Episode 30 transcript ·
Accessible and Sustainable AI with Hidde
This transcript was auto-generated and may contain the occasional robot-brained mistake. The audio is the record of truth.
Brad: [0:13] Hey
Hidde de Vries: [0:14] Hey, so good to see you
Brad: [0:15] It's good to see you too. Let's, uh, let's get right into it. What, what has you waking up excited these days?
Hidde de Vries: [0:21] lot of, a lot of things, but I think for me, uh, and we may get back to this later in the show 'cause I know you are excited about this, but it has actually been music. going back into, like, making sure I listen to all the music I should listen to, making lists and, like, figuring out, oh, that person is related to that. And, like, I've been spending a lot of time going to gigs, figuring out which albums to listen to and, like, all that kind of stuff. Like, in this age of, like, where we're trying to automate everything, like, finding out all these cool things that have been made and, and are still being made as well and then being released, uh, to keep tabs on, on the music has been keeping me busy, I think. Yeah
Brad: [1:03] That's amazing. Uh, what, what have you been your, your go-to tools and vehicles for being exposed to music, but also like what are you using to make those lists and, and kind of sift through them and… And, but also like what, what led you to that place where that's been, where your attention's been going?
Hidde de Vries: [1:24] Yeah, I've been using various tools. Like there's websites that list gigs. so when I travel, I travel for work regularly, so I try and find out like what's playing in that area. There's a lot of websites that, that do this, uh, like Songkick and like, like others. They're not always good, so I always try and find out like what are local venues and then look at their websites or frankly like their Instagram, 'cause often, you know, they don't update their websites. and like, yeah, I try and, and use that to find out about sort of live music. Um, and for listening to music, I started moving away from streaming as, as many kind of nerdy people have been doing, I think. And I actually talk to people who are like, "Oh, I've been doing this forever." Uh, like it's been a thing. But I've started, uh, collecting music files instead of, of, of streaming, uh, to make sure that I can listen to it offline and like just have my own collection.
Brad: [2:20] we should get into that, but let's, let's stick with the live music for first, because when, when did you start going to shows just as a byproduct of your work travel? You're already in a place for a thing, and you're like, "I'm gonna go do a gig." Like, like when did that come to be
Hidde de Vries: [2:38] It's been recent actually, like the last couple of years. I was chatting to my friend Dave, who is organizing State of the Browser, a really cool event in London. And he was telling me this, like, "I travel for music." Not just when he travels for work, but he will go to a different city enjoy music. Uh, and in Europe, like you can take a train from London to Paris in like a couple of hours, and it is no problem. Most people wouldn't do it. They wouldn't even leave their own city for music. But it actually makes a lot of sense to, uh, to do that because music is great. And like you'll have fun in the process because you'll end up in or like meet weird people or,
Brad: [3:16] Yeah.
Hidde de Vries: [3:16] It has all sorts
Brad: [3:17] Yeah
Hidde de Vries: [3:17] to just be like, "Yeah, that person is playing there. I live
Brad: [3:21] Yeah
Hidde de Vries: [3:21] but I'm still gonna try and make my way down."
Brad: [3:23] Yeah.
Hidde de Vries: [3:24] he
Brad: [3:24] Yeah
Hidde de Vries: [3:24] me into this, uh, and he still does, 'cause he'll message me and be like, "This thing is playing in Brussels," which is like an hour away.
Brad: [3:31] Pitre
Hidde de Vries: [3:32] Yeah, if he calls, I'll go.
Brad: [3:34] Dave is a, is a gem of a human being, that's for sure. I'm, I've, I, I, I, I will be talking to him at some point as, as well, uh, on this show. But, but I- So it's like there's, there's friendly encouragement, but then there's also this, this kind of deeper thing that I think you, you beautifully articulated, is that it comes in the form of, for me at least, I, I will need to, if I, if a, a show is coming, to the States, I will often need to venture outward because Pittsburgh tends to get hopped over, uh,
Hidde de Vries: [4:06] Right
Brad: [4:07] to… From somewhere between New York and Chicago is a tour stop usually in, like, Philly, but then they'll often, like, sort of stop in Ohio or something. And so if w- we get the kind of scraps sometimes, which is a little sad. But, but traveling for music specifically does, you're like, "Oh, suddenly I'm in a different city, and, uh, we need to eat, and we need to get dressed up." But the, the process of getting there, like, in that moment I'm not thinking that. It's, it's often, like, the chore and I'm, like, looking at, like, a calendar date and you're… I'm just, like, struggling with the logistics of it. But then whenever we actually do it, it's like, "Oh man, this is lovely. This is great. We should do this all the time."
Hidde de Vries: [4:48] And even the chore, I feel like it's more fun than like, "Oh, next week I have to do the dishes." Like, I will not look forward to that. But like,
Brad: [4:54] Yeah
Hidde de Vries: [4:55] "Oh, I'm going to Brussels to see…" Like I did a few weeks ago, went to see Massive Attack with Dave, actually. Uh, they were in Brussels, and it's, it's a lot of fun working up to the moment and being like, "Oh, yeah, we should try and find some place to eat," it makes the whole kind of music experience worth it too, because it's like you're making an effort to see this thing because it is a special thing. I wouldn't
Brad: [5:15] Yeah.
Hidde de Vries: [5:16] that effort for doing mundane things, but like listening to music, yeah. Yeah.
Brad: [5:20] Yeah. and then there is the, the experience itself, which, which there's recorded music which has been shrunken down, commoditized, put into this endless stream. But then there's, there's still the experience of being present in a room full of, of a certain kind of energy and other people and, and having that shared experience and, and that 100% will never be, you know, commoditized or, or like p- a lot of shit around it will be, but like that experience remains, I think, one of like the purest forms of, of life experience where you, you, you just know it to be good. You know that it's like this is, this is special, this is good, this is real. Yeah.
Hidde de Vries: [6:10] get
Brad: [6:10] Yeah
Hidde de Vries: [6:11] in our daily work, and this is very pure. I can't imagine, like, the musicians to have, like, a Scrum meeting where they're, like, doing all these kind of thing, these things that we do. And, and they are helpful, and they make sense for our work and stuff. But like, yeah, they're, really into that. So I appreciate, uh, artists a lot
Brad: [6:29] Yeah, yeah. I- that's my hope through a lot of this, this technological shift, is that we will returning to, to value, the artists and the people who are able to make experiences happen and to, to elevate people and deliver those, those kind of pure, real thing. I think, I think that everything else will become this kind of very amorphous stew with a lot of slop and bullshit and the rest of it just kind of, like, infiltrating and making it very difficult to separate the signal and the noise. And it's like, yeah, going, going to a live concert is all signal.
Hidde de Vries: [7:12] Yeah,
Brad: [7:12] just, like, that's all it…
Hidde de Vries: [7:13] out a lot more, right? If it's
Brad: [7:15] Yeah.
Hidde de Vries: [7:15] signal. Yeah, for sure
Brad: [7:17] and that kind of brings into the, the physical format that you're kinda talking about as kind of, like, migrating maybe away from the, the streaming platforms and into more, like, kind of here are these files and it, and it sounds like you, you kind of appreciate them or cherish them or think of them or, or hold them in a different regard than, oh, I could just, like, search for the song and have it come up. There's, there's, that's, like, a scarcity almost or, or a weight to it
Hidde de Vries: [7:47] Yeah. Yeah, you carefully make sure you get this album and not that album, or, like, you listen to the whole thing. It makes it kind of more… And I started collecting vinyl like, like everyone does at a certain age. It's like, "Oh, vinyl is amazing." I mean, I'm not old enough to have a childhood growing up with vinyl. Um, I think my dad probably had some, but I don't think he ever… Like, he would have CDs. Um, but yeah, it's, it's, uh, it's fun, but, um, it is not convenient. Like, I can't take vinyl on
Brad: [8:18] Yeah.
Hidde de Vries: [8:18] with me.
Brad: [8:19] Yeah
Hidde de Vries: [8:20] so yeah, digital files I've been, been enjoying, uh, a lot. Yeah
Brad: [8:25] That's great. Yeah, and, and in a lot of ways it's, a- and especially where you're, we're sourcing those or it's like, you know, like Bandcamp and the things and the rest of the, uh, the apparatus to kind of support artists more directly still exists. And,
Hidde de Vries: [8:43] Yeah
Brad: [8:44] I w- I was, I was thinking of this thing. I tried to, to vibe code a thing, uh, that was if you're listening to an artist just throughout your week, at the end of the week you get a report. and if you liked a song by an artist if you were playing something on heavier rotation, or you like a song and oh, actually you had, you had already liked a, a prior song by the same artist, that it surfaces their website or their Bandcamp link or their whatever and are, and are like these like off-ramps to it. So it's like the e- the, the convenience of streaming but knowing that these are, you know, fractions of a penny making their way to, back to the artist to j- to try to like just make it easier. 'Cause I, I think that the, the will is there. Like, a lot of people want to support the artist more directly, but it's, it's that friction of, "Oh, okay, well, I gotta like go Google them. I gotta do whatever." So it's like what, what's the automatic way of,
Hidde de Vries: [9:49] Yeah, if you've
Brad: [9:50] of-
Hidde de Vries: [9:50] at the merch stands at concerts, like people wanna like buy the merch and like spend money on their artists.
Brad: [9:56] Yeah
Hidde de Vries: [9:57] Ticket sales are going real quick as well sometimes, like to be there in the first second to get a ticket to, to some gig and stuff. People wanna pay money to their artists and they
Brad: [10:07] Yeah
Hidde de Vries: [10:08] wanna support for sure
Brad: [10:10] and, and with respect to, to the digital stuff, it just feels like there has been just this, this, just a fast current kind of away from all of that. It just, it really took a, it, it took a serious toll where all of a sudden those T-shirt sales and those queues really become… It's like that's, that's where the, the food money is coming from, which, which didn't always used to be the case. So I, I do think that there, there is something to, it's like reintroducing scarcity, reintroducing digital scarcity
Hidde de Vries: [10:46] Yeah.
Brad: [10:47] And, and then making it easy to, to, to, to, to cultivate that and to, to make that happen. It's like, I don't know. That, that, that feels like a fun design challenge.
Hidde de Vries: [10:57] Right. Yeah. And we need it for websites too. Like, how do you discover indie websites, spend more time on them, and make them worthwhile as well?
Brad: [11:06] How do you discover indie websites and make them work well? That's a great question.
Hidde de Vries: [11:11] Um, I just go on the indie web and then I see all these wonderful… I mean, I don't know. I sometimes look for them. I think they are easy to find if you know what you're looking for. Uh, and often you don't find an indie website when you're looking for one, right? A lot of artists don't have their own websites, um, or like restaurants or stuff. They'll be on social media instead. but yeah, I guess like I just read blogs and stuff that I just see links to people's websites or like follow them on social media and I'm like, "Oh, nice. They've got a website, and it's cool." Uh, they've got fonts that I like or like colors or… This is still very much alive, um, and people are enjoying building their own websites. so that's cool. Yeah. I, I think it's, it's great. It is a little hard to discover at times, especially with
Brad: [11:57] Yeah
Hidde de Vries: [11:58] kind of becoming less, less useful at this or trying to keep the traffic within Like, like, um, a lot, a lot of search engines have become chatbots and " Yeah, just stay in the chatbot. We'll give you the answer." but yeah, I'll still go to, to manual websites and enjoy what they look like.
Brad: [12:16] there's still the beauty of the hyperlink in that once you're kind of on somebody's website that is linking out to another person's website or whatever, it's like that, that's a, in my view, another pure experience where it's like, "Here's somebody taking the effort." Like, I, I think about that because it's like you can just make a passing reference to something all the time. I think that there's this, this effort that goes into creating a hyperlink that, again, just, just Web 1.0, just part of the fabric of it. But, but that continues to be this, uh, I think very, very special thing. That's a decision. It's, it's e- it's effort. It's, it's… There's something going into that, that, that enriches not just the body of the thing that you're reading, but that discoverability, that adjacent idea or thought or other person's thing is, is still just, it's such a beautiful technology all these years later.
Hidde de Vries: [13:19] it is, it is. And it still works, which is surprising. Like, I don't know of many inventions that are, they are technology, that actually stay around for that long,
Brad: [13:29] Yeah
Hidde de Vries: [13:29] and still work in the same way and have the same kind of benefits too. Um, I think that's pretty cool. and I actually wish that, maybe my wish for the web is, like, that more people would know how to build something on it, 'cause it is fairly simple technology, but although you need to be someone involved or a bit of a nerd to actually write the HTML. But it was always meant to be like, oh, anyone the street, you could stop them and say, "Here's some HTML. Like, write some HTML." Like, it, it, it's not complicated, but you do need to know a lot of things to actually ship a website. I think we could still make that a lot easier. What do you think?
Brad: [14:06] this is something you're, you're perfectly equipped to, to speak to this because I see There's the indie web movement that I have been, um, around but not directly involved in, largely because despite me doing this stuff professionally, I would be like, "Oh, okay," like, "Let me get web mentions hooked up," or, like, just, like, some other stuff. Like, on, you know, in certain respects I have the core ingredients in place and, and have for a while. But then there's been, like, the other stuff that is just like, ah, this is really, really cool. But it's always kinda felt a little bit like, yeah, here's your computer kit and go build your computer, your Linux-powered computer. And I'm like, "Ah." back to what you were just alluding to, where there's all of these artists, and there's all of these people, and there's all of these, you know, people doing amazing things in the world that have defaulted in the, in the local music venues and the rest of it, have defaulted to social media just because it's like, ah, okay, they, they, they, they made it easy despite, you know, you know, the fact that, that they don't own this stuff. They're not in control. They're not in the driver's seat or whatever. But it's fast and, and frictionless. And so I'm wondering about this, this new age where that HTML know-how, that CSS know-how, um, that, that level of, of plumbing isn't so much of a barrier to people who ought to have websites and, and value them and, and, and have them. I, I'm curious to get your thoughts on that because th- this gets into that, what I see is AI making this stuff that we value more accessible to more people. This is one of the, the hard tensions that I, I'm… This is one of the big reasons why I wanted to talk to you is I think that you've been very clear-headed at being able to sort of speak to a lot of the, the downsides as well as the opportunities of, of this technology. But when we're talking about here's a bunch of great people, local music venues, independent artists, all of these thing, people that should have websites, people that are out there produ- making good stuff to go out in the world, and here is this amazing technology called the World Wide Web that's still working. Hold down the fort. Amazing. And we want to continue to value it. We don't wanna be like, "I, I, like, remember that." And so that needs to stay alive these new dynamics that are kind of sucking the air- less towards the worldwide web and more towards just… It's almost kinda inwards or, or it's sucking things towards platforms or centralization or of, of s- some way, shape, or form. But it, it doesn't feel it's like, "Ah, yes, this is, these are, like, the, the boom years for, for the worldwide web." Or, or maybe, I don't know, maybe, maybe I'm just not in the right places,
Hidde de Vries: [17:11] Yeah. Well, seeing the web kind of go to, like, people who deserve to have websites like restaurants or artists or whatever, seeing them go to some place like, like Instagram where you can't even make a hyperlink. Like, they have this thing where they're like, "Yeah, you go to my link in bio, and then you go to the bio, and then there's like a of links or like services that provide lists of links." And it's all kind of awkward if you think about it. Um, but everyone is like, sure, that this is now our reality. I think that was a step away from the web that we have that isn't great. But the step that kind of is happening now is probably even worse when it goes into the big, chatbot machine where you're just kind of inside a chatbot and get information ways that isn't actually your own words. It's like reused and scrambled and mixed and, and changed. Uh, and it may not credit you. Like is, that is a big worry to me, like that people don't even have any control over how their information is presented. the flip side of it, like using this technology for people to actually make their own websites like you were alluding to, I think that's pretty cool. Uh, like if you don't know how HTML, CSS works, but you can give some prompts and describe kind of what you want, you want that link to be yellow, or you want that thing to be at the top or at the bottom or whatever. I think that kind of thing gives me more hope, like if people can be creative, uh, without knowing the exact technological, um, side of things, I think that is quite exciting. Uh, it is somewhat worrying to me as an accessibility expert, uh, seeing that because I know what could go wrong. also know what could go good or well, so we need to make sure it, it, it happens in a, in a good way or have the right sort of checks and, and balances. Um, but I'm also seeing it on a lower level where it's like- If you don't know how this API works or how to install this CMS, the fact that you can ask a very unclear question and get a very clear answer back, uh, where it goes like, "You need to be here in the docs," I think that is something that of these AI tools offer that also makes web development a lot easier. Uh, where like before you maybe have to go on a forum or write a blog post and then wait, uh, many hours or years, uh, to get an answer, you can now kind of get pointed in the right direction. I don't see answers to be correct all the time, but I see pointing in the right direction as something that's helpful in itself. and like the, the documentation use case is pretty phenomenal. Um, and yeah, I think that that could help people who, who wanna make their own website. It's still also a problem of like, okay, you have a website, you need to host it somewhere, and you need a domain name, and that needs to be paid monthly or whatever. Like, these kinds of structures are just hard for people to understand, especially if the would be something like a social media platform where nobody ever asks you for your credit card. They just want your, of your data and your whole life. you don't need a contract or something in place, and I think that's something that's really hard to, to make easier. Maybe governments should hand out domain names when you're born. Like, "Here's your passport. Here's your domain name. that's your FTP credentials, and here you go." Um, and make it a whole lot easier 'cause it's, you know, it could be public ins- infrastructure.
Brad: [20:37] I, I, I freaking love that.
Hidde de Vries: [20:38] websites.
Brad: [20:39] That would, that would be pretty wild. I'm sure that's been discussed at least somewhere, or at least at, like, a national level probably.
Hidde de Vries: [20:47] I should say I work for the government. This is not an official government announcement. This is
Brad: [20:51] This is, this is, this is you saying, "We should do this." Oh my God. Uh, but, but I do think, I do think that you're right is that there's this kind of, like, last mile thing. I actually just went through this process with, um… So my wife works as, at, um, uh, like a private counseling firm doing sort of therapy and arts therapy, and kinda working on updating their website from an old, uh, Wix-powered website, which again, no shade on them, but it's just like, you know, she's a small business owner trying to run a therapy practice, do good by the community, and it was just, like, very prohibitive for her, and it's like the, her results, because she wasn't, you know, she isn't a professional designer, developer, anything like that, So w- through a few sessions that included us kind of, like, kinda collaborating and, and running quite a bit of that through LLMs to help with this process and to kind of, like, distill her values and her goals and, like, how she wants to be accessible and inclusive and, like, all of… Like, these are critical for us. This is, you know, woven into my DNA. I need to make sure that these things go. To see that kinda translated into an experience was really, really cool. She, she was like, "Whoa, that's, that's really neat." And then we kinda got to the end where it's the, "Okay, cool, now it's time to deploy this thing." I see that as, you know, a huge opportunity to make that stuff just be, "Yep, here's the domain I want. Here's, here's the service. I, you know, just, just make me a website." I, I don't even really, like, care what it… I don't care what CMS it is. I don't care what whatever is, but I, I,
Hidde de Vries: [22:36] Yeah
Brad: [22:37] do, do need it to live at a domain, and I do need it to, like, be live, and if I say I want changes to that website, I want those changes to, to make their way out into the world. It does feel like there's like, that could evolve as well to, to become more accessible, uh, to, to folks and,
Hidde de Vries: [22:55] a lot of plumbing that, yeah, could be taken away or like abstracted away, I guess. Like people wanna pick their own domain name, they probably don't wanna set up DNS or other
Brad: [23:04] No. Yeah, no
Hidde de Vries: [23:06] because yeah, it's hard. But that's the same thing with like in your house, you may have an electrician to help you with certain parts of your
Brad: [23:14] Yep. Yep
Hidde de Vries: [23:15] don't know if I do. I, I definitely do. Like I don't know much to do with my hands, like usually I, I need to hire someone to help me out
Brad: [23:23] Yeah
Hidde de Vries: [23:24] in a house. which I'm a bit embarrassed about, but also like there is so much stuff that they know, like if you drill here, could explode or whatever. Um, I wanna make sure a professional looks at that. But somehow for the web, I feel like we should make it possible for that to work even without professional intervention. Maybe it's a pipe dream, but it would be cool, uh, because we would get a web that has more interesting content,
Brad: [23:49] th- and this kinda speaks to, uh, things like web standards where it's just like, yeah, you have a bunch of people interfacing, creating web code. We want… They would say, "I care about accessibility." They would ca- say, "I care about, you know, this stu-" You know, if you were to say, "Here's what that means," and 100% of people would be like, "Oh yeah, of course. Yeah, like that's, that's table stakes, that's a given," and you know better than nearly everyone that that's a lot easier said than done. But this gets to the point of kind of like smart defaults, and this gets to the land of, of design systems, and this gets to the land of, well, how do you just make the easy and default thing the right thing, the accessible thing, the thing that's built using the sturdy foundations so that anyone can pick those things up and assemble them and have a much better chance? It's not gonna get to 100%, but it's at least not going to be totally butchered. And I think that that matters a lot more than it did whenever we kind of first started talking about this i- idea of like, like a global design system or sort of something that's just like… Having something that's like less a de facto standard that everyone has to go through this gory process to ultimately arrive at the exact same place. It's like, man, just like,
Hidde de Vries: [25:15] Yeah
Brad: [25:15] we just park that stuff in like the right place, hit it hard, hold it up and say, "Here's the, here's the defaults that we, we use for just your basic form fields for your basic like, your basic common stuff." Like, and then, and then of course, yeah, tr- train and retrain these, these LLMs to be sort of scooping that stuff up and wielding it by default rather than whatever the hell it's doing now
Hidde de Vries: [25:42] Yeah, I love that we got to a bunch of that with OpenAI, like as an example, where there are a couple, like the customizable select kind of things. It's great that these will now have accessible defaults. Like, I think the trick in it is that people need to understand that you won't get all of your accessibility for free. You will only get the specific bits that the browser can know what to do with at build time or whatever. And then there's still stuff that you'll need to do, uh, as, you know, someone developing websites. Uh, but just to take away, like, at least the bits that the browser could easily do itself, uh, I think is a huge opportunity. Uh, and I, I hope we'll be able to add more things, but it's great to have it in places like popover and, and
Brad: [26:30] Yeah
Hidde de Vries: [26:31] where it's just like, yeah, a dialogue, making a dialogue that works is easier now.
Brad: [26:36] Yeah
Hidde de Vries: [26:38] still not extremely easy because we need anchor positioning, and we need, like, uh, all of this to work actually well. Um, I was actually, yesterday, I was building a popover prototype for WCAG. We're working on WCAG 3, and I thought one UI pattern there could be to do a little popover to hide some complexity. that is actually hard, and I worked on popover. Uh, but, like, to actually get it to work on all the browsers and do, like, the positioning without
Brad: [27:08] Yeah
Hidde de Vries: [27:08] around with libraries, uh, it's still not entirely there. We're almost there, but there's, like, a little bit of last miles that need to, uh, to happen.
Brad: [27:17] And that, that's just… I can't thank you and the rest of the people who do that work. Like, I, I think that I've always kind of from afar appreciated y'all's work, because it's like it's operating at a much deeper… Like, it's, it's not just the, "Ah, just ship it. Like, we'll figure it out later." There's just, like, so much more that goes into it, and that it is an iterative process, and there's so much coordination. There's so much, like, that needs to happen to get these things that at the end of the day feel Like, it's not like you're gonna get on, like, a giant conference stage with, like, you know, fireworks. It's, it's all this just, just kind of just that hard-earned infrastructure that just kinda shows up and that's just been run through the wringer
Hidde de Vries: [28:02] parts that aren't perfect and stuff because of all the steps that led to there, like popover being an element first and then becoming an attribute, like just learning along the way, uh, what, what would work the best and what would be the right approach. Uh, yeah, it's been a long process, but also really fun to see it happen. Um, and I think we were really lucky with a very smart group of people that, uh, were also partly working on the browser so they could actually implement it there and
Brad: [28:33] Yeah
Hidde de Vries: [28:34] what would happen. And a lot of stuff happened we was all like, "Oh, this won't work," or, "This needs some extra accessibility bits," or whatever. you learn while you do, and you need to get it right because it will ship everywhere and it's really hard to unship, uh, web platform features. Like, they will be there forever
Brad: [28:51] with the hyperlink, uh, uh, fast-forward a few decades and, and it's still doing its thing. And like, uh, like all of these things, it's, it's, they're, they're designed to, to last. There's a, this, this spirit and attitude around kind of like long-term planning, long-term thinking, and that just flies in the face of, I feel like, what we're all experiencing right now, which is just this frenetic hype cycle that has just kind of pulled us, you know, into this very fast-moving current. And I, I, curious to get your take as someone who operates along, you know, as a, as a practitioner doing work, uh, you know, and applying what you're building at the basement at, at the slowest moving kinda pace layer, standards level, all the way up through, and then kinda seeing the, the fastest pace layers starting to like kind of play. How are you kinda navigating this moment and that kinda tension between all of just the, what, what feels like a very real juxtaposition of, of pace layers?
Hidde de Vries: [29:56] it's really hard. But while we're in the moment, it's easier to keep that slow pace because you're doing is at the slow pace. Like, working in web standards is always a slow pace because there are so many different actors that are involved with it. Like the, the World Wide Web Consortium has got, I think, 300 members or something like that. Uh, they all have different opinions and different ways of looking at what they're doing. Like, we've got civil society organizations, and we've got browser makers, and we've got car makers, and, like, all these different people that are working with web technology, uh, that are trying to build AI into the web, but also UI and, like, there, there are a lot of different people doing different things. Um, but yeah, to, to try and get agreement with a lot of different people and, and make it really good, that has to be a slow pace. At the same time, like, that fast-paced world is als-also happening. Um, and we're witnessing it, and I think it's sometimes great to be at the slower pace because you also don't have to jump into every single bit of hype. Like, that's something that helped me in my, like, web career is like, I-I'm usually quite late on the train. Like, when React came out, it took me many years to start using it for the first time, and I don't really regret that because it means that I didn't have to a lot of different versions and a lot of changes and a lot of drama. I could just use it when it was at a stable kind of state, if it was ever there. Like, there's a-- know, it's easier to kind of jump on late, uh, when the AI is done. When we have AGI, I will
Brad: [31:33] Yeah.
Hidde de Vries: [31:34] that train or
Brad: [31:35] Yeah
Hidde de Vries: [31:35] Uh, I'm, I'm comfortable being late because I know that when it's later, it's probably also gonna be more mature. Um, and there's a lot of, yeah, benefit to, to kind of being on that slower, uh, pace too. And you
Brad: [31:50] I think
Hidde de Vries: [31:51] impact, I guess.
Brad: [31:52] Yeah. I, I think people really need to hear that. I think, like, like what would you say to somebody, and sort of specifically thinking of, you know, the person who's gotten started, let's say, in the field within the last five years, l- l- let's, let's say. You know? So somebody that's like still early, still trying to like find their like footing, I think that so many people are just frantically treading water, trying to keep their head above water, that it's like difficult to even think about going like, do you even need to be in this water in the first place? Or, or, or
Hidde de Vries: [32:28] yeah,
Brad: [32:28] could you be wi- like watching from the, the, the safety of the banks
Hidde de Vries: [32:32] So much depends on where you work as well. Like, there are a large part of our industry is, like, moving towards you must use AI, you must do this or follow the hype, and people are afraid to be Uh, I quite recently joined the government, like one and a half years ago or something, and one thing that's great about that is that the incentive is really a-around citizens. Like, we need to make things that are good for citizens, and maybe the parliament will get questions, and then we might need to respond to that. But, like, the general kind of gist of it is you can do things that make sense for the world, whereas if you work in the business, have to be much more focused on shareholder value in the end. Uh,
Brad: [33:14] Yeah
Hidde de Vries: [33:15] even if it doesn't have shareholders, like, profit has to be made. Um, and it is a lot harder. If you're starting out, I, I have had conversations with people who recently start out, and I, I think it is a lot harder. I always think back about when I started out and in a sense, like, it was just a lot more welcoming. Like, people were trying to help each other. You could view source and see, like, what people were doing and, like, um, pace was a lot slower as well.
Brad: [33:44] Yeah
Hidde de Vries: [33:45] so I, I try and, and empathize with people who are new, and I also realize that when I started out, I had it a lot easier. So I, I don't really know how to, how to package that in
Brad: [33:57] Yeah
Hidde de Vries: [33:58] a helpful thing to, to people who are new. Um, I do try and encourage them to of stay calm and try and, yeah, sort of not always jump on every hype cycle if you don't have to. Um, but yeah, it really depends, and I recognize that in a lot of companies, that's not a choice that you have. Uh, I know that people are being counted by the token, uh, these days. Like, you come to your yearly review, and it's like, "How many tokens have you used?" Um, it's sad, I think. Uh, and
Brad: [34:28] sad. It's
Hidde de Vries: [34:29] like,
Brad: [34:30] so sad
Hidde de Vries: [34:32] if you're in that company, I don't blame you for trying to, I guess,
Brad: [34:35] What? Yeah. But that's, that's just it, is like, yeah, everyone will, you know, w- we have no choice but to operate within the, the, the, the systems at a s- at a government level, at a company level, at a society level. It's like the way that these systems are in place, what's incentivized, what's rewarded, what's penalized, like influences human behavior at the end of the day. And when you have this just like fast, frenetic thing that's just been, you know, injected into the bloodstream of every aspect of society, and everybody is just getting… I, I, I just feel this, this, yeah, vibrating, like anxious, fast-moving energy just kind of like infiltrating every nook and cranny across the board seemingly, that it's… It feels very hard to, to paint a different picture. And I think that it, it is important to say that. I, I always think back to like, there's always the, uh, you going back to your React experience, I think, like that was kind of almost like the training ground for this, this latest shift, is like React is certainly, uh, like captured a lot of the attention and like it was well talked about, it was like hyped about, you know, it was, it was everybody kind of out there, you know, pontificating about it. But then it's like when you like zoomed out and you looked at like the, the market shar- share of, what technologies are actually powering the web, and you're like, "Oh, this is like actually pretty small." And it's just like, it just happens to be this very visible, very new, exciting… And again, like it, and it's not a knock on it, it was just, it was a phenomenon, right? Like in a, in a lot of ways. And so we kinda had to discern wherever you were on that, that sort of hype cycle, that it's just like, "Ah, like yeah, do I want to take this bait?" Or, but, but just, to just even be reminded that it's like, oh yeah, like the majority of the web is, is not this, and so maybe this isn't so do or die for me. And same way that all of the stuff that existed before AI hit the scene, all of the technologies, all of the applications of those technologies, all of just the, the opportunities, the, the constraints, all, all of that stuff is, is still there. And then this is just yet another thing. It, it, it, it, it is part of the equation now, but it's not necessarily like the entire equation. And
Hidde de Vries: [37:12] Yeah.
Brad: [37:13] difficult
Hidde de Vries: [37:14] that's an important thing to encourage folks as well, that you don't have to-- you don't always have to do this unless, like, forced by your management or whatever, I understand. But a lot of cases, it can just make sense to not do the hypey thing. I'm, uh, I'm doing a talk on AI and sustainability tomorrow where I do five tips related to AI and sustainability, and I have like a six bonus tip, which is don't, don't use it for the thing. Like, it is okay. And I wanna remind people, like, it is okay to have a use case where you're actually like, "Okay, this is gonna be an Excel sheet," or it's gonna be a couple of lines of Python. It doesn't have to be hooked up to an LLM or like something like that. Uh, and we are sometimes made to feel like that, like we must include the latest tech or we're gonna fall behind. I'm always afraid of falling behind, like when my kids come home with some funny new word that I never heard, and then it's like, "Oh, I'm not cool anymore," right? But we can't run a business like that, right? We can't just sort of base ourselves on like, "Oh, I'm afraid to fall behind." I don't know, like you need to be creative in yourself as well, and sort of, I mean, the artists, we started talking about musicians at the start, like they don't, they're not afraid they will fall behind, I don't think. They come out with cool new albums that are completely different than the stuff before. And I think, yeah, I hope that we, like in the industry also continue to realize that like we don't have to do the latest and greatest, uh, because the latest and greatest might be the thing that nobody's talking about, and the majority may not be actually doing the cutting-edge stuff like with React. Like, yeah, it was a small part, but talked about very loudly. Uh, and a lot of people are just kind of continuing onwards with a website that was much, um, much simpler
Brad: [39:03] It's interesting that the web has been around long enough for us to have some, at least some historical precedent now, and you have this idea of, like, the, uh, uh, like, the, the arc of here's server page, uh, server-rendered, uh, web pages, and then it's just, like… In, in React's heyday it was like, yeah, here's, like, a blank body document. We're just gonna write the entire web page, uh, using JavaScript. And then it took the better part of a decade to go back around and then, you know, not, not going, "Whoops," like we made a mistake, but, "Oh, here's this, like, hot new thing called hot module reloader."
Hidde de Vries: [39:41] in the history, that hasn't actually turned out to be the perfect solution. Like, the perfect thing doesn't exist and we'll, we'll continue to circle around. So yeah, often it makes sense to just, you know, chill and, and do,
Brad: [39:54] Yeah.
Hidde de Vries: [39:55] do it carefully.
Brad: [39:56] Yeah.
Hidde de Vries: [39:56] of
Brad: [39:56] Be-
Hidde de Vries: [39:57] around are, like, built 15 years ago and, and are still kind of going strong. We
Brad: [40:03] Yeah
Hidde de Vries: [40:03] Day last year, I think, and their website has continued to be the same for however many years they're doing it now. Um, and it's fast, it works, it's good. Um, so yeah, many of the cases, like it actually works to keep things, uh, simple
Brad: [40:18] Yeah. Yeah. Well, and, and that, that CSS Day talk was y- y- and you're, you mentioned it earlier that you're talking about it tomorrow, and th- this is one of the biggest topics, maybe the thing that, that had me finally reach out to you, as I was like, "I really am craving and, and need to get some great perspective on this moment, AI and sustainability," to put a finer point on it, because I think that what I tend to see is, like a lot of things, is that things just kind of, end up stratifying into the, the, you know, hard edges of the spectrum. And I think that there's a lot of nuance, there's a lot of, like, here's like the real reality of AI's impact on the environment. But then the, the things that we can and should be doing as, again, more people are just have that, that, as we've been sort of alluding to, this, this pressure to, to use it at all costs. So, so really your, your talk was so great and level-headed and showed some real, like practical things that we can do as practitioners, in that it's again, kind of avoiding that, like all or nothing, like just don't use it. But like, like you were alluding to earlier, it's like, yeah, not everything needs to be run through this. Like, this is yet another tool in the toolkit, and you don't have to reach for it for any and all things. yeah, would be really curious to get your take on like from the time you gave that talk to the talk you're giving tomorrow Like what's, what's changed? Where are your, where's your head at?
Hidde de Vries: [42:01] Well, that one, so I, at CSS I talked about making the web more sustainable. My talk tomorrow is AI and sustainability, so it's a bit more focused on AI. But a lot has changed as well, uh, and I've started to realize more and more that the impact of AI is so much bigger. Like, it isn't like 1 or 2% bigger, but it's like many times bigger, magnitudes more. Uh, and that is fascinating to me, And also if you look at the numbers, like they are expecting 20 to 30% increase of energy use every year because of generative AI. Um, that's scary because we need that number to go down, and we agreed in Paris to have that number go down. scientists are fairly clear on like the number go down, even like a little bit, is going to matter. Every bit is going to matter, and if it goes up, it's also going to matter. It's going to get hotter and like we'll get more, more disasters, uh, which makes it something that is really worth, uh, focusing on,
Brad: [43:01] Yes
Hidde de Vries: [43:01] then worth trying to find solutions for. And there are actually a lot of like practical, uh, solutions too, like where using generative AI, you can use smaller models, you can use more specific models. Like if you're doing a model that's around doing taxes, it doesn't need to know what the capital of France is, um, for instance. Like, there's so much knowledge in these, like we stuck the whole world's knowledge into a model. It probably doesn't need to know most of the things that are in there. A bit like Tailwind, for instance, like it has classes for everything you could possibly do on a website. But if you're not doing most of those things, uh, I used to say this about Bootstrap I guess, like you just get a lot of things that you don't need, uh, but with AI, like that is, uh, that is a lot bigger.
Brad: [43:48] Yeah
Hidde de Vries: [43:48] there's a lot of optimizations people can do, like they can use smaller models, they can use more focused models. Um, and that is, that is going to help. Also, distinguishing the different kinds of AI is helpful. Uh, generative AI is a lot bigger, magnitudes bigger in terms of, uh, CO2 emissions than what they call traditional AI. Like if you're at Apple Maps and you're trying to find out how to get home, is using AI, it's not generative AI. It's like way finding type of AI. When I, I studied AI in university like 15 years ago, one of my first assignments was do the quickest route between A and B, and it was the kind of thing that we had to work on then. and that is, is definitely AI, but it's a different type, and it also uses a lot less energy than
Brad: [44:37] Yeah.
Hidde de Vries: [44:38] stuff.
Brad: [44:39] Yeah
Hidde de Vries: [44:40] can be like using a helicopter to go to work. That's kind of using way too much, uh, effort to do something that, uh, maybe doesn't have as much value. Um, so yeah, finding the right focus is important, like what, what sort of things are actually, uh, impactful. Um, and, and yeah, using the right definitions is important, uh, because there are a lot of artificial intelligences that are all kind of different. Uh, and the same kind of goes for sustainability. Like there are a lot of angles to, there is the impact on the planet, but also in a lot of definitions of sustainability, we also look at social impact and profit impact. It's
Brad: [45:19] Yep
Hidde de Vries: [45:20] that perfect balance between people and profit and, what's the last one? The planet. Yeah. So people, profit, planet is the three that we often use. Yeah.
Brad: [45:30] as you were sort of talking about like the different models and a lot of that stuff, I thi- I can't help but think, it's like, boy there n- there needs to be more affordances. Like right now all of this is ju- it just shows up in a, in a select drop-down of like equal weight. It's like you almost want that, that like, you know, it, it- the key to your helicopter should be like the keychain should have a giant coconut on it, you know what I mean? You're, you're like, "Th- this is cumbersome. Like, like I'm not going to reach for this by default." But right now like everything is just like showing up. It's just, it's all flat and, and behind that are some like real decisions that, that
Hidde de Vries: [46:14] Yeah
Brad: [46:15] im- like you say, like steer things from, you know, going up or, or nudging it downwards and it's like that, those are design, those are design things. And like that obviously, you know, a lot of the, the design of the, you know, the interfaces that people are using with it, it's like that, that, that stuff is fascinating to me. But yeah, I'd be, I'm curious if like the, you've seen anything that's just like done a good job at like kind of injecting or, or kind of surfacing more of that kind of like impact to people in, in their, in their daily rhythm. 'Cause it, it's, it's, it's hard to… Shit's hard enough to just conceptualize in our minds and then like we kind of like, again, under a bunch of pressure, hurry, hurry, use this, whatever. And so th- there's just like a lot of weird tensions that I think are tugging at people and it's hard to be like, "Here's this thing that's an abstract concept. We're not good at thinking in that, at that scale at different time lengths." And it's like there's gotta be some like ways that we could just kinda like build that stuff into the, the fabric of using it.
Hidde de Vries: [47:30] Yeah,
Brad: [47:31] don't know.
Hidde de Vries: [47:31] a great idea there 'cause a lot of what makes it so hard to reason about this is that like AI is presented as like this abstract concept or like a dropdown of like choose this model or that. But kind of behind that is like this huge physical infrastructure. Like I drove past one of the largest Dutch data centers the other day, and it's huge. Like you can drive past it for minutes. It's very big. and that kind of physicality of what we're actually doing, uh, yeah, it's, it's very easy to hide that and, and not show it. luckily we have capitalism because this costs a lot of money as well, so like the sustainability impact is often related to spend in terms of money. Uh, it costs
Brad: [48:11] Mm-hmm.
Hidde de Vries: [48:12] money to run these systems because they're so big. Uh, and eventually if people run out of VC money, things are gonna get more expensive, and then we'll have our nudge, uh, that way because it will just cost too much to, to do these things.
Brad: [48:26] Yeah
Hidde de Vries: [48:26] really seen great UIs that, that solve for this. Um, very few, like some of them will have the word mini next to a model when you, when you pick it, so you can see that one is smaller. But there is very little explanation of like there are a couple of LLMs that will tell you the CO2 output or like CO2 emissions associated with your query. but that kind of thing is, is still really rare. It's also really hard to measure because a lot of the big AI companies don't wanna
Brad: [48:55] Sure.
Hidde de Vries: [48:56] what
Brad: [48:56] Yeah
Hidde de Vries: [48:57] usage is. They recently lobbied the European Commission successfully to get that data, uh, marked as, uh, business sensitive so that they don't have to share it. and that has now become part of the European, uh, regulation. Uh, they don't wanna share it, and that's a problem because we need that information and like to have it in the right level of detail so that we can actually build UIs that maybe, describe that kind of thing
Brad: [49:23] Yeah. Yeah. that's the hard rub is that it's, like, same as it ever was. You have companies doing their best to try to obfuscate and skirt responsibility, and, and that does require us, our collective investment and, and holding feet to fire and doing what we can in our own lives. But i- it's, like, it… The, the corollary to things like recycling feel apt in certain respects, right? And, and that was, like, the, the reduce, reuse, recycle stuff was all kind of PepsiCo, you know, kind of like, "Oh, we're just going to pass the buck on- onto here." And, and don't get me wrong, it's like, yeah, we absolutely should be reducing, reusing, and recycling. But it's like at the end of the day, if they're able to, like, kind of continue to, to operate in, in these ways, it's, it's… Our individual actions are going to be a drop in a bucket of, of, of a bigger thing. And so I, I guess, like, like, are there organizations or, or, places that you've seen doing good work at kind of, like, that kind of advocacy for sustainable AI development and things that, like, people can, could either, like, contribute to or, like, sign up for or, or educate themselves on?
Hidde de Vries: [50:49] Yeah, there's a lot of different ones. Like around the web, I love the Green Web Foundation. They do a lot of great work around making the web more sustainable and also share a lot of research that they're doing. There's also the Green Software Foundation that are, I think, friendly with one another. Uh, they are more focused on, on software at large. Uh, and I believe they also have some stuff around, um, around AI. Uh, and there, yeah, there are a lot of like different organizations working on, uh, yeah, sharing data and then finding out how can we start measuring what happens in data centers and make it more available so that people can make more informed decisions about this sort of thing. Uh, and that's very much needed. Um, and yeah, great to see a lot of innovation in that, that area as well 'cause it's gonna help people make better choices. And I think they're great because like you say, it's gonna be up to individual to recycle, but if your individual choice is something you do on behalf of a huge corporation, you can suddenly make a huge impact with that. Uh, because if large companies decide to use smaller models, it can reduce a
Brad: [51:57] Yep. Yep, yep. Yeah, and I think it is important for people to understand that, that like wherever they're sitting in a, you know, wherever you work, you, you are in a position of influence, right? If you're, if you're like paying attention to this stuff, if you care about this stuff, it's like you're in some way, shape, or form in the driver's seat. Like, usually we've been the same people that we have our family members saying, "The printer's broken," or, "Please fix my, my email or my Facebook login," or whatever.
Hidde de Vries: [52:32] Yeah
Brad: [52:32] like these are the same people that are kinda now going, "Okay," like, "Here's, here's AI." And I, I think that there is, there is a, you know, we have a, a responsibility to, kinda steer people and like help guide things to- towards more like re- responsible choices, and that's, that's education advocacy.
Hidde de Vries: [52:52] all, you are all bad people or whatever. It, it's we need to understand that these people are trying to do the right thing and actually give them tools to understand how they make better choices. it's the same with accessibility and something that we're learning constantly. Like, we cannot go, "You're wrong." Uh, especially like if they're individuals, uh, we need to give them the right tools to get it right. Um,
Brad: [53:15] Yeah
Hidde de Vries: [53:16] that's a fine line. Like, if people willfully ignore what they know and, and do a bad job when they should have done the right job, okay, we need to regulate that, and luckily we are for accessibility and sustainability. but yeah, giving people the right tools is usually the most productive way as well to kind of bring that change about. so we need better defaults basically. Um, and I hope they're, they're coming.
Brad: [53:40] that corollary to accessibility is, is dead on, and that was always my frustration with accessibility in general, is that it's just, like, my, my read on it, and tell me if this is right or, or, or wrong, but, like, I, I've seen it get so much better over the years, which is great, but I, I think that there's, there's this kind of very noble cause of accessibility. The web is built for everyone irrespective of their ability, and we have an obligation, and that translates to a lot of passion to ensuring that that happens, and then whenever people fall short, it, it, it, it, that passion comes out in the form of, of, like, "Let me yell at you pretty good. And, and if I yell at you hard enough, then you're gonna get it right and care." And, and I, I, I've alwa- I always found that to be kinda counterproductive, and I feel like there, you know, a lot of people, you know, mutual friends and yourself included and stuff I think have done a much better job at kind of, like, making it just feel more like this, this is something you, you want to be doing. We are all here. We're all… You know, there's, you know, progress, not perfection, et cetera, et cetera, et cetera. And I think that that, that kind of healthier attitude is, is welcome and needed
Hidde de Vries: [55:01] It is really important to stay, stay positive. what makes it so hard is that it's a, it's a human right. We, we need equality. Uh, people have the right to have the same access. Uh, they're not getting it, and sometimes for the wrong reasons, because there's not enough investment or people aren't paid to do the thing properly. Um, so I, I, I fully understand the frustration that practitioners have, where it's like, you know, this should be accessible. You're paid a lot of money to do this project, and you forgot that important bit where it actually affects my friends or myself or, you know, people in general because they have the right to get equal access. Um, but yeah, getting there is usually, like, the path is usually positivity. We have to stay positive if we don't like it. Um, and there are many times where I don't like it, where I'm like, well, know, you are awarded this amount of money to execute this project. You cannot go, "I forgot, sorry. Uh,
Brad: [55:59] Yeah
Hidde de Vries: [56:00] the accessibility." Um, need to do it. But yeah, we need to, as an industry, make that a positive thing because it's the only way to make it happen, I guess.
Brad: [56:11] and, and frankly, like, what I've seen to be the most successful, I'm curious to get your take on it, on it as well, there's something quite damning about check boxes. I get very leery of check boxes, and accessibility has always been one of those things where it's just like, yes, there are these heuristics, yes, there are these, like, very specific check boxes that need to be checked, but unfortunately, that ends up translating into, " Here are these boxes that we need to, to hit," versus something that feels a lot healthier and just more positive and is, is just, it's just kinda weaving it just throughout. It's just, just part of the language and part of the, the ongoing design and development process. It's, it's just like, it's not a, a separate thing, and certainly not just, like, a check box, but rather just, like, something that we understand to be part, an inherent part of making digital things, that it's just a quality of it that we internalize and use. And, like, the healthiest projects I've worked on are the ones that, that just have internalized that and, and, and it feels like more of, like, a… just, just an understood or a v- a value that's lived versus this just fucking chore. I don't know.
Hidde de Vries: [57:41] Yeah. I, I get the point. I mean, I'm one of the editors of WCAG 3, so technically I am a, a checkbox producer.
Brad: [57:49] You're the
Hidde de Vries: [57:50] checkboxes.
Brad: [57:50] checkbox guy.
Hidde de Vries: [57:51] I
Brad: [57:52] But they, no, no, no, they, they need to be. But I, I, I, yeah, I, th- they
Hidde de Vries: [57:55] I
Brad: [57:56] need to exist
Hidde de Vries: [57:57] Yeah, I think they need to exist, but the other hand, like, it is not about-- it shouldn't be about checking boxes. It should be an internalized thing. Um, but it's, it's really a bit of both. Like, if you, your organization says, "We've internalized it, and we've done a great job," I can use my checkboxes to see if it's true or not. Like, I can, I can double-check it with that.
Brad: [58:20] Mm-hmm.
Hidde de Vries: [58:21] but they are, they are basically like abstractions of, like, actual accessibility Like, the people making the checkboxes, it's all based on research, and it's based on,
Brad: [58:32] Yep. Yeah, yeah.
Hidde de Vries: [58:33] that are helpful and stuff like that. um, yeah, it becomes checkboxes because, you know, regulation needs to take it up, and, like, there's,
Brad: [58:41] Yeah
Hidde de Vries: [58:41] Uh, but it's, it's really a bit of both. Like, yeah, it's helpful to have checkboxes, but in the end, like, if you're in a team, probably not the, the only thing you need to do. Um, especially like understanding how people use the web, I think that's the main If you do
Brad: [58:57] Yeah
Hidde de Vries: [58:58] it's like you need to understand not everyone uses a mouse. Some people use a keyboard. Like, if you have that basic understanding, then the checkboxes will come because you will, you will meet them. but yeah, it's more about how people use the web. That's also how the W3C about it, by the way. Like, they have a page that, explains that. Um, and I've always found that one of their most useful pages where it just explains this is how people use the web in general. Like, these are some different ways. Uh, and if you know about those, it's harder to make things inaccessible
Brad: [59:30] That's right.
Hidde de Vries: [59:30] you need to,
Brad: [59:31] Yep. Yep. Yeah. don't get me wrong, like, the, the check boxes are, are very necessary. It's the, it's the, the stopping there, and I think that you just did a, a beautiful job at, like, kind of the… There's, there's the framing, there's the storytelling, and then ultimately, like, the, the kind of more just cultural transmission and, and
Hidde de Vries: [59:51] Yeah
Brad: [59:52] empathy and, and just, like, a, a, a greater awareness that you're able to just kind of bring to your work that ends up making, surprise, surprise, making the work a lot stronger. And I, I think that that's, that's… It's, you operate from just, like, a healthier and more proactive place rather than this, like, whole like, "Oh, shit, we just made, like, a big mess," and then, like, we gotta go in and fix the check boxes after the fact. Yeah.
Hidde de Vries: [1:00:18] I do,
Brad: [1:00:18] Yeah.
Hidde de Vries: [1:00:19] accessibility audits every now and then, and then I'll have the chat box where I write down failures, and it's like, this fails, that fails, that's a failure.
Brad: [1:00:28] Hmm?
Hidde de Vries: [1:00:28] negative activity in a way, and, and ideally you would just take an organization along and be like, "These are-- is how to do it." Make it positive. Like, even in, like, the process of auditing, there's a lot of improvements that people can make to making that not about check boxes or failures, but making it about, here's a code example. Dump this in here because of this reason. Uh, this is gonna make it keyboard accessible. It wasn't because of that attribute missing or whatever.
Brad: [1:00:57] Yeah
Hidde de Vries: [1:00:58] yeah, that can also be a big step in, like, positivity and, and making things actionable to
Brad: [1:01:04] I'm curious to get your take. So it's like we're on, we're through l- one lens of how AI is impacting user experience, which is okay, we're, we're creating this stuff, we're producing code, and we could, you know, point these things at our code to say, "Is this accessible?" Or like, what are some, you know, what are some things that we can be doing to, to improve this stuff? And, and that's one lens of it, and then the user experience comes out in the form of a thing that people interact with, and we guide people through flows and stuff like that. But one of the interesting things that's … I feel like hasn't gotten as much sunlight is that, oh, well, the users of our websites and apps and all of that stuff also have these tools, and that changes the game, especially, I think, through the lens of, uh, accessibility. But let, let… I'll paint a picture here. I'm curious to get your, your, your thoughts on it, because in two hours or something, I created a, a Chrome extension that's basically like a supercharged version of like user style sheets from back in the day, right? So it's like, oh yeah, like any time I'm visiting, you know, Amazon.com, like I just want to hide literally every part of the , the, the webpage that isn't just the product image and the, how much the thing costs and the basic description of it, and stores that to local storage and st- you know, there it is, and then any time I go to that webpage, I don't see any of the cruft. If I go to another webpage, a- another domain, I could change it to, you know, neon green and if, if that's my, my desire or, or bump up the font size. And what I'm very fascinated about is the, this new generation of accessibility for people that tend to get the short end of the stick in general because people like us aren't, aren't considering them as w- as well, but also tend to get like flattened into, to buckets, right? Which is like, oh, y- you know, you're a wheelchair user, you're, you know, blind person, and it's like obviously all of these things are, are spectrums and people's specific, uh, you know, disabilities or, or operating, uh, capacity or whatever, short-term, long-term, everything in between, is so hyper-personalized, So I'm curious, like, what your thoughts are on that, or, like, been going on there with respect to, like, how does AI unlock maybe a different type of accessibility for people, or hopefully a better user experience for people across the board?
Hidde de Vries: [1:03:58] Yeah. In general, I believe browsers a lot more to fix accessibility on the browser level. Like, we've been adding focus styles to websites for years, but a browser could just draw a focus ring on whatever has focus on behalf of the user, right? If they wanna have focus, that should be something that-- That's one of the checkboxes in WCAG that we could throw away if
Brad: [1:04:22] Yeah
Hidde de Vries: [1:04:22] that functionality. Um, I had a small, like, we had a small group at W3C that worked on this for a little bit. Um, we, we stopped doing the work for, for various reasons, but there is a list that we can make of things that browsers could allow for. Also, like, killing elements, like you were saying on, on Amazon. Um, just like, I don't want that element. Just I don't wanna see it, or I want it to be bigger or smaller. There are things, like without AI even, that browsers could do on behalf of the user. Um, I think they should be doing it 'cause they are user agents. They should be agents for the user, uh, and do on behalf of the user, do whatever the web designer didn't want. Like, the web designer wanted huge fonts or, like, some light colors on light colors. The user might not want it, and the agent can do that for them. Um, so You know, bring the user back into the user agent.
Brad: [1:05:17] Yeah
Hidde de Vries: [1:05:18] I think there's a lot more customization that can happen. Uh, I've been thinking about this in terms of alternative text, for instance. Uh, we all do our best to write down an alternative for an image, uh, the way that we intended the thing. what I often find when I talk to blind friends is that they may be interested in very specific things in the image that I didn't think of. like if you're shopping for a dress, and I'm giving this example 'cause I heard Leonie Watson talk about this a few weeks ago at a conference. She was saying, "I'm shopping for a dress. I want that dress to be black, so I wanna check that it's black." And she had a couple of other things that I don't remember, but there were some specifics that she wanted to know about, uh, and she could ask her AI if they were in that image. Uh, and it was great. Uh, and the original author of that webpage could have never written down those details specifically, um, because maybe the next person will want different things about a dress or find different things important about it. Um, so I think that sort of, like, personalized alternative text is just one example where some system that can describe what's on an image is really helpful, uh, and can be customized in a way that would be impossible for any human to, to write down, because it
Brad: [1:06:35] Yeah
Hidde de Vries: [1:06:35] much on, on the use case. Uh, maybe you don't wanna know about the color of the dress. You wanna know about the material, or you wanna know if it's suitable for evening or whatever people wanna know about dresses. I don't, I don't-- never buy dresses. I don't know. I always pick the wrong ones for my daughters. Uh,
Brad: [1:06:53] But, you're onto something there because it, it's like there, there's fa- there is fashion and there is a whole language around fashion and dresses. I- I don't, I don't know dresses very much either, but I, I will say my, my first job whenever I moved to New York City was working for a fashion e-commerce company, and we would sell these, these multi-thousand dollar dresses, like boutique dresses, like d- you know, designer things. And that, and that wa- it was fascinating because it's like, "Who's gonna spend $5,000 on a dress?" Turns out people spend $5,000 on a dress on there. And, but, but there's, there is so much, like, specific language that goes into, you know, a garment and, and that is very, very difficult to, to, like, distill into, like, a single aesthetic thing. And I, I do think that we, this new technology does unlock the ability for these experiences to be a lot less inert than they have been to date, that for me is particularly exciting. And that's all, of course, that the raw ingredients, the, the, the default, the standards, all of the, you know, tho- those checkboxes, all of that stuff, like, kind of needs to be there, and then there's this, like, layer in front of it that just, like, can supercharge that accessibility and actually, like, take it from, like, a, "Yes, you can survive this experience," to, like, "Oh, this is freaking lovely, and this is, like, tuned for me, You know, my cousin's, uh, quadriplegic and has, like, the nature of her accident, she has, like, kind of, uh, difficulty with, with one of her hands. Like, like, she, she can use it, but it's like the, the proximity of, say, like, the, the primary call to action button, which there is no standard for, is, is, is hard to come by. So it's like the, th- just to even think about, yeah, whether that's additional browser settings or whatever that says, it's like, I want the primary call to action button for these types of pages, if I'm shopping or sort of something like that, to, to show up in a place that is comfortable for me, that is, that is more reachable for me or so, so… Like, that stuff is super cool. Like, I think that that just feels like, oh wow, that is like- Could make a real material difference in a lot of humans' lives.
Hidde de Vries: [1:09:22] Just gonna make sure people can do the thing they, they wanna do. this is also part of, uh, voice control interface on, on the Mac. You can have, like, a grid overlay and say, like, it will put numbers on all the elements, and you can just say, "I wanna press element 15," where number 15 appears, like that sort of thing. Uh, it's great. Like, yeah, it also makes people less dependent on how an individual UI is designed.
Brad: [1:09:47] Yeah.
Hidde de Vries: [1:09:48] I
Brad: [1:09:48] Yeah
Hidde de Vries: [1:09:49] great that people design websites, and all the respect to the people that are doing that work, like, like us. Uh, but also they are lower priority than the users who just maybe wanna do their thing or make it bigger or smaller or whatever. so it's, yeah, it's gonna give people more agency and, um, I think that's great. Yeah.
Brad: [1:10:09] Yeah. That's, and again, it's, it's, it's the complicated stuff around this is like you could kind of like see all of this promise and potential and very real improvements in, in, in people's lives. And then there's also that, that balancing act with the, uh, you know, our responsibility to make sure that it's not just we're just YOLOing it, us, us all into Armageddon
Hidde de Vries: [1:10:36] Yeah, and that's the thing, like, from an ethical point of view, I think, like, if you're weighing these different values, it's like if someone needs this as an assistive technology, they-- that's a different thing than someone wants it because they just wanna see what it looks like if you put an elephant's head on a tiger or whatever and, and make an image generating based on that. Uh, that's a real, like, tangential difference to, uh, tangible difference. Yeah.
Brad: [1:11:02] Yeah. Yeah, yeah. Yeah, and it's like, and it, I think, uh, kind of bringing it all home to, to just like people's… But, you know, we are able to make our, make decisions, and so it's, like, kind of making the right decisions of, like, it's, like, kinda using… For something that, that could be just, like, a, a simple search or a query or a little sort of script or writing two lines of code or whatever it might be. It's like, you know, may- maybe err on that side and, and not just kind of throw the helicopters at it, uh, by default. I think that, I think that's, like, really kinda pragmatic stuff that we could be doing. And to, to employ this technology to sort of help, you know,
Hidde de Vries: [1:11:44] Yeah
Brad: [1:11:45] I think, kind of advance the, the promise of, of the web. I, I, I… Maybe that's, like, a good kind of thing to point it at is, like, it's like, okay, like, yeah, if we're talking about pointing this thing at something, let's make it, let's point it in the direction of betterment. And we've done… You and, and, and your colleagues and everything have d- have done, like, a really great job at defining what that betterment is and what those standards are and, like, what we're all sort of, like, striving for, and I think that that, that feels like a good thing. A, a, a relatively good application of this technology is like, "Hey, how can we just make sure whatever we're producing, with or without LLMs, is, is going out the door and is, you know, sturdy and resilient and gonna last a long time and is accessible for anyone who's, who's accessing it?" the only last question, we started our conversation talking about music. uh, what is some music that you want more people to know about?
Hidde de Vries: [1:12:51] Yeah, this is such a hard question because there's so much great music out there. something that I've enjoyed recently, and this is Dutch, so it will be really hard for anyone listening who doesn't understand Dutch, is a group called Iceland Uh, they are named after the country, but it is a metaphor for the Netherlands being a country that's too cold towards, for instance, immigrants, uh, making immigration too much of a problem and, like, closing borders. Uh, so they call the country IJsland, um, and they make all sorts of somewhat silly music, but also quite serious about, you know, what is AI doing to why capitalism is bad. Uh, they call capitalism the C-word. Uh, so yeah, lots of, lots of fun, um, in that music. It's kind of like punk hip hop. I've enjoyed that. Um, and, uh, if I can name two, I also really like James Blake's latest album called Trying Times, which is about, like, of love and, like, lots of, lots of good themes. but also I think he's technically, like, a great musician, uh, doing magic with instruments that I don't understand. You probably do 'cause you're a musician too. I don't know what he's doing with his synthesizers, but, um, I think it
Brad: [1:14:09] He, he is doing a lot of crazy stuff right out of the gate, but I, I think that the thing that impresses me most about him is he's able to really convey a mood better than nearly anyone else that I've ever heard. It's like, it's like, oh my God, uh, you, you feel something listen- listening to that. But I'm sure Iceland makes you feel something as well, and I, I will give that a, a listen. That sounds like some good, like, good, like, social commentary and a good, like, kind of bookend for, for our conversation and our responsibility as well, so that's great. Fantastic. the last thing is where can people, follow along with, with you and your work, and where are you hanging out?
Hidde de Vries: [1:14:52] easiest is probably hita.blog, which is my website. Uh, I also am on various platforms with HDV, uh, because my first and last name are a bit tricky to pronounce, so I just stick with the initials. HDV is where I'm at most of the time
Brad: [1:15:10] Beautiful. Beautiful. Well, thanks so much for taking, uh, taking the time to chat, and, uh, go enjoy the rest of your, uh, your evening over there, I suppose, right?
Hidde de Vries: [1:15:19] Have a great rest of your day, Brad.
Brad: [1:15:21] All right. Take care, man. See ya