you know, taking down notes on a note notebook. Having this always on companion would just be a great way to funnel out a lot of let's say thoughts for later intensive work.

21:14Interesting. So, then that then you know, tying it back to our AI search and visibility, that means mentions is way will become way way more important than just the citations because as you are as this new form factor opens up, you're not really going to be following those links around, but you'll of course hear those brand names or get those brand names mentioned, which will then create that brand recall in your mind. So, I think then

21:43Yeah, citations is independently it's important, but then the mentions are way more important then.

21:49Absolutely. Yep. So, in my like conversation in the afternoon, I there was a mention of this framework called Temporal alongside launch LangChain. And I had never heard of Temporal as a as a framework. Then in my next que- question to GPT Live, I asked more about Temporal, right? So, and then it could it could fetch information and now I'm a little more aware about what Temporal is. So, a click in the voice mode is basically asking more questions about it, right? So, dive deep can still happen in that mode. But absolutely, I think mentions are going to be a lot more important.

22:25Perfect. Again again, amazing insight. I'll also give it a shot and then yeah, happy to happy to discuss more about it next time. All right, let's move on to our next big news coming out this week.

22:39Cloudflare.

22:39This one is out there. Yeah. So, Cloudflare came up with a couple of interesting updates. first of all, I think we learned that Cloudflare can actually mistakenly or accidentally block AI bots. So, Google bot was blocked and and it could hurt your organic visibility. And then they also there were also news about Bingbot getting blocked as well. So, this comes at the heels where AI search is becoming more and more important. So, what do you think like that what should be the brand paradigm? Then of course we can talk also about how they are now trying to provide more tools or more control over what gets crawled on particular publishers website.

23:28Yeah. So, so Cloudflare is basically the in some sense the the gateway to roughly 20% of the internet. It's a lot of these publisher websites. Even us at Airmeet we use Cloudflare. It helps us against DDoS attack. and essentially it has become like a firewall for us, right? And a lot of content delivery also happens through it. So, it has access to each and every request that is coming to our to our to our website. And it it started to classify the traffic that is coming. Especially the AI traffic which is, you know, where the user agent is a bot, right? It it had started to classify that AI traffic or bot traffic into three buckets which is search basically search which is basically trying to index the your website into search index. Then training where let's say AI lab is coming for specifically to your website for, you know, token collecting tokens for training. And then third is AI agent where basically the AI agent is coming on behalf of a user, right? Now, what they are doing is they are basically they have give out they have

24:45They have given out a date which is September 15th to all the publishers who are using Cloudflare that they can come and explicit you know you know explicitly set the policy access policy on their website. So, by default they would disallow the training traffic and the agent traffic, right? The search indexing traffic is allowed, but by default they will, you know, disallow the training and the the AI agent traffic. Which which as publishers we can go and like depending upon our let's say our unique situation, we can allow it. So, it's not very clear whether the Googlebot. Now, for us for any publisher to ensure that they are in a Google or a Bing search index, is it precisely search or is it like the training traffic also? So, it's not very clear, right? So, the advisory out there is that people should explicitly mention Googlebot and Bingbot in the allow list. There's also an allow list which overrides your explicit setting on these three buckets, right? So, you might still block general training traffic, but allow some you know, some training traffic, right? From from some from let's say some providers. So,

26:03As I said, the advisory is to basically make sure before 15 September 15th, you must not be blocking, you know, at least the Google and the Bingbot because they are in a way the foundation for a lot of AI search that might that might be using their index. So, that is one, and the other related cloud cloud player update is that it is also ushering in this era of pay per access for the web publishers. Like so, today like you know, up till now, there was there basically essentially two business models on the internet for people who are publishing content. First was revenue-based monetization is like ad-based monetization where you have ads on your website, and then the second was subscription where you basically subscribe users can subscribe for a monthly, you know, for a month and then access your content. Now, Cloudflare is basically building the infrastructure for publishers to also monetize via paper access where after you have set your allowed disallowed for these three buckets, for all the bots which you have allowed, you can then set whether certain pages are

27:22Only available for, you know, paid access where the AI agent which is accessing, let's say with a premium set of content, premium set of information beyond those, you know, on those links, you could use a protocol 402 protocol is there where you can basically the web server can point them to a payment infra where they could explicitly make a payment and then they would be allowed after a successful payment, they would be allowed to access that page. So, this would this would be a great, let's say, I think it's a good news for the entire ecosystem and democratizes you know, the in some sense, Cloudflare democratize, you know, the the the unique content creation that continues to happen on internet in the long tail, right? you know, it's very difficult for the AI labs to basically reach out all these long tail content creators, right? So, this is a great way for Cloudflare to become like a middle exchange where the discovery can happen through search index and then, you know, through paper use, their content could also be part of the training and publishers also make some money.

28:36To me, this is a very, a monumental development. I think in the last few episodes we've talked about how with the whole LLMs coming in and then we heard the news from UK where they are like publishers can opt out of it. Like there is legislature being built around this so that to give more rights and more control to publishers. I think what Cloudflare did actually has a real-world implication and gives more control to publishers immediately. So, they can monetize also and they can of course disallow as well. with the content that they don't want to you know surface on AI searches or AI overviews and let the users or it could be you know it could be their gated content, it could be ungated content, but whatever at least it gives control back to publishers. Again, how much does that monetization you know cover for the loss of organic traffic that they are getting or loss of their you know ad traffic, but still at least there is some path to a better more controlled version and more publishers getting more power as well. But then my question is to the other side. Now, for the LLMs, now one part is training. So, let's talk about a publisher. Let's talk about a media

29:53Site which is industry media site. So, let's say an industry media which is into procurement. Let's say they publish content every day, they publish videos, they publish content every day. And then today the Cloudflare this control goes live where now LLMs can no longer train on their data. But up till this point, all the training all the training data already has their all the old content. Now, how does it matter if they even stop a training bot or training crawler now? now of course they will still allow for a search crawler or search bot which can still show their content in AI overview I would assume. But it would not train. So, how do how does that still solve a a a publisher who's already whose content is already gone into the training models already? Like how how does that change things for them?

30:51So, the search the the classification is basically three buckets, search, train, and agent, right? So, search is when the index indexing is happening. And in that mode, I think we assume that, you know, the AI labs will be ethical. And obviously there's there's going to be confusion between, you know, the crawling that is happening through search and that same could be used for training. And that is the confusion which I was highlighting earlier and that's where, you know, accidentally Google bot might also get blocked. So, you have to go and explicitly allow it. but your question, I think the See, for domains where new data keeps on getting generated, right? this could be let's say a lot of pharmaceutical publications or it could be, let's say, you know, for your example of industry industry trade news, right? So, a lot of regulations keep changing, new information keeps getting published. So, the publisher can control the indexing piece by just giving let's say just enough information, not the whole information, but just enough information which is part of the page description, which is part of, let's say, the you know, the the typical SEO, you know, tags, etc. But the bulk of the the

32:08Content could still be withheld with let's say a paid subscription, right? So, as a user, either you pay or let's say when the agent is coming in the training mode or the agent mode, right? So, they can also play pay to access that content. So, as a publisher, you you can do that kind of a like, you know, that mechanism is open for you.

32:27I understood. Understood. Yeah, that could be something which yeah, which can give again, something which gives control back to publishers. But yeah, this as this space evolves and I and looks like with Cloudflare starting on this path, there would be other other services also which would probably try this path out which will give control back to publishers. but, yeah, I think this is a great step in the right direction. It needs to help publishers control what gets indexed and how their content is surfaced on the LLM search.

33:02Yeah.

33:04Yep. but, coming back to one of our last items or from news from this week is around now as we've been talking about EEO, GEO getting more and more attention and businesses being more as they're seeing drop in organic traffic, businesses are getting more and more eager to jump onto the bandwagon. And this has led to some black black hat techniques as well. And we are seeing fake agencies reselling wrong links. So, like what has what is the extent of this problem or or do you see this getting worse? What's happening right now in the desperation to get featured on those AI overviews or get featured in LLM res- mentions in LLM responses, what sort of black hat is now becoming popular which might eventually be very detrimental to a brand?

33:58Yeah. So, I think So, a lot of businesses are getting sensitized towards AEO and then they now have dedicated budget for that. And whenever there's a budget, I think you know, the there are a lot of takers for it. So, you rightly said there are a lot of these black hat agencies which are popping up. And essentially, they're selling bundles, backlink bundles, or let's say boost bundles where they're just selling, let's say, number. The the the sheer quantity of backlinks. And not necessarily the relevance of the backlink. And what this is resulting into is entity sort of dilution entity dilution or entity confusion is what you know the industry term is where let's say a certain company for example Airmeet let's take Airmeet right if Airmeet gets backlink backlinks from let's say B2B marketing web blogs association blogs right and let's say VC and tech tech news all of it is relevant for us right? But if the bundle that we were sold to sold on and that bundle has backlinks from let's say recipe websites or bundles from let's say you know teen teenage fashion focused website all of it

35:15Will just confuse the the AI which is now very smart to quickly understand that this relation is sort of not organic this is something that you're paid for. And it might just dilute the the let's say the authority or the respect that that the AI understood. So AI engines also assign some importance to every domain for the subject that they claim that they are an expert in right? So I think this entity confusion and dilution is is something that is hurting a lot of vendors sorry a lot of businesses who are getting victims of becoming victims of these black hat techniques.

35:57Absolutely so this is one example of how what kind of black hat techniques we are seeing in the market. So any non-relevant directory links pointing to your site or your site pointing to them could definitely hurt your brand because the AI does not perceive your brand to be relevant for the links that you're getting traffic from or you're getting a backlink from. So just for brands publishers just be mindful of you know not falling for such techniques. So, this is just one of the examples which is around backlink building which has been known to be very effective in SEO world and now in AEO world as well. But, ensure that you're staying relevant here. Ensure that you are still using you are still maintaining the context. So, if you're you're posting links or getting traffic, ensure that it's relevant for your users and how your users are coming onto your place. So, yeah, what we've heard is that this these will grow like again, just going back 15 years, SEO when it was getting more popular, there were people were trying to, you know, cross that line and trying to see what gets penalized, what doesn't get penalized. But, since for agency doing this, it doesn't harm them because it's they are they don't

37:15Have huge skin in the game. It but it's your brand which gets impacted. So, just be mindful, build for the long term, build in the right way, and ensure that you're not falling for such low-hanging and, of course, most of the time, you know, across the line sort of traps.

37:33Yep. And the other thing that the, you know, this this whole thing could be could be used in a negative way to sort of, you know, launch a sabotage attack on a certain brand as well, right? Where the brand itself is not doing anything, but the competition has hired a bunch of these black hat agencies to sort of get, you know, dilute Basically, it's a dilution attack, right? Now, that can happen and I think Google has a way in Google Search Console, you can definitely, you know, deprioritize or remove some backlink from the from the Google index for that are that you do not deem which are relevant. I'm not sure what is the overall let's say the the AI ecosystem's equivalent of that. But, the only right action that the brands can do is to continue to have more and more relevant backlinks. Right? So, if you if you keep improving or keep adding to the positive set of backlinks, the dilution can obviously be, you know, reduced as well as the density could be improved.

38:35Perfect. Sounds good, love. Yeah. I think this was quite an insightful episode learning a lot about some of the new things which are coming out and then of course always the space is evolving super fast. So, good to keep a tab on all the changes which are coming. But thank you, love, for Thanks for spending that time. And listeners, we are looking forward to hearing some feedback questions from your side. Let us know if there is something specific that you would like us to cover in our next episode. Thank you.

39:07Thanks.

39:08Thanks, everyone.