Transcript of Interview With Luke Wroblewski

Interview With Luke Wroblewski

This interview features Luke Wroblewski, managing director at Sutter Hill Ventures. You can watch it on Luke’s profile page.


– Hello, again, I’m Joe Welinske, and I’m the conference director for ConveyUX. And we have our next event coming up the last three days of February. And it’s gonna be in-person in Seattle and online. And one of the fun things I get to do is talk to the many speakers that will be at the event. And today I am pleased to be speaking with Luke Wroblewski. Hello, Luke, how are you today?

– I am doing quite well, hello, hello.

– Well, I’m talking from Home Office in Bellingham, Washington, which is just North of Blinks Headquarters in Seattle. Where are you talking to us from?

– I am in Palo Alto, California, just about 30 minutes away from my home in Los Gatos, but Bellingham would be a great destination for some mountain biking, so we’ll have to make that happen one day.

– Well, yeah, it certainly is. And I enjoy taking advantage of kayaking around here and hiking and the biking as well. So you’re definitely right about that. Well, yeah, I’ve been fortunate to have known you for quite a long time. You know, one of the things I’m able to appreciate from a distance of your work is, that you’re very prolific in your sharing of ideas, you know, talking at events, but also, the books that you’ve written and you regularly contributing, you know, online articles on a regular basis. I haven’t been as caught up with you recently, but for people that may not be as familiar with you as I am, why don’t you talk a little bit about your background, some highlights of that and talk about what you’re up to now.

– Yeah, these days, I feel a little like a dinosaur because when I talk about like how I got started and all this stuff, it was in the mid-90s when I was working at the National Center for Supercomputing Applications. And the big project over there was NCSA Mosaic, which turned out to be the first graphically, the first widely distributed graphical web browser. It turned into Internet Explorer, Netscape Navigator, and then away we go, right? So back then, I was building products for research scientists ’cause that’s what the web was for. And from there, went over to eBay, went over and spent time at Yahoo, kind of got to live the whole web RUSH, Web 2.0, RUSH Mobile. And then I went and did a couple of startups, sold one to Twitter, sold one to Google, then hung out at Google for a good number of years. And now, I’m kind of back building new things, but building companies as opposed to just products. So I work at a firm called Sutter Hill Ventures and we do a whole bunch of incubations of companies where we’ll start them out from zero and try to get them to grow all the way up to becoming big great companies. And so, it’s kind of what I’ve been doing all along, but yet in a different context, which is fun. It’s nice to change from academia at NCSA to growing internet companies doing the, during the boom to startups to this.

– Well, you did a surprisingly brief encapsulation of what I know is quite a bit of work over your career to this time. So it’s so good on you with that. But there’s definitely a lot packed in there that we didn’t have time to get to. But let’s talk about the topic that you’ll be presenting at the Conference. So the title is “How AI Ate My Website.” So you know, why don’t you tell us a little bit about how that topic came about, and what we can expect to learn from it.

– Yeah, well, maybe that’s why I was brief in the first part so I can sort of tie this part into that. As I mentioned, I’ve kind of gone through, hey Let’s get everything on the Cloud. And I don’t think it’s any surprise to people that nowadays the chatter is all about let’s make everything AI. And so, what I did in the past around all those other people will call ’em like platform transitions or what have you in computing. But when really, dramatic new capabilities came around, there’s a process of learning what they can do and should do and how they should do it, which I find super fun and interesting, this type of person who always wants to be learning. So what better way to learn than to try and hit the Reset button multiple times and you start to see similar patterns, right? So when the web came around, everybody was trying to pour desktop apps to the web and then when mobile came around, everybody’s trying to pour websites to mobile and this pattern just repeats over and over again. So in the early stages, when these things, these platform potentials come out, there’s this tendency to just repeat and rehash and try and port and you don’t really get to see the true potential of stuff unless you dig into it, start playing with it, try to build things, try to design things. And that’s when I feel like you get opened up to the new ideas. It doesn’t happen instantly, right? And there’s no guarantee that that happens. But being in the thick of it is where I find those opportunities kind of show up for me, and to come back again to some of the early intro stuff. You know, that’s why I do all the writing you described at the beginning to kind of think through this stuff. Half of that is just discipline and making myself understand things and then write it down. And since I’m already writing it down, I might as well, share it. And if I share it, maybe somebody sees it and asks me a question or points me at something different, that expands my learning. And so, I’ve been doing the same kind of thing with this AI transition now. there’s also the benefit that I’m no longer a Google and I can do a lot more public stuff than I was doing before. But that’s the process I’ve been going through is just getting inside of this technology, trying to build things with it, trying to understand what it does, how it does it, and then when I learned things, putting them out there in the shape of articles, talks, like we’ll be doing in February and other formats. So the talk was a long-winded wind-up to the talk description. But since I cut the other thing short, the topic is, I have over these years built up 2,000 articles, three books, like 370 presentations, 100s of videos. And so, all that stuff sits on my website, like it sits on so many other websites on the internet, right? And if you wanna find some information from me or from anyone else, what can you do? You can like type keywords and hope something comes up that’s easier to do when the entirety of the internet is your canvas. When you’re on a specific site, it’s a little bit more hit or miss. You can browse stuff, maybe it’s here. Maybe it’s here. Maybe it’s there. And you can kind of make your way through and try to read and understand what’s going on. And the other thing that you can do in real life is you can say, “Hey, Luke, what do you think about NCSA Mosaic’s design?” Right, and I would think about what I know about it, process some thoughts, and respond to you. And with some of the AI models and the generic technology we have right now, we can actually get pretty close to that last modality, right? So what I did, this has been almost a year ago now, is just sort of spent a bunch of time playing with these models, working with a few folks to ingest all that content on my website, do a bunch of processes to break it down, to make it more recompilable by large language models and segmentation models and all kinds of different models, and then recombine it when people ask questions. And so, the talk is really just an overview of that process, both from a why did we make the interface design decisions that we made, but also why did we do this? How did we do it, I don’t go too deep technically, but I talk enough about the technical stuff for you to kind of peel back the covers and get a sense of how some of these technologies work and why. And then I also walk through a couple of experiments and changes and what we learned from it, what have you. Yeah, so hopefully, that’s a good description of the talk.

– Well, yeah, you know, what you’re describing there where you’re able to take your own work or a certain repository, and then build up an AI version of it using, you know, a lot of freely and relatively inexpensive tools. I think that’s really something that, to me, is extremely interesting and valuable in and of itself. Just the idea that it doesn’t necessarily require a huge organization to develop this. So, you know, I, the fact that you did that for yourself is interesting at one level, but it also means that it’s possible for people all over the world to do it.

– Well, I had help to be clear, right? It’s not all me, but yeah, certainly not a large organization. We’re talking like two to three people tops. And so, yeah, it’s doable and I think over time it’ll probably get even easier, right, as more and more people go through this process, and this is the same thing, right? When you first try to deploy a webpage in the nineties, like, what the heck is going on now? You can just open up your phone, right? And push a button and you have a webpage. Or you can even just talk and say, I want a webpage that’s this, and out it pops. So there’s definitely this curve, right, where things become more and more accessible, but what those end products end up looking like, is often shaped by what happens in the early stages, which is also why I really want to get involved in the early stages. ’cause that’s where a lot of decisions get made that later, everybody has to live with, right? And so, yeah.

– And it’s probably if I… I wouldn’t be surprised if the term AI ended up being the most common new word of 2023 or something like that. I mean, you’re certainly in a great position where you see startup companies getting involved in that technology ones that you work with as well as ones that just surround you. You know, I think one of the things people, you know, maybe outside of that are wondering is, how do you really know like what is probably gonna be relevant. You know, what is something that is, you know, maybe ending up being something that doesn’t come to fruition? How are you kind of positioning yourself in your strategy right now with respect to AI?

– Yeah, well, I think it’s like every other kind of wave of technology, right? I mean, you can see the waves off the shore, you know, they’re coming in, you know, they’re gonna break and hit the water, but if you wanna surf one of ’em, you either have to be there. I mean, you have to be there right in time. You can’t be too early and you can’t be too late. And that is non-trivial, right? So I think the big picture, will our interactions with computers be changed significantly by the advances in AI? Absolutely. Like, it’s already obvious and it’s gonna happen. How exactly that pans out, when it pans out, and who kinda defines what that looks like to be determined, right? If anybody knew that, then it wouldn’t be fun A. And B, like they basically have a crystal ball in the future and they could probably answer even more interesting questions than that. But I think, you know, that’s the part that’s kind of fun is trying to get that timing ride so that you can ride those waves. And that’s the magic of surfing is, actually, being on the wave, not having them smash over you, or what have you.

– Well, you’ve certainly, you know, always been in that mindset. Your book on “Mobile First,” I think was one that, you know, helped a lot of people to change their thoughts about converting desktop software into mobile apps and, you know, your approach just completely, you know, went from mobile as the start of that. So I imagine those same types of, I don’t know, new kind of thinking applies here as well.

– There’s similar patterns, right? I mean, if I told you how many times I heard people tell me back then, “Oh, nobody’s ever going to do blank on mobile,” right? And like, now the opposite is true, right? People are like, “Oh, nobody’s not gonna do that on mobile.” So things change. And at the time, it was same sort of thing, you could see that mobile was gonna become a really big and important thing, you know, in the like early mid-ish 2000s, but exactly how it was all gonna play out, who knows? But that, again, that’s the fun part, right? That’s the ride and that’s the thing that you get to enjoy and be part of if you get in there early enough as opposed to too late.

– Well, it, at the conference, we have a lot of very experienced practitioners, but we also have new people that are, you know, coming into the profession. I mean, if you think back on your own career when you were just leaving the University of Illinois after working with Mosaic, you probably had no idea you end up or have the career path you had here. There any thoughts that you might wanna share just about maybe your journey or how it developed that you think might help others that are just getting started?

– Yeah, I mean, if I look back at it now, I would say the fact that I was always interested in learning and kind of sharing what I learned and using that as this feedback loop was probably a big part of what moved me from one thing to the next thing. Even now, like just the call before this one, I was talking to a digital artist that I discovered on the internet and was doing super interesting things with 3D and AI, and I just reached out to him. We had a conversation, I learned a ton of stuff, right? So I think that simple process can really take you pretty far, right? If there’s something you’re curious about, take the step to learn about it and you’ll be surprised how many people are interested in teaching you and the amount of stuff that you can get from it that then takes you to the next thing.

– Well, thanks for taking the time to share your thoughts about your topic and other things, and I’ll look forward to seeing you in a couple of months in Seattle.

– Sounds good, thank you.

– All right, thanks a lot, bye-bye.