Latest / The Interline Podcast / Live Shopping, MCP Intelligence, AI Sustainability And Lending Risks, And Mum Fragrances In Fortnite
Transcript
- speaker-0: So welcome to the edit from The Interline, the show where we run a quick fire analysis on our pick of the most important fashion and beauty technology stories from the last seven days. I'm Grace, the social editor, and I'm joined by Ben, the editor-in-chief, and together we have less than twenty-five minutes to give you our analysis on the stories that we think really matter. speaker-1: Hey Grace, good to see again. It's really hot in the UK and a whole bunch of other places again at the ⁓ at the moment. So if you're watching this on video and we collectively look a little bit frazzled, that's probably the reason. ⁓ I'm rocking my linen shirt for anybody who's listening to the audio version of this, because it was the the best combination of smart and relatively airy I could put together. It's about thirty one degrees in this room, I think. speaker-0: So a big headline this week was that the live shopping platform Whatnot has raised an additional five hundred and forty-five million dollars in Series G funding, and this is bringing their valuation to twenty billion dollars. This is apparently the biggest raise in live commerce. So I'm interested to get your take on this news generally, and also if you think Whatnot is in its own lane in terms of live shopping. Since its mission is really to help small businesses grow, or do you think it is competing with the likes of TikTok Shop for instance? speaker-1: So we've covered live shopping before in I think it was StockX's expansion into it. ⁓ and I think we've talked about the sheer scale of the live streaming, live commerce sort of market ⁓ in the past. Now, this is an interesting one to me because it's clearly a very big race and it's clearly a company w that has a lot of ⁓ support and belief behind it. They claim that the live shopping market In I think that they define it as the US, the UK, and Europe as worth an estimated 20 billion. ⁓ so roughly in line with what their current valuation is. And they they currently have like 60% market share of that. Now, it's difficult for me to pass this. So I've read through ⁓ Whatnot's own sort of state of ⁓ life commerce, I think it was called, report for 2026. I've read through a bunch of the analysis of this. ⁓ and it's tricky to understand. what the definition of live shopping is here. ⁓ like it your your question's pretty apt. Is it its own vertical in the sense that we define it as dedicated apps that are exist solely for the purpose of people selling through live video? Do we define it as encompassing the entire social media landscape with all of the platforms that have added Transactional or referral ⁓ functionality. So TikTok shop is the most prominent one, but there's a fair amount of commerce that's integrated into Instagram and everything as well. You and I talked, I think it was last week or the week before, about how every social network is a selling and shopping destination at this point. ⁓ now I I wouldn't pretend to know how that's gonna play out. ⁓ unless I'm missing something. There's no way that Whatnot owns 60% of the overall potential market globally for people buying things through social commerce. For one thing, this excludes China, which is a a a major kind of live shopping destination. It Has to definitionally, I think, exclude TikTok ⁓ and ⁓ exclude Instagram. And if you stack this alongside some of the ⁓ kind of record-holding live shopping events, so the cosmetics brand P. Louise here in the in the UK, they've set and then beaten their own kind of live shopping record a couple of times with selling something like 2.7 million. G VP worth of merchandise in a twelve to fourteen hour window on TikTok shop. There's a huge amount of scale here. And I think the question for me is if there if it's dedicated applications, is that where individual sellers and micro brands and ⁓ kind of smaller brands go? That seems to be the case between what's ⁓ what not's communications or Does it actually start to then pull some of the larger brands who are setting records and who are shifting huge volumes of merchandise through the big established social channels? So for me, I think the thing to watch here is not necessarily is live shopping big? It's very clearly big. It's who owns it? Is it likely to be owned by dedicated companies who ⁓ who literally just do shopping? with a community aspect, or is it likely to end up being owned by social channels that are increasingly making shopping a big fixture of it? And I guess the final consideration for me, as people who listen to this regularly will know, like a non-social media user in ⁓ in my personal life, seems to me like if you if people value selling with a community element to it, which I think is what Motnut's kind of edge is. The cynic in me says, Eventually you bolt more and more community onto the selling, and the selling app becomes a social network in its own right. And ⁓ this converges with TikTok and converges with Instagram. But unambiguously, like week over week, we continue to see a huge amount of money, a huge amount of interest, and a huge amount of brand investment flowing into live and social commerce. I have an interview on the Thursday interview show coming up sometime in the next couple of weeks. Which is about exactly this, the creator economy, ⁓ and the size and the scope and how to measure the value of ⁓ investment in social commerce. speaker-0: The next story is about a shift that could be happening with intelligence providers. So we've seen this week that Edited has released Edited MCP, so a server that users can connect Clawed Desktop or their own agent to, for instance. And this basically pulls from edited resources and data, but the user doesn't actually have to be on the platform. We've also seen something similar to this with the Adobe plugin for ChatGPT that effectively brings Adobe tools to ChatGPT. LGBT work and also codecs. So my question with this to you is do you think we're gonna see more intelligence providers and technology companies do this? It seems like there's a choice to be made between people selling licenses of your own software which contains your intelligence, or selling that intelligence wherever people want it. Should those be different business models or the same thing, do you think? speaker-1: This is a really interesting one. ⁓ so MCP just just does a bit of groundly model context protocol for anybody who's not deep in the ⁓ AI weeds. I think that that's common parlance enough now that everybody understands it. ⁓ it just think of it functionally as a slightly enhanced version of APIs and ⁓ tool calls for ⁓ for databases for AI agents. Now, edited is ⁓ a retail intelligence platform, I think is the way that they would put it. ⁓ I used to know ⁓ the One or two of the founders there, they've they've since moved on. So it's not functionally equivalent to like a forecasting service and platform like WGSN. It's something more tightly aligned to brands wanting to analyze the competitive landscape ⁓ and to make fairly quick decisions based on a huge amount of different variables. So the intelligence that's packaged into platforms like edited is the value, right? ⁓ if you if you are a subscriber user of edited, you are there because you want data to back up the choices that you are making about the products you make, the assortments you build, ⁓ the channels you start them through, the prices, the markdown strategies, and so on. Now that doesn't mean that the edited interface doesn't matter. 100%. If you're an edited user, I'm sure there are people out there who like sitting down and interacting with it. What it does mean is that edited and a bunch of other companies, which I'll get to in a sec, have kind of started to reckon with the idea that the interface is much less important than what it gives people access to and what it contains. And right now, as I if I understand it correctly, unless I'm missing something, ⁓ edited MCP allows people to ⁓ pull all of that data into, as you said, AI agent harness of choice, provided they are existing edited users or customers. ⁓ there is no kind of secondary tier between full platform users and MCP users. The same is True of the Adobe plugin for ChatGPT, but with a slight twist. So I think you can do some of the basic functionality through the natural language interface in ChatGPT without an Adobe account. But if you want access to any of the generative tools to Firefly and things like that, then you need to have a pre-existing Adobe subscription. Now, you know better than anybody that the Adobe subscription model is very confusing and the packages are very confusing. But what I very strongly believe that we're going see is a new class of licenses for yes, intelligence platforms like Edited, but also creative tools like Adobe, ⁓ supply chain management. You name it, if there is value in the data that it contains and the questions that it can answer that exists independently of the interface, we're going to see a new class of licenses which are just This kind of MCP access. I think that's pretty inevitable. Because think about it from Adobe's point of view. Would you rather sell a hundred for argument's sake, $33 a month ⁓ licenses to Creative Cloud? Or would you rather sell like the full suite so you can go download the desktop versions of Photoshop, ⁓ Illustrator, Lightroom, and so on? Or would you rather sell 10,000 five dollar licenses that just give people access to the core features through ChatGPT or through their agent of choice. You you would rather have the latter, I think, just from pure monetary terms. The thing is you don't have to choose between the two. ⁓ I think we're gonna see much more stratification. You're gonna see people who are desktop users, and then you're gonna see people who are MCP headless ⁓ chat interface users of these kinds of products. Intelligence and creative tools are a good First step, but you can stretch the definition through to transparency platforms and traceability and life cycle impact assessment. You can trace it down to I don't know, dis the line planning, ⁓ forecasting. There's there's just so much that you can put into a MCP that doesn't require people to have a full license. And it significantly expands the audience for your platform and your tool in the process. So I think we're going to see a ton more of this. And you're going to see some kind of cheap and easy attempts where people just go, hey, we've just made a few features available through ChatGPT. And then you're going to see some more serious attempts, which I think Adobe's is a good example of, where people are saying, we have so far only been able to sell software to people who are interested in sitting down and learning it and interfacing with it in a desktop app or in a web application. Now all of a sudden we have a new class of users in a potentially much, much bigger addressable market than we thought we had. speaker-0: Next, I want to talk about an interesting piece we saw in the business of fashion this week. So the piece reveals that AI chatbots. Could be making inaccurate, sustainable fashion recommendations. So, as you know, LLMs are pulling from simply the words most frequently used to describe fashion brands, rather than doing independent investigation into the truth of those words. So, Business of Fashion partnered with Quilt AI to analyze responses from ChatGPT, Claude, and Gemini, and when prompted to suggest. Clean, plastic free, and sustainable fashion brands. They found that though AI was suggesting some well known sustainable brands like Patagonia, for instance, it was also clear that the results were simply based on language, therefore not totally accurate. For instance, Allbirds was among the results that we know isn't true. So this story was interesting enough in itself, but it also came across our desk at the same time that we saw an announcement from Anthropic. And the announcement is that all text generated with Claude is going to carry this invisible watermark in the future. So this is to comply with the EU AI Act, but a lot of people have some very strong opinions about this. So what do you think about this story? Do you think there's a story about what AI looks like in terms of written content? And how do you think companies should be thinking about how they create that content in the first place? speaker-1: Hello. Yeah. So I'm I w I would always encourage people to go read our AI report for a whole bunch of reasons. ⁓ for for this one, I'm gonna pull from a piece I wrote about I think I titled it like turning down the temperature and the AI discussion. And in there I I made two recommendations for people. One is everybody should at some point in their life sit down and interact with a like a proper frontier model. That is fully connected to all of their enterprise technology estates. So a clawed Opus V or what have you at the time of recording with connectors to everything that you do. Because you will quickly find that if it's if it if LLMs are given the right context and the right information and the right grounding, they can give you accurate and trustworthy answers. So the assumption there is that your own internal business data is up to date and is accurate. And if you connect it to an LLM, you will get up to date and you will get accurate answers. The other thing that people should do is they should run like a really small local model with no access to the internet. And they should ask that some questions about something that they're interested in or something that they care about. Because that then gives you the very unique experience of being lied to with like supreme confidence of of receiving back a reply that is it sounds believable, but is just based on ⁓ on on inference. It's not based on any kind of source of truth. The reason that I suggest people do those two things is that LLMs don't know anything. They d they don't go out and in ⁓ there's n there's no investigation, there's no verification, there's no like in order to determine whether Patagonia, for instance, is sustainable. You or I would do a reasonable audit of ⁓ reports and everything else. We would look at word of mouth. We there's there's a whole bunch of variables that we would take account to decide whether or not Patagonia was a sustainable brand. An LLM is doing it based on ⁓ the references to Patagonia that exist within its training data and whatever information it can gather at the point of inference. So at the time that somebody makes a query. ⁓ and at the time that it then does its own ⁓ web search and fan out search and whatever else it does to return information. Now, most LLM queries default to fairly quick web search. They default to ⁓ to fairly fast returns because ChatGPT, Claude, and so on are optimized for giving people quick answers. If I was to say, recommend me a sustainable menswear brand, and then the the answer takes fifteen minutes because it involves a big set of research and fan out. I'm just gonna move away and do something else. That use case isn't there. ⁓ if we're doing the kinds of research that you and I do, then great. Then you can go away, you can do something. now what this means is they pull from the most readily available and the most heavily kind of search-ranked content. ⁓ and a lot of that is based bas it it's a combination of Reviews, recommendations, brand language, backlinks, like the the usual type of stuff that makes up the content of the internet. None of that is independently verified or audited. And what this allows companies to effectively do is to make claims in writing, because the the written the written word is the most important part of this here, either explicitly or accidentally through ⁓ kind of AI generated copy or so on, that is not backed up. It doesn't have to be backed up in order for LLMs to repeat it as authoritative. I saw a really interesting experiment that somebody had done on ⁓ Substack where they had made a fake deodrant brand. They've put together a bunch of basically meaningless when you combine them ⁓ sustainability claims like magnesium-free. I I f I forget the specifics of it. ⁓ and they'd written a Substack article about it. They put a website up. ⁓ and it took not a terrible length of time for ChatGPT to start recommending that product as ⁓ as a sustainable deodorant. Now that's the prime example of the fact that there is no independent verification for this. Now to link together this story with our previous one If you were to pair ChatGPT or Claude with some kind of MCP tool call to an index of this stuff, where people have actually done ⁓ inspections and audits and certifications and all of that was rolled up into an intelligence source, then great, you're getting actually more authoritative answers. But we we live firmly in the era of you can just say things. You can just claim to be something that you are not, and LLMs will pick up on that and they will repeat it as ⁓ as gospel. Now, the the way that ties into the Claude watermarking is I know for a fact a lot of companies will do product detail page copy. They'll do ⁓ a lot of consumer-facing ⁓ written content through AI. It will be based on product attributes and technical specifications and design intent and things like that. But it doesn't take a lot for an accidental certification or an accidental wording or something that would class as greenwashing or something to ⁓ to start to slip through. And then if LLMs pick up on it, before you know it, you can end up in a situation where you have a product in the market that doesn't meet the claims that you are making accidentally. And then those claims are getting repeated verbatim ad infinitum by LLMs ⁓ thereafter. So it's a very, ⁓ very complicated space. But the the upshot of it is don't tr don't trust ⁓ LLMs to do this kind of research on your behalf, whether you're an investor or a consumer. speaker-0: Keep my question brief on this one because I know this is an area that you've been watching, and I also know that you've covered it in one of your recent long form podcast interviews. So I saw some news this week from the US government that they've relaxed the financial rules that have helped de-risk investments since the 2008 financial crisis, specifically for data centers. So I wanted to know what's happening here, and yeah, can you explain what's going on? speaker-1: Yeah, I'll also keep the answer brief because this is a prime example of my kind of ⁓ cynical attitude manifesting itself in ⁓ in the real world. So the specifics of this story are that data center ⁓ owners and builders can now issue what are called in the US by the Securities and Exchange Commission, ⁓ like asset-backed securities. So derivative products ⁓ that people can ⁓ invest in and purchase. That have underlying pegs to ⁓ an existing asset class. The reason that this is ⁓ tied to the kind of 2007-2008 subprime mortgage ⁓ crisis and so on is that that was precisely the same thing. So ⁓ mortgage-backed securities were, you know, everybody just go watch the big short or read ⁓ read read the book ⁓ tied to that. But the essential idea is you can you can do a lot of speculation built on top of an underlying ⁓ asset class if you believe that asset class has a future. And the the two thousand eight financial crisis was I think took a lot of people in the financial world somewhat by surprise, just based on the fact that mortgages have always been considered like a very stable asset class. There was a lot going on under there about kind of rating different ⁓ tranches of ⁓ of investment classes higher than they should be, which a whole lot of corruption that that was behind it. The the underlying assumption here though is that data centers represent that stable. that kind of stable asset class. That's that's what the the SEC is is claiming here. And therefore they should be able to offer these derivative ⁓ products on that basis. Now, I'm not sure any of that is true. ⁓ we time will tell at the moment. Like data center usage has been fairly flat. Data center usage is definitely on the up. ⁓ there's a lot of rebellion against the build-out of ⁓ of of data centers, particularly in the US, which is where I think this this friction comes from, but also here in the UK. This is an example of what people talk about when they talk about an AI bubble, which is this idea that it's potentially a house of cards built on the assumption of future use and the build out of capacity to meet future use that is assumed and forecasted rather than guaranteed. If you then stack derivative financial products on top of that, you're making a whole layer, a whole series of layers of assumptions that it doesn't take a great deal for that house of cards to collapse. ⁓ the implications that would have for AI companies and for the the general economy are pretty profound. ⁓ so I'm I'm sure it's something that everybody, CFOs and so on in fash in fashion, are watching. But even if you dislike AI, it's something to continue to be aware of. speaker-0: So the final story is the news that Joe Malone has launched a fragrance with an in-game experience in Fortnite. So the activation is to promote their new Sea Salt Bergamot Cologne, and it's a two-week in-game experience that apparently will give consumers a more personal and immersive way to discover the fragrance through exploration and in-game storytelling. Now, we've talked about gaming being the best way to reach younger generations. In the world of social media bands and all these things. But in all honesty, it does feel like Joe Malone, to me at least, is more of like a mum brand and not something that belongs in a game like Fortnite. So I wanted to get your take on this. Do you think this kind of play to target younger consumers through gaming is viable for brands that already have a pretty clear defined audience? And I also know that there's been wider news from Fortnite's parent company, Epic. So why don't you talk about that as well and yeah, give me your thoughts. speaker-1: Yeah, so your read on Joe Malone and my read on Joe Malone are basically the same. ⁓ it's not a young person's brand, hasn't been a young person's brand for for a while. At least purposefully cu cultivated that that audience. very clearly this is the result of some kind of market investigation and reporting that says there is a target demo we can go after here. ⁓ and the best way to reach it is to work with a ⁓ an an agency. So this is not an official epic collaboration ⁓ of of the kind this is this is part of the Fortnite creator platform side of things. So they've they've partnered with somebody to build out this ⁓ experience with the goal of reaching the kind of audience that you and I talked about before that is not going to be on social media for too much longer in in the UK and elsewhere when social media bands come into force for ⁓ for under sixteens. Now I I don't see any reason that this play couldn't work. There's the the the the whole of f fashion, fragrance and beauty is riddled with stories of brands that became a little bit fusty and did a successful job of appealing to younger demographics over time. It's also filled with examples of where that didn't work. ⁓ so I can't I can't make any kind of cultural assessment on whether this is something that will play out. I think it's interesting that Fortnite is what they turn to because ⁓ Fortnite has had a drop-off in engagement ⁓ and ⁓ and in players. and Epic have also had a bit of a time of it recently. ⁓ they've they've conducted a bunch of layoffs. They acquired ⁓ the platforms ArtStation and Sketchfab about five years ⁓ five years ago, which is the other news that you mentioned. And they've now offloaded those, ⁓ they've now sold those as of this week to ⁓ to someone else. Now those were purchased, I think, on the assumption that Real-time 3D was the future of the web. Right. They were they were they were built out on that assumption. So the the original vision for something like Germalone would have been come to our own website or across other channels and interact with real-time 3D in some way there. ⁓ now the bet is very clearly from Epic's side and from brand side, let's focus back in on the core experience and the core engine and the core IP where people actually are. I think we'll just have to see if Fortnite engagement and numbers do go back up in the way that Epic are predicting. ⁓ I think you'll see more companies taking this like agency build route rather than doing direct partnerships with ⁓ with Epic and similar platform holders in the in the near term. The last thing I would encourage people to do for this one is ⁓ ages ago, I think towards the tail end of last year, I did an interview with ⁓ Marcus Holmstrom of the gang. They're a ⁓ they're a company that does these kinds of builds for brands in ⁓ in Roblox in particular, but ⁓ Fortnite and other platforms. I'd encourage people to go listen to that episode because it contains a lot of I think insight into the justification behind these kinds of choices. But yeah, that's a that's it's a brand story and a technology platform shift and acquisition story as well, which is always an interesting one to bring us to a close.