Hosts Eamon McErlean and Joe Devon interview Victor Tsaran, Google Design Platform Accessibility Lead about his early experiences growing up blind in Ukraine and his first interactions with computers in the US. He emphasizes the critical role of foundational accessibility in AI and design systems, advocating for user-driven interfaces and warning against a potential digital divide if accessibility is not prioritized. Victor also shares insights on effective public speaking and the importance of role-modeling to foster empathy for accessibility, highlighting the transformative power of technology in empowering individuals with disabilities.
OUTLINE:
00:00 Opening Teaser
00:43 Introduction
01:27 How Using A Screen Reader Is Like Looking Through A Straw
03:38 What Are Earcons?
04:31 Growing Up Blind In Ukraine & Moving To The US
08:53 Victor On Being A Musician
10:11 Why Did You Start Working With Computers?
11:57 Launching A Computer Center For The Blind
14:11 Victor's Advice On Public Speaking
16:12 Communicating The Importance of Accessibility To People Without A Disability
19:09 Role As Google Design Platform Accessibility Lead
21:56 Recommended Courses For Learning About AI
24:27 What Role Does AI Play In Alt Text For Images?
26:31 Future of AI's Role In Accessibility
28:25 What AI Tools Do You Use For Coding?
32:34 Concerns About AI and Accessibility
34:36 The Trolley Problem - Ethical Dilemmas In AI / Waymo
40:19 Victor's Book Suggestions
41:53 Things Sighted People Say To Blind People
45:31 Making Art and Music Accessible
47:27 Wrap Up
--
EPISODE LINKS:
Victor Tsaran: An Introduction to Screen Readers
https://youtu.be/izrC4R7SsH4
How Blind Users Experience Youtube | Victor Tsaran | Talks at Google
https://youtu.be/ewCSjiZ7Cdo
Earcon Demonstration
https://youtu.be/sv_swxXGc5s
TEDxSiliconValley - Victor Tsaran - 12/12/09
https://youtu.be/BsJB73c38yw
Braille Writing Devices: Slate and Stylus
https://youtu.be/50XCmWMiM-4
CompuServe
https://www.reddit.com/r/vintageads/comments/yaa0ik/compuserve_internet_access_in_1980_only_500_an/
Material Design
https://m3.material.io
https://m3.material.io/foundations/overview
Service Now - Horizon
https://horizon.servicenow.com
https://horizon.servicenow.com/guidelines/accessibility
Victor Tsaran's Music
https://open.spotify.com/artist/605ZF2JPei9KqgbXBqYA16
https://music.apple.com/us/artist/victor-tsaran/344507896
https://www.youtube.com/@VictorTsaran/releases
Vanderbilt University
https://www.vanderbilt.edu
https://engineering.vanderbilt.edu/bio/?pid=jules-white
Andrew Ng's AI For Everyone
https://www.coursera.org/learn/ai-for-everyone
DeepLearning.AI
https://www.deeplearning.ai
Gemini Canvas
https://gemini.google/overview/canvas/
CodeMirror
https://codemirror.net
Visual Studio Code
https://code.visualstudio.com
GNU Emacs
https://www.gnu.org/software/emacs/
Cursor
https://cursor.com
Windsurf
https://windsurf.com
Waymo
https://waymo.com
Moral Machine
https://www.moralmachine.net
Technofeudalism by Yanis Varoufakis
https://www.penguin.co.uk/books/451795/technofeudalism-by-varoufakis-yanis/9781529926095
https://www.wired.com/story/yanis-varoufakis-technofeudalism-interview/
Emma Donoghue - Author
https://www.emmadonoghue.com
Ep 9 - Matt Ater - Senior Vice President, Vispero
https://youtu.be/n6NGlE3qM3E
Ep 2 - Jennison Asuncion - Head of Accessibility at LinkedIn, Co-Founder of GAAD
https://youtu.be/X4QsrxskWHU
James Teh at NV Access
https://www.nvaccess.org/about-nv-access/
OSARA: Open Source Accessibility for the REAPER Application
https://osara.reaperaccessibility.com
Victor Tsaran's Links
https://www.linkedin.com/in/victortsaran
https://www.youtube.com/@VictorTsaran
https://x.com/vick08
- I like good alt-text. However, if I do want to understand the image better, for example, if somebody says picture of a music instrument, That may be enough until I realize, you know what I really would like to know what that music instrument looks like, you know? Then I'll just press a shortcut key and then I want to get a better description. 'Cause then they can have a conversation about the image. That's where the power really comes in. So I would maybe just to sort of summarize the conversation. I would say that alt-text is just the beginner of a conversation. And for anything more, this is where alt-text ends and I think an AI begins.
- Welcome to episode 11 of "Accessibility and Gen AI," the podcast where we talk to the people shaping the world of accessibility and artificial intelligence. I'm Joe Devin and joined by my co-host Eamon McErlean. Today we're speaking with my friend Victor Tsaran, the Google Design platform accessibility lead. He is a TED Talk speaker, accessibility pioneer. And I think I can now safely call you a fellow Vibe coder. Your work has reached far and wide, and Victor, it is my pleasure to welcome you to the pod.
- Thank you. I'm happy to be here as well and talking to you Joe after I don't know how much time, but no matter still happy talking to you.
- Well, why don't we start with this in your talk, I think it was your TED talk, or no, it was your talk, how blind users experience YouTube. You described, and I love this, you described using a screen reader as looking through a straw. Can you elaborate on that analogy for our audience?
- Yeah, that's a pretty old paradigm by now, but back in, I wanna say '90's, when people were trying to explain what it feels like to use a screen reader to someone who is cited, and I just preface it by saying that when cited people look at the screen, they generally get this privilege of seeing screen as a whole or perhaps parts of the screen. But there's obviously some sort of... They can see the layout of the screen, you know, in the much sort of broader picture. When somebody who is blind is using the screen reader and they're trying to understand the layout of the screen, you have to piece the layout one by one. And what that means is, in order for me to understand layout of the screen, I have to use a keyboard, first of all, because I can't use a mouse, then I need to use a tab key or some kind of arrow key or some sort of keyboard shortcut key to get to a certain place on the screen. So that means I can't just like understand the full layout of the screen. I have to get there by moving from one element to another. And so people sort of compare this, it's as if like you were trying to look at the screen through a straw, and the only way you could understand the full layout is by moving straw across the screen or up and down. And so that's where the whole concept and whole idea of a straw came from. And I just kind of reuse it because I think it very well illustrates what it may feel like for someone who is cited to try to understand the experience of a screen reader user.
- Yeah, I mean it might be old, but I really think that's an excellent way for someone that has not, you know, seen a screen reader in use to kind of picture it. And how about you-
- Maybe today we would modify it by saying like, looking through a paper straw? If you wanted to be-
- Oh, there you go.
- Conscious, right?
- And how about earcon? What is that?
- Yeah, so earcon is, again, it turns out it's a very old term as well. From 1980s I believe it was coined by Japanese sound engineer. And that's basically, it's a bit of a tongue in the cheek, but you know, we have icons, you know, for people who can see I-C-O-N and then someone thought, wait a second, if I want to give a cue or a representation of something on the screen in sort of visual form, but translate it into sound, why don't we just call it an earcon? And that's kind of where the term stuck. And it's pretty heavily used, I would say today especially, I wouldn't even say it's used by users today as well. So if you tell a blind user an earcon most people will probably understand what that means.
- Thank you again, thanks for making the time to join us today. We appreciate it. Love having you on. Before we dive into the deeper accessibility and gen AI technology discussion, can you share with our listeners a little bit of background just about, I believe you were born in Ukraine, and just how you grew up and how you kinda transitioned over here and into this role?
- Yeah, thank you. I try to be brief. So yes, I was born in Ukraine back at the time when the Soviet Union was just sort of collapsing or on the verge of collapse. And so I got to see, I would say a little bit of what it was like to live in the Soviet Union. Obviously, you know, my parents see so much more of it than I did. I grew up first in the eastern part of Ukraine than my parents moved to the west part of Ukraine. And in both cases I actually stayed at a boarding school for the blind. And so it was kind of interesting to compare the differences even though it was the same country. But, you know, boarding schools were run by the same kind of Soviet system. And one highlight feature of those boarding schools was the fact that you were kind of boxed into sort of society's understanding how you should be studying, what you should be studying, where you'd go after you're done with your studies. So in that respect, you know, things were no different from one boarding school to another. And then everybody knows what boarding schools are like. So then after I finished my school sort of high school, I entered a philosophy department in the State University in Aviv, which is in the west of Ukraine. And then while there, you know, and slightly before that even I was kind of always interested in music, but you know, while studying philosophy also, you know, tried to be a professional musician. You know, I was performing with the band on the stage and I won a festival in Ukraine in the genre of acoustic music. And while I was at the university, I was lucky to get a scholarship to come and study in the United States in Philadelphia. There was an international program run out of Overbrook School for the Blind, which the school still exists, but the international program does not. And basically the program was designed to bring students from different countries and kind of give us the necessary skills to perhaps go back and establish computer programs for blind students in our respective countries or just get educated and do something great with our lives. And this was the first time beginner of '90's when I came to Overbrook in Philadelphia. And many things happened to me there. First of all, I met my wife, who is now my wife. We met back then and we're still together and we traveled the world together and it's like separate story. There was one big thing in the second big thing that happened to me is that this is the first time I got to interact with computers. I've not even touched a computer keyboard before. I didn't even know things like speech synthesizers existed. And obviously this was the age when `the internet was just starting, you know, the internet era was just beginning , Compurserve, you know, things like that, AOL and all that stuff. So it was a complete change in my life. And since that experience of the school in Overbrook, I decided that computers is going to be pass in my life, that I'm gonna take both career wise but also in my personal life. And so after that, you know, I finished my computer studies degree, then my wife and I traveled in Southeast Asia. We were teaching computers and establishing computer centers in a few countries and let's see where do I go from there. And then eventually I, you know, decided that it was time to enter, you know, proper working class, so to speak. So I ended up applying for jobs first working in DC, then at Yahoo and PayPal and today at Google. Let's see, maybe I'll stop there. Maybe if you want you can ask.
- No, that was awesome.
- Sorry, that's just trying to sort of be as brief as I can.
- No, that was fantastic. Before we dive down the computer technology track, I do wanna ask your music skills. I believe you're even a published musician on Apple, is that correct? What the name of the album, was it Apple or Vanilla Bean Fields, is that correct?
- Yeah. Wow, that's cool. I should come some vanilla beans. That sounds very cool. Thank you.
- I love it.
- Yeah, music was always a big part of my life and you know, as I mentioned even before I entered university, I was already, you know, I won the festival in Ukraine and then I sort of played with a band and so music never really left me and I never left music and that led me also to participate in a few open source projects that, you know, whose goal was to make music software more accessible for blind users. But yeah, I do still play guitar and piano. I recorded, you know, a few albums, all of that stuff is on Spotify, so I still keep recording. although today I'm still struggling to record an album. I've been just doing more singles because I was like, you know, who cares whether it's an album? Today everything is in playlists anyway. But yes, I'm still pretty, I would say actively, I don't really play out as much as I would like maybe perform 'cause it's just not enough time. But I certainly keep playing at home and recording and try to contribute to the music accessibility as much as I can.
- Can I ask why computers? Was there anything specific?
- Yeah, so like I mentioned Overbrook, you know, when for the first time I interfaced with the computer, I just realized what an amazing vehicle it would be for a lot of my dreams, you know, including that I suddenly got access to a lot of the information. I could do a lot of the things independently, including recording music. I mean before if I wanted to use a professional recording, be it mixing board or you know, hardware or software. I probably most likely needed sited assistant. I mean, not to say they were not blind engineers before the age of computers, but it was certainly much, much harder. You really had to know what you were doing. With computers, I got access to music recording, I got access to music composition, you know, I got access to my job, you know, I can stay employed and so on and so forth. So I mean, this is kind of obvious, but you know, it wasn't obvious back then in the '90's when I just started, you know, before I even came to the United States. And once I did get to experience what computers can do for me, I was just, there was no question no going back.
- I can't help but notice when you mentioned CompuServe, they used to call it I think like the CompuServe information service, and they spelled the service with a dollar sign because it was so expensive. I remember my parents were paying a lot for that. And now with AI it's very similar because my monthly bill for AI coding is definitely a strong reminder of CompuServe. So you made that little connection for me. But again, there's one more thing I'd love to cover before we get deeper into the computer side and AI and all that. You created your own school for the blind, is that right?
- I mean, it wasn't school for the blind, it's pretty close. It was a computer center for the blind in Ukraine. This was actually... So after I came back from Overbrook in Philadelphia, as I mentioned before, I just got this idea, wouldn't it be really cool to teach other, you know, blind kids in Ukraine what I learned, you know, by then, now I know what I learned back then it wasn't that much, but it was way more than, you know, a lot of people in my country had access to. And so I decided, you know, it would be so cool to do something similar in Ukraine. And so my friend from the university and I started the Tree of Life Computer Center. We were lucky to get a funding from, and I have to mention USAID wink wink, not gonna go any further. But thanks to USAID we were able to start our program. And yeah, the program lasted for, after I left back to the US it lasted for another, I know, five or six years. Eventually they moved to producing audio books. It really had a very important beginning just for me because it kind of gave me an opportunity to experience what it feels like to help other people learn computers and become more independent. But also it helped other people in Ukraine as well, especially blind students. So yeah, that was an amazing experience. And I mentioned eventually later, my wife Kara and I would travel to Southeast Asia and we helped to establish similar computer centers in Cambodia and some other countries. So see this was a monumental experience for me. I can't, you know, possibly overstate it.
- Yeah, I mean that's incredible. And I love looking at builders of all kinds and you know, you have the mind of a builder that said, okay, I've learned something over here and I wanna, you know, pass that along to others. On a similar note, I'm part of the National Speakers Association Board here in Las Vegas, get to meet lots of public speakers and having run lots of tech community, engineers often struggle with public speaking. And here you are, you give TED Talks, you're the very first speaker for Global Accessibility Awareness Day. And I had a friend of mine who's also a really great public speaker and he could not believe how good you were. So if you'd like to share anything for engineers that might be nervous about public speaking or just speakers that want to know how to get into doing a TED Talk, I'd love to hear your experience and public speaking thoughts in general.
- Yeah, no, that's really cool. I mean, maybe I'm too modest, but I don't consider myself a good public speaker. I guess being on stage, you know, with music helped me to feel a bit, you know, comfortable talking to other people, but also I just try to be myself. You know, what I learned, at least one thing that I learned and the thing that tripped me up quite a few times is that if you don't know the material that you're gonna present on very well, there's gonna come a moment during the presentation where you eventually gonna screw up. But if you feel comfortable with what you're talking about, then the next big thing is just to relax and just be yourself and try to talk to other people about what you know, just as if you were talking to your friend. You know, there's obviously other nuances like, you know, making sure that especially there's obviously some specific nuances when you are blind, right? Because obviously I can see reactions of the audience. I always have to figure out a way to engage the audience by throwing a joke or two. And if I hear people laughing, okay, I guess at least they're still with me, you know, if I don't hear anybody laughing, people are sleeping or whatever, you know, so there are a few tricks that I kinda learned along the way as I was doing these things. But the first and most important for me was just to try to know the material that you wanna talk about, relax, and then just pretend you're in the living room talking to your friends. That helped me quite a lot.
- Good suggestions they are. One of the toughest things I've always found, either public speaking or informed, or even in meetings with larger groups, is how do you communicate the importance and the criticality of accessibility to those individuals that do not have a disability? How have you navigated that? Because, you know, sometimes that can be tough to do, to increase that awareness and empathy.
- I feel the role modeling does help a little bit. You know, I don't feel bad referring to my own experience when I talk about accessibility. And so that does gimme a little bit of an advantage in the sense that, you know, people can see what's possible. 'Cause I'm standing right in front of them, I'm telling them my story, and through that I'm hoping to build this empathy bridge that they can walk as slow or as fast as they want. But now they have someone to refer to, you know, they talk to their friends and say, you know what? I just saw this amazing presentation that was done by a blind Guy and he talked about accessibility. Why don't we think about making our product more accessible? It sounds like a much more powerful story as opposed to them saying, you know what? I read the Wikipedia article that said accessibility is important for everyone, you know? So I feel like the role modeling is a way for me to get people's attention. And then obviously again, just given the contrast, even during the TED Talk that you mentioned, Joe, is one thing that I've finished the TED talk and I said, look where I was before I got access to computers, and I think I was holding a brush or something like that.
- Slate.
- Yeah, slate, right? The braille slate.
- Oh, on the brush too.
- Yeah, and for the people who may not know the slate is basically how people used to write on a braille paper. You'd have a slate, which is like a metal grid, and you would put a piece of paper in between the two grids and you would use a stylus to push the dots through. And so this is how I used to study, you know, throughout my high school, I went through university typing on that thing. It was a lot of physical labor because, you know, you can imagine six hours punching the dots, you know, in paper it certainly made you feel tired. And then, you know, and in the right hand, I had an iPhone, and I said, look, thanks to this technology, I'm able to stand in front of you today and deliver this talk. And, you know, again, I think this leaves an impression. So I strongly believe that anybody who wants to bring accessibility into their life or make accessibility, you know, important for what they do in their work, figure out a way of getting a user with a disability to talk to your friends, to your colleagues, because this is probably one of the most powerful ways of introducing accessibility.
- Yep, speaking of work, would you mind sharing with us, I'm the user or the listeners, what you do today, what your role is?
- Yeah, so I work today, I work at Google. My current role is an accessibility lead at the Google design platform. Most people will know design platform by our library called Material Design. So Material Design is a design system, but it also comes with a set of components for different platforms. So if you want to build a website or you wanna build an app, you might reach out for Material Design because it already provides everything you need to have to build a product. And so my team's job is to ensure that this comes with accessibility built in. So our ideal situation is that, you know, you grab Material Design, you install plugin or framework, or you download the library whichever way you want to get to it. And our expectation is that you're gonna get accessible app or accessible website without too much work on your part. You may still be, you know, doing design and, you know, introducing your own little enhancements or whatever you're trying to do. But at the core, we hope that accessibility will already be there for you, so that you have to do as less work as possible. And, you know, I've done some other things at Google, I work on, you know, help out with other AI initiatives and things like that. But my current role, as I mentioned, is to lead the accessibility for Google design platform.
- It's so critical, Victor, like it really is, fortunately our own design team at ServiceNow, we recently released a new design system called Horizon. And one of the key attributes of that was ensuring that accessibility was built into the elements and built into the components right from the start as a foundation. And it sets the tone, it sets the tone the whole way through.
- No, absolutely agree. And I think we might get to this topic as well when talk about AI, because I do personally believe in like foundations, you know, it's, again like a house. I know we keep saying this sounds almost like a broken record, but it is so true. The better the foundations, the least space there is for people to make a mistake, not just for accessibility, but in general, right? And so this foundational aspects are so important, like, and I'm kind of happy to be contributing to Google's leading design system, Material Design, so that we help developers avoid making unnecessary mistakes.
- And that's awesome. I didn't realize after playing with Google, you know, with their Material Design a bit as a non-designer, it's good to know that I know somebody on that team that helped with the accessibility.
- Yeah, there you go.
- So now let's move to AI and talking about foundation, you've taken lots of different courses and have certifications, things like AI for everyone, which is something that no matter how deep or not you are in AI, that will help you a lot. That was a great course. What really was your favorite course and what would you recommend to others and what has sort of helped you go deeper into AI?
- I found the courses, I started very sort of, you know, very humbly by going to Coursera and see what I can find. A vendor, let's see, what's the university name? Geez, I should remember that, I forgot. Vanderville University, I think out of Tennessee? here is a professor, Jules White, and he does these really cool courses and I just like his style of teaching very sort of approachable. And so I like pretty much most of his courses, like on computer vision on ChatGPT, he was talking about there's a course, how to use Zapier. And then I moved on to deep learning stuff. And the team, that was a bit harder because their courses rely or you have to see the screen to understand what they're teaching. So then I was kind of employing AI a little bit to help me describe what's on the screen. That stuff was very interesting. But they went a bit more into like what to do with LLM, what are the good testing frameworks? What are some of the AI principles to keep in mind when designing with LLM and for LLM, you know? So a lot of the I would say computer engineering stuff that was very insightful to understand the background behind LLM because if you remember initially, it still is a very much a mystery for many people. But I think especially when you were trying to get into AI, people just throwing all this terminology around, it's like, what does this mean, what does that mean? PFT, overfeeding, you know, whatever, you know, people see. It was like, I have no idea what they're talking about, you know, inference, one person says one definition, the other person says the other definition, I was like, okay, I have to go to the source and figure out what these people are talking about. So that's why I was kind of decided just to take a few courses just to get myself acquainted familiar with, you know, with the field. And then obviously a lot of the other stuff, you know, it really comes from practice. But yeah, I would say definitely Vanderville University and deep learning AI, those guys are doing some cool stuff.
- Oh yeah, deeplearning.ai. And since you brought up using AI to help you with the images, I'm super curious about this, because a lot of practitioners who don't really need to use the alt-text have a ton of opinions about whether it matters or doesn't matter, but on the ground here you are, do you use AI to alt-text all images you run into? Or do you appreciate a good alt-text? What's your actual experience on the ground?
- Yeah, you know, this is one of the areas where we could probably talk for another two hours. I like good alt-text. However, I do have to say, there's so many AI tools today that I don't stress over it. If there's no alt-text, I can quickly press a shortcut key and get a description. The only reason I do like when people provide alt-text because when I'm reading, you know, some article and I just want, you know, maybe I just press, like read all something like that. I certainly don't wanna be, you know, unnecessarily reaching out for AI just to understand the meaning of every image because alt-text sometimes could be succinct enough for me to understand what's on it. And so yes, alt-text is still important. However, you know, I wouldn't bend over and say like, oh my god, you know, AI is so bad or so good about alt-text. You know, I think it's just another tool in the toolbox, right? If I do want to understand the image better, for example, if somebody says picture of a, I don't know, of a president or whatever, you know, picture of a music instrument. That may be enough until I realize, you know what I really would like to know what that music instrument looks like, you know? It's not enough for me to hear that it's a musical instrument. Then I'll just press a shortcut key and then I want to get a better description. 'Cause then I can have a conversation about the image. That's where the power really comes in. So I would maybe just to sort of summarize the conversation, I would say that alt-text is just the beginner of a conversation. And for anything more, this is where alt-text ends and I think in AI begins.
- And that's the beginning. Where do you foresee it going to, like, as it relates to accessibility in gen AI, the bigger picture? Do you believe it will be a true benefit for those individuals with different disabilities and if so, in what way?
- Yeah, no, I mean, I think inevitably, I mean I think this AI's here to stay, what I think AI is gonna enable us to do, and this is kind of what I'm really excited about, is to eventually let the user decide what sort of interface they wanna interact with. You know, so for example, if I go to, I don't know, cnn.com, I wanna be able to press a button, say, you know what, remove all the extra stuff from the page and I just want to see whatever in sections are interesting to me. Maybe I want to see those sections in a list view. Maybe I wanna see them in a tree view, maybe I want to see them with headings or whatever, right? My dream is that the users should be the one deciding how they process information, how it's laid out. Now I realize of course, it's a very challenging thing to say, especially that many designers pour their sweat into put, you know, putting together these websites or building these designs. And of course for them it might be difficult to let go that the user might be in charge, but I think this is the future I feel we're going towards where again, the drivers are the users and less so makers. Makers maybe enablers. So they could, for example, build a system, but be it a design system or let's say a, you know, website or whatever, that is flexible enough that does a shape shifting so to speak, you know? But it's the user to say, you know what, I just want this and I don't want that. And that will be perfectly fine. I think it'll be perfectly acceptable in the future I think.
- Now let's move on to vibe coding. What AI tools do you use on a daily basis? And especially when it comes to vibe coding? 'Cause we've spoken about this on DMs and stuff here and there, so I'm super curious what your tool set is.
- Yeah, so I use Canvas quite a lot and I do have to mention foundations again, because both ChatGPT and Gemini have a pretty accessible canvas. And that a lot of it has to do with the fact that the editor that I think everybody's using underneath, whether it's CodeMirror I think, has added some accessibility features. And now the result of this being is that whoever incorporates, you know, CodeMirror is gonna get accessibility even if it's not a 100% accessibility, but it's gonna be relatively usable. And I love seeing that, you know, when you see competitors, you know, using these different features, but they rely on some foundational framework or some foundational components, like, that's so cool. You know, I can recognize it, I know what the shortcut keys I can use. I know that it's, you know, gonna do this. Similar situation and I'm sorry, deviating a little bit, but like we have similar situations with Visual Studio Code, because it's used today across different AI products. And when you see that, you're like, oh, okay, so I can expect some level of accessibility. But yes, Canvas is my main one. I also use, this is gonna sound very archaic, sorry about that, but I do use for like more, not so much advanced stuff, but if I wanna work with like multiple code screens or multiple code bases, I use Emacs, because I find it still with Emacs Speak, Emacs Speak is a audio desktop as the creator calls it, T. V. Raman. Audio Desktop that allows you to use Emacs Editor. And the reason I like Emacs is because it still like gives you that feeling of like you're working with text, but it has a lot of the advanced features that all the graphical user interfaces have. So those, I would say my two. And then, you know, we see at Google we use some internal stuff, which, you know, it's irrelevant here, but yeah, those are the interfaces that I use primarily for vibe coding.
- Well, if you use Emacs then your OG, that's for sure. And you mentioned the VS code, you've got Cursor, Windsurf and probably a couple of others built on top of it, but unfortunately from what I heard some of them, the stuff that they added, the UI stuff they added tends not to be very friendly with screen readers, unfortunately.
- Yeah, and great point you bring up, you know, again, we talked about foundations, but then again, the foundations you still have to take good care of foundations, right? And so I'm hoping, you know, I think you've seen my tweets on the Cursor Editor. I do hope that, you know, eventually the team will find the time to make their edit because the foundation is there, you know, when you open it up with a screen reader, it talks to you and like it finds some toolbars, you can find some buttons. But yeah, the stuff that they've added obviously clearly needs a bit of work and so.
- I have been in touch with them about it actually, and they're very open to being shown what's going wrong. And I was gonna get one of my folks to just mock it up, but if you ever want an intro or whatever, if you wanna do a little video or something together, happy to bring it over to them.
- Sounds cool.
- And Victor, did you say your wife Carol?
- Yeah.
- Does she work in the accessibility arena as well?
- She does, yeah. And also we're both Googlers, so. We don't have to keep secrets from each other.
- Is she heavily into the AI world and vibe coding and like yourself?
- Maybe less so she works for place Google Play, so, but also on accessibility. So I share lots of things. We have lots of things in common, talk a lot about the same things, but you know, who doesn't work with AI today, right?
- Do you have any concerns, like bigger picture concerns with accessibility in AI?
- Yeah, I do have a concern. And that is the... We sort of touched on it a little bit with Cursor. I do see proliferation of AI platforms that do not take in accessibility into consideration. And I am kind of worried about it for several reasons. Under the heading of digital divide, the AI platforms and AI tools are going to become... They are becoming already, you know, tools for people to get jobs. These are the tools for play and work and fun and, you know, whatever, for people to start small businesses. And if we don't take good care in terms of accessibility of these tools, I'm just foreseeing the issue where some people may not be even able to get jobs or might not be able to do AI for everyday needs because they can't get proper access within these tools or to these tools. And that is one of my biggest concerns. There's such a proliferation of these tools and things just changing every day. And I'm just a little bit concerned accessibility might be, again, one of those games that we'll have to do quite a lot to catch up. And so I'm doing everything I can to make people aware that, you know, you're not just building a beautiful or fun AI tool, but you are enabling people to participate in the future of our society. And so it's critically important that you think of accessibility.
- Completely, completely agree. And that's one of the things that Joe and I and many guests that we've had on this call reiterate that same thing, that, you know, accessibility, the reality is accessibility wasn't afterthought for many years for many companies, but with the speed that AI is moving up, we cannot afford to make the same mistake again. We really can't.
- So Victor, you have studied philosophy. And one of my favorite problems, or let's say ethical dilemmas, I even spoke about this at... I gave a lecture at USC about self-driving cars and writing algorithms where you have to make some difficult choices like the trolley problem. Perhaps, can you just explain what the trolley problem is for those who may not know? And do you think that there are any algorithms like Waymo or any other self-driving car that you think they're actually writing in these kind of choices?
- Okay, so Joe, I have to ask you actually, could you explain to me what the trolley problem is? Because this is the first time I hear about it. That would be a fun element, maybe like in a sentence or two.
- Yeah, so the trolley problem essentially is, it's just a hypothetical situation where you are a conductor, a train conductor, and a train is coming. And if it continues on its path, it's going to kill five people who are tied to the tracks. And if you switch the, what do you call it, the little handle there, then you'll veer it off to another path. And there's one person tied to that track. And so if you do nothing, like, first of all, it's sort of a dilemma because either you kill five... Either five people will die or one person will die. But if you take action, are you actually killing the one person, right? Think about that person's wife or child. And there's tons of different variations on the trolley problem. And anyone, I highly recommend you watch the show called "The Good Place," which is a comedy, and they have a nice real visual of this trolley problem. So when you're doing a self-driving car, and you're in a situation where you detect children in front of you and you have to stop suddenly, let's say it's raining and you need to stop. And if you go straight, there's a bunch of children. If you go to your left there's a bunch of old ladies. And then if you go to your right, there's, you know, another set of people that you know, or dogs or you know, whatever it be. But basically, does anybody have to write in an algorithm for these self-driving cars to figure out.
- Yeah. And as a philosophy person, like how do you analyze it?
- Yeah, no. So that's a really cool, very cool. Thank you for setting this up. So my experience with Waymo so far has been, we haven't gotten to trolley problem, thank God. But it is pretty good at like, it'll actually give you a message yielding to pedestrian. I think it's trying to be a very good citizen. I actually think if we have more Waymo on the street, it doesn't be Waymo could be any other, you know, car. Provided, of course they're Waymo standard. We gonna have better safer streets than we have with humans driving. I mean, at least from my experience, if we are gonna have such, you know, cars that are great citizens, they're kind, they're, you know, they're smooth riding, they're whatever. We might actually end up with a much better experience than with some of these emotional human beings, you know, who had too much coffee or whatnot. So I think we might not even ever get to the trolley problem in the first place, you know? So that's kind of my hope. So, yeah. But so far what I observed with Waymo has been really, really quite an amazing experience. And it's taken years, you know, 'cause I remember when my wife and I were tried Waymo like, I dunno, it must have been seven or eight years ago, and it was still like very buggy and it was still jittery and it's just, you know, wasn't as great of an experience compared to what it is today. I mean, it's honestly, it was hard to believe. It's probably one of my biggest... If somebody ask him what's my most impressive innovation of 2025 most innovation experiences, that probably would be the one. I mean, Wayne has done an amazing job, even in terms of accessibility as well.
- Yeah.
- So I don't know if I answered your question about the trolley problem, but I hope-
- Well I just wanted to hear your thoughts.
- Very, very good picture. It was, it was good.
- I just wanna add, there's a website that some folks put together where they ask the questions to the audience, in order to inform the people writing the algorithms for these self-driving cars. I'll find the URL and we'll put it in the show notes.
- Yeah, and you know, one other, just last point on this, I wanna say. When people start bringing like, what will machines do or will not do, I always say, okay, let's go back to the baseline, okay? How is this any better than what humans do today, right? Like, you know, if human was in the same situation, human would have to take a decision, right? Like, you know, there is no way that human could have gotten out of that situation. So I think, you know, this whole idea of like neural, you know, us trying to replicate human brain, it's an interesting thought even from that point of view because I think a lot of the thinking that goes into some of the robotic technology or some of the gadgetry is precisely can we build a better human, right? It's not like they're trying to, I don't know, invent something completely new, but it's like, can we build a less irrationally impulsive human, you know, that can do better than humans can, right?
- Yeah.
- So you seem to be keeping yourself busy, right? Between the public speaking that you do, the certification and training that you do, your love for technology, playing music. And I've also seen that you're an avid reader. Any book suggestions, either business technology related or personally that you'd recommend?
- You know what? Okay, I'll start with maybe technology book, which is, well it's not exactly a technology, but it is technology related. It's called "Techno feudalism." It's written by a guy, I can't believe his last name, Janis and last name escapes me right now. He used to be a minister of finance of Greece. It's a very interesting thing-
- Is that a good thing or a bad thing Victor? Being the minister of finance for Greece?
- Well, I think it's a good thing. I mean, if he got to minister of finance, that's, you know, you gotta be pretty smart, right? To get there. . But you know, he's a very smart guy. He's got a lot of interesting views on things like cloud capitalism and what is gonna be the future of post-capitalism. You know, I don't have to agree with every point he made, but his way of thinking is very, very interesting. Definitely challenges a lot of today's, you know, ideas of where we're going, what technology may mean for us. I thought it was a very, you know, thought-provoking read. You know, other than that, to be honest with you, Eamon, you know, we primarily read a lot of the literature just like fiction, you know, so like the, you know, Emma Donahue, who's a great writer from, I think she's originally from Ireland, but now lives in Canada. Honestly I could probably, we read pretty much one book every week or every two weeks depending on the length. So there's so much I could talk about literature, but, you know.
- Yep. She's awesome. She is.
- We had Matt Ater on recently who's at Vispero. And we asked this question that really, it came from a social media trend where people were saying dumb stuff that cited people say. What are some dumb things that you've heard had sighted people come tell you or do to you or around you?
- Yeah, you know, I probably have many of these, but like one thing that just immediately comes to mind is I remember when our work that I mentioned in Vietnam, Southeast Asia, in Vietnam, I think one time I, you know, was coming up the stairs in the hotel and the person who was trying to help, they didn't quite know how to help a blind person first time probably in their life. Plus they didn't speak good English. So he just came up to me, he just lifted my foot and put it on the next step. And I thought it was-
- Oh my God.
- It was kind of cutesy. I loved it because, you know, well, you know, if you think from that person's view, this is the best they could show me that they really care. And I look at these things positively really to be honest with you. Like, I rarely treat these kinds of attitudes as dumb things. You know, I think actually to me it's always an expression of empathy. It's just that different people express it differently, you know? I mean, mostly what makes me laugh are the expressions that people use. Like, for example, you know, sort of saying blind will say like, sightless or even heard experiences, like people without eyes or, you know, things like that. It sort of makes me chuckle. You know, I grew up and since the very... I would say childhood, I just was never really offended by these things. It's more like, okay, you care, you know? We can work with you, you know what I mean? My attitude has always been more on the positive side. You know, I'm willing to accept any strange definitions or what I'm not willing to accept is when people make up their mind or they say something or they call you something and then they judge you by what they think, you know, you are. And that is where my tolerance kind of stops. But until then, until you're willing to work with me, and if you are willing to listen and if you are willing to, you know, learn that there is another way to think about somebody who is completely blind. I'm a happy, you know, colleague, happy partner, and I'm happy to work with you on that.
- I once was with Jenison, my co-founder, who of course you know, and we were at a hotel, and the person that was checking him in said, you know what, I'm gonna upgrade you to this other room. And by the way, it has a beautiful view of the lake. And he said, oh, that'll come in so handy or something like that. And she was so embarrassed, but it would a good laugh for all of this.
- Yeah, you know, but again, right. I mean people, she did it without even thinking twice that she's saying it to a blind person. But the gist of it, like the source of it was good, right? Meaning the person yeah, exactly. So she wanted to do something good, she just didn't think second step, you know? For that maybe you need to play chess at least.
- As we start to wrap up, 'cause we've taken enough of your time today, but is there anything we didn't cover that you wanted to share in regards to either projects you've worked on or lessons learned throughout your career and journey or possibly things that going forward.
- I mean, in dangerous slope, but maybe one thing I would say that really gives me a lot of pleasure is to work on making art accessible. And in my case it's music. So I've worked with Jamie Teh, who some people might know, one of the co-founders of NV Access and just an amazing all around person talent, very talented engineer and just my friend as well. And we used to work on this set of scripts for Jaws called J Sonar, which enabled us to make Cakewalk Sonar, which was a very, very popular recording software back in the day for Windows who made it fully accessible. You know, as a musician you could use, completely bypassed a graphical interface and you could just record just like any other engineer. And then I've contributed to OSARA, and also Jamie actually runs the OSARA project. Which enables you to use REAPER Digital audio workstation. So it always brings me pleasure to work on these sort of initiatives when, you know, a utility or software that you help build or contribute to, helps some musicians around the world to record, compose, and just express their creativity using the mainstream software. And there's coffee of course. I love coffee.
- Coffee's always good, always good.
- Yeah.
- So it's so funny you mentioned Jamie, literally, I was talking to a few people that he would be a great future, maybe the next guest for us. And I was just trying to find somebody that knew him or his co-founder. And you're nominated now, so.
- Cool, I would be more than happy to introduce you guys. Yeah, he's a great musician too. You know, lots of good things I could say about Jamie, so I think he'll be a great addition to your podcast as well.
- Well, thank you so much, Victor, this was a blast. Really enjoyed the conversation and yeah, I'll catch you on social media vibe coding.
- Yeah.
- And hopefully see you soon.
- Cool. Thank you so, so much to both of you. It was my pleasure. Great chat. And we'll catch each other somewhere. I mean, it's the world where it's difficult to avoid each other unless you really want it. But even then, it's not easy .
- Love it, thoroughly enjoyed it. Thank you Victor so much. Take care.
- Thank you. Bye-Bye.