Latest / The Interline Podcast / PHP, Generative Tech Packs, And The State Of Clothing Repair
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 a little bit less time today because we do have less stories this week, but we're still gonna talk about all of the interesting stories that really stood out to us. So Ben, how are you doing? speaker-1: I'm jet lagged is how I'm doing, which is also the reason that we're running a slightly truncated edition of this. So I was in in Colorado from Sunday to yesterday and got back quite late in the game yesterday. So you and I only had chance to float a few stories across our collected desk instead of five. So slight slightly truncated version this week. We're gonna be doing three rather than five headlines. but I am here in spirit, even if I'm not a hundred percent here in body. speaker-0: I know we've covered cybersecurity a couple of times already, but we couldn't really ignore it this week since it became a headline in the news that I don't think we've seen before. So I'll let you fill in the more technical details. But in essence, there was a long dormant vulnerability in PHP, which is a scripting language that's used in a lot of web applications. And some estimates that I've seen is that it's part of something like 70% of the web. Now, it sounds like AI had something to do with the discoverability of this vulnerability or the development of exploits for it. And this story was absolutely everywhere this week. It was the top headline in the BBC and other news sources. So tell me what happened here and what you think the implications are for fashion. I've seen some government websites and other big news sites being taken down as a preventative measure. But I don't think we've seen any brands or retailers doing that. We did also see that payments and logistics are being disrupted in separate outages for Google Cloud earlier this week as well, so let me know what you think. speaker-1: Yes, I know we've hopped on about cybersecurity a few times here. We did it last week around physical retail premises. We did it in the wake of the hugging face attack. This is the first time I think that it's percolated into the very top of the mainstream news. And while it didn't directly implicate fashion and retail in the sense that no brands and no retailers took their websites down, it did impact a lot of government services, a lot of other places. I'm talking specifically about the PHP exploit here. Then earlier in the week, a Google Cloud outage was behind Revolute, the fintech slash just recently accredited bank, got prominent here in the UK, along with shipping and and a bunch of payment services and things were taken flying by Google Cloud. I'll put a pin in the Google Cloud one because it I think that's a fairly sort of everyday outage of the type that we've seen before, where you have Amazon Web Services, US East One or whatever it is, that that famous data center that when it goes down takes down huge amounts of the web and web applications like Spotify and things like that. We've seen that before. What we haven't seen before is something on the scale and sudden severity and accelerated like patching timeline of this PHP thing. So Sachi Nadella, who's the CEO of Microsoft, he said this is something that we've not seen in a very, very, very long time. It's not not as long running as like Y2K, but a very, very similar scale of severity and action needing to be taken. Now, it doesn't seem like it led to any big issues or any big breaches or any big hacks, but that's gonna take time to play out here. So PHP correctly, yes, is the scripting language behind a lot of web applications, primarily content management systems and things like that. There is PHP mixed into a bunch of other stuff, but WordPress runs a huge volume of the web. The Interline runs on WordPress. And PHP is the main scripting and execution language behind that. The vulnerability apparently it's existed in PHP for 10, 20 years potentially. The actual ability to exploit it seems to have been expedited. It seems to have been discovered By AI, by a kind of point cloud of different companies. OpenAI took credit for it. Google took credit for it. a company called Wiz, which is part of Google, took took credit for it as well. But you have so you have something that's existed and been an issue for 10 to 20 years, hasn't been able to be exploited. Suddenly, thanks to AI, is flagged, discovered, patched within the space of 24 to 48 hours, it seems to have been. Within that 24 to 48 hour window, though, the standing advice from governments, a lot of kind of web infrastructure providers and things was to take websites offline, to literally just say, okay, do close this down for 24 to 48 hours until the patch is ready and and deployed. In the interim, people use Cloudflare and other things to try and mitigate what. CloudFlash says there's millions of attempts to break into websites from this. Like massive severity issue with like full remote code execution on people's servers and stuff like the highest grade of cybersecurity risks that there can possibly be. This is a testament to a couple of things, I think, for fashion. One, it's lucky that Shopify, for instance, or these kinds of things don't incorporate a huge amount of PHP because Nobody needed to take down their s e-commerce storefronts. Nobody needed to take down their their public-facing Shopify pages or what have you. But this is a vulnerability and an expert that has to be patched like server by server. So if if you work with a big kind of infrastructure provider for your website content, then you're fairly safe in the sense that they will have already done the patching for you. That was the case at the Interline, we use Cloudflare as an interim solution. And then the patch was deployed by our distributed cloud provider within twenty-four hours. If you are the kind of company that is developing your own applications, which is increasingly true now, you have I've when I was in America this week, spoke to a few brands who were like, I've we you know, we've reduced our reliance on software. Provided by third parties for things like market analytics, competitive analysis, those kinds of things for our own homegrown setups and scrapers and what have you. If you have that setup where you're developed something for yourself, you developed a tool, the big issue that people have been saying for a while with these kind of like homegrown or slash semi-bibecoded sort of setups is. They're hard to maintain. And that was feels really abstract when you're putting something into place. When you're like, okay, I've developed a platform that does all my market analytics stuff for me. I've developed a platform that replaces our CRM. I've developed a platform that replaces part of our PLM or something along those lines. How hard can it be to maintain? I have a fixed number of users, I have a fixed set of activities, I know the volume of data that's going to be passing through it at any given time, but surely it's fine. It's fine until you hit a scenario like this, where if any measure of that is built on PHP in this case, or it could be Rust in the future, it could be Ruby on Rails, which Shopify is primarily built on. There's nothing fundamentally different about those except for the Rust is is is a little bit more secure in that respect. All of a sudden, you have full responsibility to patch this yourself. You have full responsibility to close this yourself. So a very weird week in cybersecurity in the sense that. The Google Cloud outage is just a reminder of how reliant everybody is on centralized infrastructure for the web. And then you have the PHP thing, massive story because it can't just be fixed in one central location. If it was an issue with Shopify, if it was an issue with Visa or what have you, it would be fixed in one central place. What you have is an issue in an underlying language discovered and exploited by AI that has to be fixed. in each individual instance. And as AI cybersecurity issues become more prevalent and as brands develop more of their own homegrown solutions without necessarily understanding the full architecture underneath them, you're going to see more of this start to happen. And I've seen people say, this is the start of a terrifying era for the web in general, where exploits are discovered and exploited actively within the matter of hours and You have a s tiny window to defend yourself against them. I don't know how true that is, but this is definitely the biggest instance of this that I've seen in the news headlines in a long, long time. And it's something that should be a further reminder for fashion about the importance of cybersecurity. speaker-0: Next, we have to talk about a really interesting development from the fabricant, who I know you've spoken to recently. Now, we all know that there is this technical skill gap in the industry, and this has meant that European textile manufacturing has shrunk significantly in recent years. However, in collaboration with London College of Fashion, the Fabricant has developed two new digital tools that could prevent things from shrinking further. Now, the first is an AI-powered tech pat tool, and this has been developed in collaboration with a handful of real garment manufacturers. The second is a sewing pattern tool, and this essentially converts a sketch or a photograph of an existing garment into a 2D pattern. Now, both help to put pattern making back within reach of small manufacturers. And they also are trying to help to assist this knowledge technical skill gap. Now I wanted to get your take on what this story means to you, and also if you think that this is a good approach or not to this issue. speaker-1: Yeah, you're right. So I have spoken to the fabricant a couple of times, actually. I did their CTO, Marco Marchese, was on my my other show, the weekly interview show a couple weeks ago. And we did start to talk about some of this, like AI pattern making, AI tech creation. We didn't go super deep into it. And then Adriana Pereira, who is their one of their co-founders, she was on our Mailroom episode a little while ago talking about AI native brands. These are people who know what they're talking about. These are people who understand the industry very well. I've not seen this solution in action. Neither have I yet been able to pin down a demo from companies like Fermat or Flora, like these these the kinds of organizations that are saying, okay, we've done a lot of generative work in visualization and in image creation. We now think we have the foundations to go and apply our learnings around AI and our industry knowledge to generative pattern making to some extent and generative tech pack development. This seems to be something that the industry is converging on. We've we've talked about this before. It just it seems like that's the next step for those workspaces. The challenge is going to be that I don't think there's much about image generation that teaches you anything about pattern making. And n neither is it neither is the other way particularly true. These are these are two fairly isolated worlds from one another. When we talked about Apogee from Lectra the the week before last, I think it was, that's a bit more grounded in this, in the sense that Lectra is a company that spent a lot of time working on product design, development, and direct production. So they have, in theory, a stronger foundation to build some of that stuff on. Now, there's two parts to this. One I agree with and one I'm not sure about. The one I agree with is traditional tech pack creation. which is done in or around PLM and involves assembling manu I mean a tech pack technical specification just it's a PDF that captures all of the essential information you need to produce a product. It's it's the equivalent to an IKEA manual for making a garment, points of measure, sewing operations, materials, colors and so on. In a lot of traditional setups, tech packs require rework at the factory. So when the fabricant talk about Manual tech packs, unstandardized spreadsheets, iterative sampling processes. I'm using words from some of the coverage here. They're not wrong. That is a process that happens. brands put together tech packs, they send them to their manufacturing partners, and their manufacturing partners end up having to either rework them, reverse engineer them, recreate them from scratch, which is what leads to the average sam iterations per style being like three or four at the very least, because It's not an exact science and you started from an inexact brief. Now, anything that could do something about that, like 3D, digital product creation, lots of tools have been pointed at shortening that gap, removing iterative samples and getting everybody talking the same language between creative and technical design. I just don't know that AI is going to solve it. Again, I haven't seen the demos. I need to see that. I'm I'm not 100% convinced at this point in time, but I remain open-minded. What I'm not sure about is what this does for skills, because the reason that tech packs, patterns, and things get batted backwards and forwards between the different actors in the supply chain is that the skills to do these things are rare. And hard to acquire and highly specialized. pattern making is a very scientific and artistic discipline that only a few people learn nowadays, and only a few people really have. That's true at the brand end, and it's true at the supplier end as well. Way more brands than you think don't do their own patterns anyway. They don't own the patterns, they don't own the blocks. They rely on their manufacturing partners, vendors, to do that work for them. I don't know that. Generating this stuff does anything to bring skills back. Again, maybe I'm missing something here, and the generation is a way of teaching pattern making to a new generation or to an existing generation. This to me feels like an instance where you're automating something because there aren't enough people to do it at the volume and to the degree of like accuracy that you need. Great. Interesting use case for automation. I hope it plays out. The bit the gap for me is I don't know what that does to then encourage a new generation of pattern makers. I don't know what that does to to close the knowledge gap or close the skills gap. It closes the execution gap, hopefully, in the sense that if a generated pattern is right first time or if a generated tech pack is right first time, you don't have to sample as much. I don't know how anybody ends up knowing more about the art and the science of either of those things in the process. To me, it seems like you end up knowing less. But time will tell, I think. speaker-0: Finally, we have to talk about a report done by Circular Fashion News on clothing repair. The report explains that there's been an increase in searches for repair services in recent years, and this is because consumers have become more interested in fabric and quality and materials. Also, at the same time, different brands like Arterics and Burberry and Barber are investing a lot more in their personal repair services. And also there's a new V rule in France that could mean that repairing clothes could become cheaper and therefore more accessible. Now, this is all super positive on the surface, but as we've spoken about in previous weeks, the fact that a lot of people can only buy super fast fashion in the first place means that actually repair services aren't really going to be utilized all that much because there's no real point in them. So I wanted to get your take on this report and this story and if you think repairing clothes is going to become more common. speaker-1: And so I'll back this one back to you again as so as somebody who shops way more than I do, you do you make use of repair services? You're also a prime candidate in the sense that you live in London. So a lot of this type the the a lot of the save your wardrobe, these kinds of services, they work best in big cities where there are service partners who can go and deliver on repair and and so on. The the fact that Patagonia and and companies like that have their own repair services as a testament to the fact that there isn't a big repair network out there. How does this play out for you? speaker-0: So repairing my clothes is always something that I've always had on my to-do list and I really want to be more intentional with it. But I've only actually started doing it in recent years, and I've only ever done it with more expensive pieces. So I have some freight jeans, which are a bit more expensive, and they had a hole in the bottom from walking around. And I'm getting those repaired, but I definitely don't think I would care to do it for a cheaper piece, but it is something I want to do more on. speaker-1: Yeah. I w I would tend to agree. I've not had much stuff repaired except for I had a pair of red wing boots a few years ago and I had them resold because there was not that 'cause the sole the the rubber on the the at the sole had worn away. The rest of the boot was great, still timeless style. And so I had I had that done two, three hundred quid a pair of boots. they're not they're not cheap. I wouldn't have a pair of Zara trousers repaired, for instance. it's a and I think that that behavioral pattern does seem to show up over and over and over again, which I think puts a burden on the companies that sell cheaper items to understand that their consumer is gonna recycle and hand back more than they're ever gonna get invested in in repair. I think I think I think the brands understand that. I think the technology and service providers catering to this sector also do seem to understand that. They seem to understand that the market is like higher mid-market premium and luxury upwards is where repair tends to tends to show up. I do think it's a bigger factor. I think the resale market is going to be the primary driver of this. So I talked to when I was on stage in Paris at Premium Vision, talked to Hannah Crump, who has also been on the show here before from Business of Fashion. And she was talking about just the growth of the resale market in general. the fact that financial incentives are usually a driver for a lot of this kind of behavioral change, right? And from that report that you that that you put across our desk, the key bit that jumped out at me was that repair ends repair is valuable if what you intend to do is resell the item afterwards. So I think I I plucked out a quick statistic. Repair, inspection, washing is something like 22 euros per item. and then each item resold for 73 euros. This is all done on average. The majority of them, close to 98%, resold for more than it cost to repair them. Which is, again, like it tends to be logical in most of scenarios that we're used to. If you own a car and you want to resell the car, you better keep it maintained. You own a house, you want to resell the house for more than it's valuable. It's in your interest to keep it maintained and up to code. So I think the the financial incentives and everything there make sense. It's the resale market that's gonna drive repair for me, I think, more than individual people like you or I saying, I've spent a lot of money on this particular piece and I want to keep it and therefore I'm gonna invest from that point of view. Now, I found this report a little bit complicated to read because it's a sub stack with a PDF. And the PDF is made up of sources from other substacks and PDFs and reports which are themselves synthesized from other there's nothing necessarily wrong with that. There's just it's clear that there's a darth of like first-party data here when it comes to how repair factors into what the brand gets out of this and what the consumer gets out of this. Because the resale market that we've just talked about is primarily gonna go through channels like Vestiaire Collective and so on, if we're talking about high value items. Or for lower value items, it's going to go through vinted and eBay and what have you. The brands are not the primary beneficiaries of that higher value that we talked about there. It's it's the consumer that is. It's it's the the impact on the brand that I think I would like to understand a bit more about here. Luckily, you and I are in a position where we can go get some first party sources. So I think we bring Hasner or Medi Save Your Wardrobe and and some other folks onto the show at some point soon and we quiz them. about the the brand impact of all of this because there clearly is one and I just don't think it's properly documented. But yeah, I think light repair is on the rise. I think I will probably do more of it over time, but the b primary behavioral change is going to come from people who want to secure high value items as a store of value that they then transfer afterwards, it seems like to me. I don't know. Do you do you sell clothes on afterwards? I'm curious. speaker-0: I used to do it a lot more. I think I do it less now because I try to buy less and buy better so I buy things with the intention of keeping it. But I can totally get why people would maintain more expensive pieces to one day sell. speaker-1: I'm just I'm just lazy, is the reason I don't sell things on. I think that's that's the primary reason. Maybe I am a target for one of those like wardrobe quantification apps, which will help understand what you have and at some point assign value to it. I do have the CEO of Wearing on my interview show coming up sometime in the next couple of weeks. So let's try and knit knit a few themes together there and see if we can figure out a bit more about how all this plays out. For now, Grace, thank you for accommodating my jet. lag adult brain for for this week, slightly shortened edition. it's been fun talking to you as always.