Democratizing AI Education: A Conversation with Timothy from the AI Launch Lab

Automatic transcript of the recording (English); speaker labels are automatic. Timestamps open YouTube at that moment.

0:00Sacha: Today, I have a very special guest with me, Timothy, the Co-Founder and Managing Director of the AI Launch Lab. He's also a member of the AI Advisory Board at Dawson College with a passion for combining AI, new business models and economic paradigms. Timothy has extensive experience in the AI field. I'm really, really excited to have you on today, Timothy. Welcome.

0:21Guest: Hey, Sach. Great to be here. Thanks for the intro.

0:23Sacha: So, if we can get started with the first question, for people who don't know the AI Launch Lab yet, could you explain what it is?

0:32Guest: Yeah, it's a non-profit organization that started in 2019 and our mission is to basically enable students, but really people in general across Canada to get into the field of AI. And our main focus right now is on AI training and internship programs for college students, although we are going to open up beyond that soon.

0:56Sacha: That's awesome. And what inspired you to build this company and why do you believe it plays a very crucial role in the development of the AI here in Canada?

1:07Guest: We got started out of a student association at Concordia University in Montreal. So, in 2018, I went back to school or I guess around that time I went back to school. My background is in finance, actually. And for a couple of years prior to that, I was thinking, man, I don't know if the field, if the jobs are really what I thought they were going to be when I went into it. And so, I went back thinking, maybe I'll do like a master's degree and try to, you know, bring together finance and AI in some way. I had no idea really how to do that. And I got involved with the student association, Artificial Intelligence Concordia. And that was really exciting and, you know, learning about everything that was going on at the time. AI was still a hot thing then. And, you know, there was just so much that was happening. So, we heard from a lot of students at Concordia that they wanted to get into the field. And there weren't a lot of options for most people. So, I mean, you could do a online course or you could maybe be lucky enough that, you know, you can get into a master's degree or PhD to go that traditional route.

2:14The online course really wasn't enough for people to really get into the field or to learn the skills that were needed sufficiently. You know, you have to do a lot beyond that. And a lot of people need some support through that process. And then, you know, the numbers of seats for doing a master's degree, even today, it's just tiny if you compare that to the amount of demand, both on the student side as well as on the employer side. So, you know, we didn't have any major aspirations of trying to, at the time, you know, change that dynamic, but we wanted to just help in whatever way we could. So, we based our program a little bit on the District 3 residency program. District 3 is Concordia's incubator and accelerator for startups. And they had a residency program where they would have students from different backgrounds that would be grouped together and work on projects for startups for a summer. And so, we thought, let's try to do something like this, but focused on AI. So, that's how we started. Initially, we were focused on trying to get projects with District 3 startups. We were fortunate to get, you know, funding from the Concordia Student Union, Concordia Council on Student Life, and also from some of our startup mandators that did projects with us.

3:26And then also from District 3, we were really fortunate to have them as a major partner initially. So, that was really cool. Then the pandemic hit, you know, and then a lot of budgets got pulled back. So, we thought, like, how are we going to keep doing this? And then we saw that there was actually grants from the Quebec government to do these types of training programs. So, we set up a nonprofit organization that we could use to then, you know, apply for those. And luckily, in 2021, we got some. So, you know, we thought, let's focus on, you know, really the education side. We'll have, of course, we want to provide value to the companies as well as the students. But having kind of one funding source kind of made it a little simpler. And so, we did that for a couple of years. Then we partnered, we were very lucky to partner with Austin College. So, in, I'm not sure exactly what year they started, but around that period, they invested millions of dollars to transform their entire curriculum, basically every different program, every major, not major, every faculty to incorporate AI to some extent into the course content.

4:34And so, that was really exciting. It kind of made a big splash in Montreal at the time. They started their advisory board. I was very lucky to be selected to join that. And so, through these kind of collaborations, we started working with them on a college-focused program. And so, in 2021, we were still focused on undergrads as well. But we, in 2022, made the switch really just to focus on college students. And so, this is a change kind of from when we started. We thought we would just be serving, like, the undergrad, I guess, demographic initially at Concordia, because those are the ones who, like, reached out to us. And, but we were really lucky that actually, Joshua Bengio shared our, one of our social media posts. So, that got a lot of applications from all over the place, not only to join our program, but also to join the team. So, you know, our founding, our co-founders basically came from University of Montreal, came from Mila, and other places.

5:37And so, we were super, you know, thankful and lucky to have that catalyst. But then, also, we had people applying who were graduate students, people who were not students. We had an executive from Hydro-Quebec in our program. So, it was, like, amazing to have all of these people together working on projects. You know, we, everything from, like, CGEF students to have an executive at that level and, you know, postdocs, et cetera. So, that was really great. But the problem was, we didn't really know what we were doing with regards to our curriculum. And so, people did feel like, especially if they were at the beginner side, they felt like it was difficult for them to keep up. And so, people felt like they were being left behind. So, that also gave us the rationale that, look, if we focus on college students, we'll really be forced to learn how to train students who have a minimal background in this. So, we did that for, you know, for, since, I guess it was 2021, 2022, focused on college students.

6:39And then, this year, we were really lucky to have a new partnership with MyTax. So, MyTax is, especially in the startup and tech ecosystem, I'm sure a lot of people know them, they fund a large portion of the R&D internships across Canada. So, with them and Dawson College, we have this new initiative called AI Accelerator Programs. And that is designed mainly for college students, or I guess it's only for college students, although, of course, sometimes, you know, companies want to hire interns that are not just college students. So, that's the reason for maybe allowing some who are in university as well to join. But really, the focus is college students because the timing is right for that demographic to get into this field. The tools have become more accessible and more powerful. Just an example, last year, and so we do these events, obviously, with Dawson every year, hackathons. And last year, we did a hackathon with them. And, you know, of course, students are very motivated, and many of the ones who joined for those types of events already have been working on this stuff, you know, in their programs, but also on their own.

7:48So, it's a very specific, you know, kind of motivated group of people. But last year, I'd say it took maybe a day and a half, you know, of course, some exceptions, but a day and a half for students to really get their data together and to get it to be usable, to kind of process it and clean it. This year, by using LLMs to help do that work, it was maybe a few hours to get to that level. So, it's a massive jump. Of course, all the problems with LLMs in terms of, you know, them, you know, hallucinating or, you know, making mistakes, even with that and students having to troubleshoot and figure out, like, oh, why is this not working sometimes or to ask for help? It's a massive jump that we've seen. So, we think that not only, you know, of course, the hype is there and the demand is there, but now these students at that level can actually deliver value for companies. And so, that's why we think this is going to be a big deal, not just for the students who join and the few companies that are, like, already interested in hiring interns, but in the long term, something that will be, like, kind of Canada-wide and make a big change in terms of how companies hire and how they think about doing AI projects.

9:02Yeah,

9:02Unknown: that's a

9:02Sacha: big deal. And so, the AI Launch Lab, you guys work on diverse projects, right? It's from, I've read XR to Energy, and you were mentioning that students work on real-life projects, right? I'm curious, how do you guys go about selecting these projects?

9:20Guest: So, our process is always changing, you know, depending on what partners we're working with. Now, with Mitacs, it becomes a little bit different again because they obviously have a huge amount of them, a huge network, and companies are already used to working with them for internships. So, we're very lucky to have their support in getting new projects. But excluding that, I think in general, it's just we want to have companies that see this not just as, like, get a cheap student to, you know, to do, you know, this, like, work that we don't value. We want them to really perceive this as not as, yes, we need somebody to do this work, but also we want to invest in the community, invest in the ecosystem. They see it as something that they want to provide value to beyond just what they get out of it. So, I mean, that's a little bit vague. It's usually just kind of, you know, meeting with potential partners, seeing how they, you know, what is their thought process, what are they looking to achieve, how do they think about their interns and their employees, and just see if it's a good fit, you know, because even if intentions are good all around, it doesn't necessarily mean that it's going to be the perfect fit, and we want to provide the most value and make sure that, you know, our program is really well suited for each project.

10:38So, you know, we'll kind of have a discussion and talk with our technical team and see, like, is this feasible for the students to achieve in that period of time? Is it a good fit in terms of the curriculum that we have, et cetera? But in general, you know, what really makes, I think, our programs exciting, or one of the main things that makes it exciting is it's such a wide variation of projects, and they're always, like, really cutting edge. I mean, you think who's going to hire a college student to do, you know, state-of-the-art stuff, but every AI project is state-of-the-art, right? So when you have an intern working on it, they have to figure out, like, oh, we don't know, has this been done before? How can we figure this out? Is there some new tool that can solve this problem? So that brings a lot of value both to our program but also to our students like, wow, I'm going to get to work on a computer vision, an agriculture project, or a self-driving truck project, or a crypto and AI project. And, you know, that's a big thing that kind of excites them and makes them want to join the program.

11:42It's not just about here's a boot camp, I'm going to learn from a textbook, or I'm going to learn, you know, I'm going to go to store shops. It's really about the project that makes most people excited and motivated to learn and to achieve the results that they do.

11:56Sacha: Yeah, and I've read something else that also probably excites a lot of people is that the program is really focusing on experiential learning, sorry, and mentorship, right? So I'm curious to get your take on why this approach is important also in developing AI talents.

12:16Guest: So for us, it was just we thought it made sense initially because, you know, learn by working on projects, it serves two, you know, buckets or two problems, you know, you both get to learn, but you're also getting work experience that's actually relevant. You know, unbeknownst to us at the time, you know, this is a big trend in education that's already been ongoing. You know, every major educational institution is investing big in experiential learning and in these types of internships slash applied AI projects with companies. And I don't want to act like I'm an expert on it. I'm definitely not. Definitely, we have those people that we work with at Dawson that understand exactly why this is so important. I would just say that it's always been something that students have seen as valuable. That's not to say it's not without controversy, because, of course, there's the potential issue of, especially during our training programs, before we had the internships, the paid internships that are tied to it, the idea of like, oh, I'm learning, you know, and I'm not getting paid to do this work.

13:20And, you know, is there a sort of exploitation coming out of that? So for us, you know, we always had to make sure, like, look, the value has to be there for the student first and foremost. They make the decision, you know, do they want to keep coming? For them to keep doing this. So that's always our main North Star. Like, the value has to be there. And then some for the student to feel like this is, you know, valuable enough. And so that controversy has kind of been, has to some extent held back a little bit of, like, universities and colleges from doing this in a big way. But for us, you know, we think about, like, you know, if we're serving that, you know, we're not having that problem of, like, the student feels they're putting in more than they're getting out, as long as that's fine. You can't pay students to learn. Students pay for education. Obviously, for us, we want it to be accessible. So that's why we have it free. I don't know if that'll ever change. But if, you know, we definitely want to make sure that the amount of money a student has is not a barrier for them to learn.

14:23That's really important for us and our team to make this accessible to people who might not otherwise have these opportunities. But really, I guess, getting back to the point of why is working on applied projects so important, you know, you're not learning stuff that's already out of date. And especially in AI, things are changing so quickly, right? It's like on a monthly basis, sometimes a weekly basis, there's a new model. There's some new tool that came out that just changes the game for some tiny portion of the ecosystem or of the field. And in some cases, it, you know, changes everything, you know, if it's a new LL model from meta or from OpenAI. So that's also why, even though we have our curriculums and we, you know, our teams kind of develop a blueprint for students in terms of how are they going to go about executing on a project and what workshops are we going to provide them at what time, it has to be adaptive and dynamic because, you know, not just in terms of what tools are coming out and what models are coming out, but also students can, in week two, see like, hey, here's another way we could do this.

15:27And then that changes the entire path of the rest of the time that they're going to be working together. So our team has to kind of pivot and change up what they're going to provide for supporting students along the way. So there's that. And then, you know, it's real world stuff. So, you know, of course, it's been a cliche, I think, for as long as I've known is that, or been a student or whatever. Or, you know, you don't really learn the stuff that you need to know when you get your job. You know, you learn all this stuff and it shows, okay, this person's disciplined. They can do work on a timeline and, you know, whatever it is that that says about somebody. And then you get into your job and it's like you have to learn everything almost from scratch again because the company has its own processes, its own, you know, its own frameworks, et cetera. So getting as close as possible to what they'll actually be doing, of course, is valuable. And so that's one of the big reasons why we think it's important. I'm sure I'm missing stuff, but that's part of it.

16:22Sacha: Yeah. And I think the value is being perceived because you've been successful in recruiting and engaging over a thousand participants. I think I read on one of your blog or one of your posts with a lot of members coming from underrepresented groups. So I'm curious to get your take on how you guys have been successful in recruiting so many people or students and why is diversity so important to you?

16:50Guest: So when we started at Concordia at the Student Association doing these programs, initially we were faced with the problem or I don't know if it's a problem, but we were faced with the issue that, okay, you know, is this going to be an 85% men program? Which is just reflecting the field, unfortunately, you know, today, well, the last numbers that I've seen statistically in terms of, you know, who's in this field in Canada, maybe a couple of years old, the data is it's 85% just as one metric. But I mean, you can kind of spread that across every different diversity, you know, group that you're talking about. And the thing is, that's gone down since 1990. So we actually had higher levels of diversity within the field in 1990, despite the efforts that happened in the last several years before a little bit of pushback. And the other thing is, it's also an issue of accessibility. So, you know, if you happen to have, if you're on the path already, you know, if you're in university and you're in, you know, computer science and you're on track and you're going to do a master's degree, I mean, you're going to be fine, you're going to get into the field.

17:58I mean, for the most part, those diversity rates are not very good in traditional educational institutions, especially in terms of graduates. But, you know, we really wanted to have kind of a different path and an additional path, right? So for people who had trouble getting into the field, especially because, you know, we really believe that at some point, whether it's two years, five years, 10 years from now, the our AI competent in some way is going to be much, much bigger. You know, I remember seeing a data point or it's from, it was from IBM, I think in 20, 2018 or 2019, I forget what year. And they said that the demand for AI developers in Washington, DC alone was something like a million, it was something crazy, which doesn't even make sense, right? And like, I don't know how, what their methodology was. And then of course, demand is affected by supply. So if you can't possibly find it, then is it real demand? Anyway, but like, that was just an example of like, there is demand to do an AI project, but most companies in this, when we started in 2018, it was just PhDs.

19:03Like companies were like, we were looking for PhDs. They didn't even consider hiring anyone who was not a PhD or above. And I think it's still largely the case. It's changed a little, you know, from AI scientists. Now it's, you know, there's AI developers and stuff, but not as quickly as I thought it would in that period of time. So a lot of what we're trying to do as well is to show that actually that's not necessary for every AI project. Of course, for some, you need somebody with that kind of knowledge and background. But also sometimes researchers in academia and scientists in academia are not necessarily the best for doing AI projects. You need somebody who has a different mindset than doing kind of that sort of research in academia, which is, has a different timeline, has a different workflow and, you know, in terms of the expectations. So I guess I'm going dancing around the question about diversity, but I would just say like, we're very proud and happy that 75% of our cohort has always been from these underrepresented groups in tech. And we hope that, you know, and we're working always to continue that and to try to do better.

20:07But it's, you know, we think it's important for the field to have people from different backgrounds and different domains as well. And, you know, it's just when you have different people with different perspectives in the room, it can make a difference in terms of how you build and what things you think about when you're building product. So, and there's a backlash on this. And I didn't know, you know, initially when I thought like, oh, diversity, of course, everyone agrees with diversity. And then I learned actually in tech and in startups, there's actually historically been this counter position to that, which is that actually you want everyone to have to think similarly so that you're all rowing in the same direction. And that's a funny metaphor, I guess, to think about, because it's like, oh, yeah, you're rowing and you're going fast. And the idea is you all go in the same direction. And, oh, we're in the wrong direction. Okay, everyone change. And then everyone rows in the other direction, right? You know, I don't know. I guess that can work in some cases. Hey, there's like a, there's like a rock. Like, let's, let's turn back. I see the rock. Right. So, I don't know. I, I, I see that there's both sides in some cases, but I think in general, especially for society, it's important to have people with different backgrounds, different perspectives in the field.

21:14And so that's something that we're, we're really passionate about.

21:18Sacha: And you know what, to, to go back to your earlier point, a few weeks ago, I was talking to one of the co-founders of Move.ai and he was saying that for them, it's all about young people. And they were talking about this and they were hiring a lot of young people. And it's very, very interesting because this technology or this industry of AI is really mostly about young people. At least that's, that's what they were saying, right? Like a lot of the leaders in the field are young people. And I'm curious to get your take on it. Why do you think so many companies still to this day are reluctant to hiring or trusting young people that, you know, come out of either bachelors or even programs like DI Launch Lab? Why are they so reluctant to, to trusting these young people? Because to, to Move.ai's point, like it is really maybe one of the use cases where young people are actually the leaders in this field, right?

22:12Guest: Yeah. So, I mean, that, that's, it's an unfortunate thing that people who are above a certain age find it hard to get into this field. I mean, we, we have been lucky to have people in our programs that have been over 40, over 50, over 60 in some cases. And of course it's a self-selected group. So that's not representative of, of the overall demographic, but certainly the people who have been in our, in our programs who've been, you know, older have been as enthusiastic, as curious, as, you know, willing to adapt and, and try new workflows and all that. But of course it's natural. And I don't want to act like I'm an expert or whatever. I'm just kind of what I've, what I've seen in the years that I've been doing what I'm doing is that, of course it makes sense that it's harder once you get entrenched to a workflow and a way of doing things. You're less likely to want to change unless you're forced to. So, you know, that's going to be part of it. I mean, the other thing of young people is the idea of you don't have a family and you just kind of work, you can work 80 hours and it, you know, it's not, you know, the work-life balance is not necessarily the same.

23:18It's kind of a different mindset. But again, some people who are older, many people who are, you know, work long hours and all that. But I think in general, you know, I was, I was reading this on social media. There was a survey from Goldman Sachs about investment banker interns. It was from a few years ago, I think. But for some reason it was, it was being shared again. And they interviewed a small number of interns about their job satisfaction and kind of, and it was, it was like, you can't even believe, like, okay, I knew people who went into investment banking when I was in school and I knew how brutal it seemed. But like, this was next level. Like, this was like straight on, like, what, you know, works like, what is your satisfaction with work? It was like one out of 10. I mean, people were like straight up, like my, you know, my body is hurting. Like, it was brutal. It was really brutal. So there is, and I'm not saying this is what, you know, companies are expecting necessarily in AI or whatever, but I think there's a little bit to that, that it's like, it's easier to expect this type of grueling work ethic from young people that you don't necessarily value because they're replaceable and there's a steady stream of ones that you can just, you know, bring in.

24:31You know, I think certainly that's probably going to remain. I don't see how that's going to change. You know, part of it is psychology, right? In terms of what, when you're older, you just, you know, you know what you want and you know what you don't want and you're used to certain things. And, but, you know, 2030, I think everyone, almost everyone is going to need to be working in this, you know, using tools in this field, at least they're going to need to have, you know, similar to how you needed to learn Excel spreadsheets in the 80s and 90s. It'll be like that. So, you know, again, that's not what companies like Move AI are hiring for at the moment. But I think it'll, you know, there'll always be demand for the top level and the young people who just came out of school and that sort of thing. And that's a, that's a certain profile. But I think our hope is that, that many companies will be willing to take chances on people who don't fit those criteria because it is the case that there are people who don't fit that, that mentality.

25:35And, and, and increasingly people are being forced to change. So they will, they will change because there's no choice. And, and so, you know, companies should give them a shot.

25:48Sacha: Earlier in this conversation, you mentioned that Dawson College allocated a significant budget, right, to becoming the center of excellence in AI in, in, I don't know if it's Montreal or Canada, but could you tell us how your company, so the AI Launch Lab is helping them achieving this goal?

26:09Guest: Sure. And, and actually, I just wanted to add one more thing to the last question, which is that, what did I forget? No, I was going to say, okay, let me come back to that. So Dawson, we, we were lucky to do some projects with them, some pilot projects with them that got funded by the Quebec government, by the Ministry of Economy and Innovation a few years ago. And those were really about, you know, experimentation and, and trying, you know, new methodologies for training students, you know, in regards to AI. So, you know, we were lucky to be, you know, partners with them at, you know, at that time. And, you know, they've, they've, they've invested, you know, as you stated heavily into, you know, training their professors, how to incorporate this and, and, and changing their curriculums. I don't want to speak too much about it, because I'm certainly not, you know, I can't, you know, represent them, and I can't do it justice in terms of the amount of work and effort that they put in, and in terms of what impact that's had on, on, on their students.

27:16But, I guess part of it is that it is, it is beneficial to be able to have an external program that you can send your students to, that they can get some sort of extracurricular credit and in the future, maybe some, some course credits for, rather than trying to force your professors to do things, because that doesn't really work. So, so part of it is just, you know, it's, it's difficult to get a big ship to turn. Not to say that they haven't turned, but that is always a challenge. And then the other thing is, you know, you have to get buy-in from everyone. And, you know, our, our team has bought into it, of course, because we joined, because we believe in what we're doing, right? But not every professor, even at the institutions that really are passionate about this, believes in it. So, you know, anyway, this is just one thing that, that I think is part of it. I don't know if this is their rationale necessarily. I think for them, maybe it's more about like trying new things in a way that's, easier without the red tape that they would need to do, you know, but, but, you know, you can have a student who can go through the same program in any, in any program, in any institution.

28:27And if the professor has gotten really excited and, oh, you know, let's try using LLMs, then they can learn about LLMs and how can LLMs apply to their field, you know, and, and where are the problems with it and where are the benefits potentially. And another professor might not want to touch it at all. So there's not necessarily a standardized outcome. Anyway, that's just kind of my perception. It might be totally wrong, but I think that's part of it. I think that as much as institutions, traditional institutions want to invest and, and, and, you know, adapt for where things are going, it is pretty challenging for them to do so for many, many reasons. And that's why I think there will be new educational institutions that will increasingly take some of the pie, not to say that like, universities and colleges are going away or whatever, but like, there is a need for new ways. And yeah, I mean, it's nothing new. This is always happened.

29:27The question is, is it going to be like, you know, the newspapers where they get just totally disrupted? I don't know, you know, education is such a challenging place to be. It has been for so long. There's no money in it really. You know, when we were trying to do this in 2018 to 2020 and, and, and 2021, there are a lot of companies, for-profit companies, nonprofits that were kind of trying to get into this and then they all gave up because it's just, you know, it's just not easy and the funding is, is not there. And so anyway, I'm a, I'm a big believer that like at some point, whether that's 10 years a lot of what education is will be different. I mean, that's nothing new to say that. I think everyone, most people kind of perceive that. And, you know, there will be some major new organizations that have arisen out of that. And so I guess, you know, if we can keep surviving, maybe we'll, we'll be one of those that's well known and, and hopefully then we can have some impact on the broader AI field.

30:28If we can get to some sort of a scale that we have, you know, that we can have input and have things happen.

30:37Sacha: And I hope so too. One of your recent posts, in one of your recent posts, you mentioned a new AI accelerator program. Could you explain what it is, what this new program is about and how it will bridge the gap between academia and the industry?

30:53Guest: So, um, I guess the, so, you know, we already had training programs that we were doing, um, and we had students working on projects with companies. And then when we pivoted to college students, we're like, okay, let's, let's kind of reset. Um, we, we stopped doing projects with companies at that, at that time, because we figured students weren't ready to kind of provide value at the level that, that was needed. Um, and after a couple of years of doing that, we thought, okay, now we think they're ready. Everything's ready, you know? Um, so we think college students are able now to provide value to companies when working on AI projects. I mean, AI projects can be everything from like, let's do something new from scratch. Obviously that's difficult, um, to, you know, let's use existing tools, use, you know, open source LLM models, you know, large language models to create a chat bot for, um, you know, some medium sized enterprise. Um, so the, the capacity is there. Um, the tools are there, the accessibility is there.

31:54We think the value is there. So that's on the industry side, um, on the student slash academia side, there's huge demand. I mean, it's, we do almost no marketing, right? And there's, of course we have our partners that help, you know, in terms of sending students our way, but like, there's just huge demand to, to get into programs like ours. And, um, so in terms of how big we can get, it's not the student, you know, that it's not the student amount that is the barrier to growing. Um, it's, it's really more on convincing these partners and industry partners in particular, um, to try doing this with, um, with interns. Um, but also there's lots of companies that hire interns that are college level anyway. And so we think that our programs, um, add value there. So for many companies, you hire an intern, um, you have to do a lot of handholding. You have to, you know, spend a lot of time, you know, your, your highly paid employees are spending time, um, explaining how to do things, explaining how they should be done, um, answering questions.

32:59And that's super costly. Um, and, and so our programs cost the same. You just pay the intern. It's subsidized by my tax. And on top of that, you have our team of experts to kind of manage that process. So we're the go between, um, you know, insofar as the companies want to, if they want to be completely involved in every step, that's fine. It's great actually. Um, you know, but many companies, it's, it's great for them to be able to say, okay, here's what we want to have done. Here's how we want to do it. They talk to us. Um, they can check in with the students whenever they want. Uh, we, we ask them to check in, you know, maybe once every two weeks or something with the students to make sure there's that connection and, and, um, you know, that, that students have a lot of, of something, uh, with the company. Uh, but for everything else, like our team kind of manages the students, sees where they're failing and, and, and, you know, you know, it gets into that to kind of help them to, to move forward and, and to, um, not guarantee, um, that, that the students will be able to execute, but really increase the chances that, uh, value will be delivered. And, and so I think all of those things combined in terms of helping students learn and get into this field, um, helping academia or academic institutions in terms of, you know, here's some skills that, that your students are wanting to learn and you don't necessarily have, um, enough available at your institutions to, to, to, to provide to them.

34:20Uh, so we can kind of help with that. And then, you know, in, in terms of helping industry to not have to manage, um, their internships as much. Um, so, so I think all of that, there's a lot of room for growth, uh, for these types of programs, not even just saying us. Um, and, and I'm sure that there will be a lot of that, um, over the next several years.

34:39Sacha: Hmm. On a different kind of topic, you know, responsible AI is also talked about a lot these days, and I'm curious how does your company and the AI Launch Lab integrates ethical AI, responsible to this program?

34:53Guest: So when we started, uh, we didn't think about this at all. Um, we just kind of deferred to the companies that we worked with, like, what do you want to do and how do you want to do it? And then we just tried to help in whatever way we could. Um, of course, pretty quickly, we realized that not only is there a need to ensure that these projects are done ethically and, you know, data privacy and data security and everything else that I'm not qualified to speak about, but our team is. Um, but also it's a value add. So some companies maybe have a complete understanding and know how of, like, how this should be done and what are best practices and what are new, uh, you know, methodologies to do this in a way that not only is better for society, but better for their company as well. It reduces, um, litigation risk. You know, there's, there's so many benefits, uh, it reduces, um, uh, government, um, you know, in terms of like what laws will come in. Um, if you're at the forefront before these laws come in, obviously that's beneficial rather than having to go back and, you know, like redo everything like years later.

35:58So for us, it became a big priority. We wanted it to become, um, a main value add that our organization brings to the table. Again, we're not trying to force anything upon companies. Um, but that also has to do with fit. You know, if some, if a company's like, oh, we don't, we don't care about, you know, where we get the data from or how we're securing the data. I mean, then, you know, it's not necessarily a good fit for us, but, um, but it's more, we want to be able to provide value if that is beneficial, not just for the students, but also for the companies. And I think that's increasingly going to be important, um, you know, for us and for anyone who's involved in this field. Um, it will become a different differentiation point at some, at some point, you know, I don't know when consumers will necessarily be thinking about it, but, um, I can't imagine, you know, especially cause right now, like the field is just get to AGI, right? It's just, it's a scale scale. I mean, I'm obviously exaggerating, but like, um, and, and there's so many things that fall by the wayside. I mean, all the ethics people being fired, you know, from companies and, um, at some point, I think that will kind of swing back the other way.

37:04Um, and it'll become important for open AI and I'm not to speak for them, but like for anyone who's involved in, you know, startups and bigger companies to be able to say like, you know, very much like Apple does, right? Apple claims their privacy first and all this stuff. And some people might say it's not necessarily the case or not, but that's something that they try to, um, differentiate their company based on. And I think that'll be the case with whatever AI ethics, you know, entails. Yeah,

37:32Unknown: I'm fully

37:32Sacha: agree. I've noticed that trend also seeing recently with Apple Databricks, they all kind of take that stance where data resides, you know, on the consumer side, which I think is good. Uh, but we're seeing that, that trend, it needs to come from big players like that also to set the example for the broader industry. So I think that's, that's really cool. And I, and I fully agree with your, your take as well. And hopefully the pendulum will, will actually swing back and, you know, people will see more people in ethical AI, you know, get rehired and stuff like that. I think that's, that's really important. Um, so for your company, for the AI launch lab, what are some, some, uh, ambitious goals that you see or that you guys set yourself for the next five, 10 years? Uh, could you share some with us?

38:18Guest: Sure. It's always, um, it's always, uh, you know, I guess there's risk in sharing, you know, what, what your kind of, uh, internal goals are. Um, but you know, when they say like when you, um, if you're like, oh, I want to get into shape this year or whatever, uh, they say it's good to like tell people to like post it on social media. Like I'm going to do this. Cause then you're like, oh wow, now I got to do it. Right. So I don't know. I'll just say like, um, you know, our partners, so if we want to have an impact as an organization and, and certainly it doesn't have to just be us, you know, anyone else who's doing this work, uh, that I'm aware of or not aware of, I hope they're successful, uh, because there's just a need for, for the types of things that we're doing. And if anyone can do it better, that's great. Um, you know, we're a nonprofit, as I said, like, we're just trying, we're, we're in it for the mission, not to, you know, to, to, to become rich or whatever. Um, but, uh, so, you know, uh, if, if it's the case that this is going to have to be, you know, in 10 years or whatever, that, that the majority of people, maybe all, everyone will need to be AI competent in whatever that means at that time.

39:23Um, you know, you're talking about not just hundreds, you're not just thousands, right? This is like, you need to have, everyone needs to, to learn these skills. Um, maybe, you know, at some point it won't be necessary to do a training program in the way that we're doing. Um, uh, I, I don't know. I think that the field will keep advancing. So there'll keep being a need. Um, but of course, like the tools, maybe that exists now at some point, it'll just be like, oh, anyone can just use them in the way that most people can just use like Excel now without doing a training program to learn it. Um, but, um, you know, Dawson has 30,000 students, something like that. I mean, we have to get to tens of thousands of students. So we, we really, um, have plans to get to tens of thousands of students. Let me say five, at least five figures. I wouldn't say tens, but like, we want to get to that scale, um, by 2030. And, uh, you know, that's a daunting task, but I mean, in order to have the impact that we want to have in order to bring in the amount of people into the field that is needed, um, if we want to have a plan to do that, um, that's definitely what we're striving towards.

40:26Sacha: Great. So last questions. I want to get your personal opinion on that. What happens if we get to AGI?

40:35Guest: Uh, so preface with, I'm not an AI expert. Um, I'm new to this field. I'm, my background's finance, but, you know, of course I'm addicted to, to this topic, just like anyone else is. And, um, you know, uh, you know, it's, it's, it's really hard to have an answer to that because obviously there are people who say that it's kind of nonsense to even ask this question. Like, um, I, I am a little bit partial to the idea that a lot of the talk about AGI is a bunch of people who have been in this field for, for years and now people who have not been, you know, for decades and, and who have been obsessed in their little circle talking about what's going to happen when the singularity comes. And, um, it's, it's almost this kind of religious thing, right? It's like a cult-like thing. And, and these are the people who have become the leaders in the field, right? I mean, who was working on this when nobody cared about it? It was people who were like, oh my God, AGI, AGI, right? So, um, but you know, it's, it's pure, it's, I mean, it's extrapolation, it's speculation. Um, we don't know what this means.

41:35We don't know if it's possible. Um, you know, you can look at a chart and say, yeah, if this chart continues, then I guess at some point we'll get there. Um, maybe it'll go like a hockey stick or maybe it'll plateau. I mean, who knows when the plateau is and when the hockey stick is. But anyway, um, I think it's, it's exciting to think about. It's scary to think about. Um, I think a lot of it is, um, not worth thinking and talking about because it's not relevant to what, what work needs to be done now. It's not relevant to the harms and the risks that exist today. Um, so I'm a big believer in focusing on like current harms and risks versus future harms that may or may not happen. Um, uh, but I, you know, I think it does make sense it forgetting about super intelligence and the idea that like this thing is going to become smarter than everyone and take over the world. And, um, you know, I, I think it does kind of make sense that these tools will, you know, I was listening to the, I don't know if I mentioned it already, but like there's podcasts from, um, uh, I don't know what podcast it was, but it had Ethan Mollick on it.

42:37Um, who's kind of a, a big name influencer, um, in, in the AI space. And he has his kind of four, uh, trajectories for the future. And, and one is, you know, kind of this AGI super intelligence thing. Um, and then the other is like, actually we just, you know, hit, hit the ceiling or hit a wall and, and, and nothing happens beyond today. Um, but even in that scenario, I think he said that, you know, we have, and Mark Zuckerberg said that came out yesterday as well, that there's still years of just using what's already there, implementing basically. Right. So there's still a lot of room for that, uh, to, to, to kind of use what's already possible. Um, and then the other scenarios are you get this incremental growth, you know, every year, it gets a little better, a little better. And I forgot the other, the other scenario was somewhere in the middle. Right. And, um, you know, it's, it's a safe, probably it's a safe bet to assume that it's somewhere between we've hit the wall and it's going to be incremental increases. Incremental increases over the course of a decade get to, you know, that gets you far. Right. So, um, I, I don't think it's, um, outlandish to believe that, um, a lot of office work, a lot of, you know, um, white collar work will get taken.

43:49And the idea of, is it replacing or is it augmenting? I mean, um, you know, I, I just don't think that we're getting those hours back. You know, the idea that we're all going to be like sitting and lounging while AI, AI is doing the work for us, I think is nonsensical. Like I, I honestly don't understand these people don't seem to understand economics to me. I don't know how, like barring some massive political change that we suddenly decide that people don't have to work. Uh, I mean, I, I just, that's, I don't think that's happening. So, you know, we're going to be working as much or more, and we're going to have these AI agents working alongside us, I suppose. Right. And, um, you know, I think one metaphor that I was thinking about recently is, you know, the switch from, and this is also from Ethan Mollick's, um, podcast. He mentioned how in the initial transition from, um, I don't know, whatever came before, like manufacturing. So kind of artisans, right. Or whatever, like blacksmiths and whatever. And then we went to like this factory type situation. Um, initially it wasn't that everyone was just an assembly line worker, just kind of a cog in artisans who knew how to best use these machines who were figuring out, oh, like it can do this, but here's this way that we can like connect it to this other thing and make it do this thing, you know?

45:03So for a time, and actually there's a connection with the Luddites in there. That's also pretty interesting. If anyone reads up on it, the idea that we have of Luddites is that they were like anti-technology. And, and actually a lot of what they were was not anti-technology, but anti how technology was beginning to be used by factory owners. So they were actually parts. I think I'm getting this right. Many of them were working in factories as those artisans, and they were helping to make these things work and designing new, um, you know, uh, ways to make them work in the factory. But increasingly factory owners wanted to produce the value of their labor. They wanted to go from it being this artisanal type of work, um, to take away all of the individuality of the work, take away all of the skills that you're needed so that you're replaceable. Right. And that's really what they were against. And so we're kind of in the stage right now where the people who will succeed in, in AI, I guess, broadly are these kinds of artisans, right? They're not, it's not, we're not at the stage where it's a factory, but probably we're going to get there at some point.

46:09Right. So that's, I think, rather than the idea of AGI, um, increasingly it's like, oh, okay, well, you don't have to answer your email. Cause AI will do that for you. Um, for some time you'll have to monitor it and, you know, oversee it or whatever. And eventually that you won't need to do anymore. Um, so what are the tasks that we're going to do is I think to me, what's more interesting, what will be left. And that's like, I don't know, anyone can kind of speculate on that for, for years, people were saying, oh, the creative stuff is what's never going to get taken away. And, and, and I always was a little bit suspicious of that because I thought like, yes, creativity, I get it. I, you know, I don't understand the brain and you know, I'm not, I wasn't like an expert or whatever, but like also, you know, we have what we listen to on the radio is not like, you know, these are not like necessarily the most amazing composers of music, right? They're putting out a product that does well in this distribution medium. So can you have AI copy that? I mean, to me, that made sense. So anyway, so gone, we've gone in that direction. And actually now the, the, uh, the, the expectation is the creative stuff is going first.

47:14That LLMs can't do. Um, you know, uh, so I don't know, but, but that's, I think the, the question and the scary thing, if you, if you're thinking about it is like, what, I don't think there's an answer. Like, what would you teach your kids if you had kids or whatever, like to what, what should they learn? Um, you know, the idea is like liberal arts is now like the thing, but like, I don't know, are people hiring, are companies hiring for liberal arts people? I mean, it's, I think it's like, you need to race to get to the top of an organization as fast as possible before the, the lower levels just get stripped away. And this is kind of how I was feeling about finance, like years ago when I went back to school, because I would apply to jobs on LinkedIn and on LinkedIn, when you have like the premium thing, it'll show you like how many people applied and what is their education level. And I was like, oh my God, this is like an entry to mid-level job, like low, you know, like a low level job, like nothing high up. And it'd be like 400 people applying to this job. And, you know, 30% have, uh, you know, an MBA and of those 15% also have a, like a, you know, I was just crazy.

48:14Right. And I was just thinking like, basically, if you're in the field already at a certain level, you're fine and you're actually doing really well because efficiency and productivity is up and profits are up, but who's hiring like the lower level. And, and, and so this is happening in computer science now too. Right. So that's another reason why I think it's so important that we can do whatever is possible to get these people who are entry level, who are getting into the workforce in, in any domain to try to get skills that can get them in, because I don't know, like the, the, the lower rungs of every organization are getting stripped away. I mean, we're seeing this with layoffs at big tech. They're like, oh, profits were up 30% or whatever it is, you know? Oh, and we laid off another 20,000 people like quarter after quarter after quarter. Anyway, I think that that's, that's the thing that I'm passionate about as to whether we're going to have some sort of a God AI thing. I don't even know where to begin to think about it.

49:10Sacha: Interesting. And you know what you just mentioned makes me think of, I think it was Jeff Bezos that mentioned in one interview that a lot of people are asking or building their business strategy based on what's going to change in the next 10 years, right? And for him, it's more about what's not going to change in the next 10 years. So he's looking at that and then he's asking himself, okay, so how can I build my strategy? So for example, he was mentioning something like, okay, so now customers will never not want a product, you know, delivered faster. And so that's how he's building his strategy and implementing. And I think that's a good question to ask also in, in AI. So what's not going to change? And then you can look at that. And that really made me think about what you, what you said with, with this, with, with this quote from, from Bezos. Anyways, I want to get one last question in, and it is really about your own personal hopes or concerns about the future, or even like the current state, because you said that you'd like to look at the current state of AI in society. So I'm curious to get your take on that.

50:10Guest: So, I mean, the current state is, I think, you know, it's unfortunate that there is such a, it's, there's so much inertia in just getting to scale, which of course, it makes sense economically, like you need to get the biggest, best possible model. And that's the, that's the straightforward way to get, you know, to get there is to just spend tens of billions, you know, billions or tens of billions of dollars on, on training that. But, you know, the problem of course, is that like, if you don't, which we don't really understand how these models work and why they do what they do. And, you know, there is work going into that, but it's, it's, it's tiny, right. In terms of relative to the amount of money that's going into like, let's just make these things bigger. Yeah. I mean, that's, that is a scary thing, right? I mean, it's, there's nothing new to talk about. It's been said for years and years by people who've been, you know, in, in AI ethics and, you know, AI generally is, you know, if you, if you have, if you increasingly outsource more tasks and decision making increasingly to AI systems, and you don't know the process and how, why they come to the answer.

51:23I mean, that's, that's, you know, it reminds me of this example of the, the correct, the Apple card. Do you, I don't know if you, it was a few years ago. And so Apple had this credit card with Goldman Sachs and I don't really know the back end at all. And, you know, I just kind of read the articles or whatever that came out about it. But, but I think there was a famous founder, I forget his name. He, he said that like him and his wife applied for the Apple card and his wife got way less credit and like just on worse terms than him. And, and, and he, and also Steve Wozniak, the co-founder of Apple also had the same exact issue. So both of them, it was like, you know, I think both of them, their, their finances were together, like they're married. In theory, according to what they were saying, there should have been no difference in terms of what was offered to the, you know, their wives and to them. Right. And, and they got these hugely different offers and they went public with it. And, and I don't, I don't think Apple ever really came out and said what went wrong. They just kind of like fixed it for those two people.

52:26And then never, we never heard back. I could be wrong, but I don't think anything really, obviously Apple's not going to say like, you know, what went wrong. Cause it's just bad PR, I guess they just want it to be buried. But you know, this is like the thing and it's, it's going to happen everywhere. Right. I mean, Amazon had this with hiring and then they just gave up on using it for, for hiring. Cause they were like, oh, it's, it's being discriminatory in terms of who, who wants to hire. Um, and a lot of people say it's just, oh, it's the data that they use. Right. And, and, and that's probably part of it, but I mean, we don't really know. Right. Um, so that is a pretty scary place to be like, of course, human judgment is also bad. Um, I, I worked for a company, uh, a couple of years ago, um, that was focused on small business lending and they were trying to, uh, well, I guess one of the projects they're working on was on small business lending and trying to kind of reinvent small business lending. Um, because small business lending is difficult. Uh, many small business owners can't get loans, et cetera, et cetera. Uh, credit scores are, are pretty bad. They're not, they're not an accurate, um, indicator of whether somebody should be getting credit and all this stuff.

53:31And, um, and, um, so I did a lot of interviews with small business owners at the time to try to figure out like, what are their issues and, and, and, you know, what should we focus on in building a product? And, and some of them, uh, interviewed in, in the South of the U S, uh, and not just the South of the U S, but like that, that, um, um, um, you know, we, we didn't interview some people with, with these experiences and they, they, um, they'd say like, oh, I, I faced racism. Like people would just like straight up be open about it, you know, at, at banks. Um, and so it's not to say like, oh my God, AI is suddenly going to cause these problems that don't exist. Like they do exist, but like, you know, I don't know, you can try to stop racism or you can try to have policies in place to limit it. And, um, you know, but the problem is that in, in, you know, banks have in, in order to the risk of those types of things happening bias from their employees, the process for getting a loan has become so automated in a sense, so systematic that actually it's just, it's just, you know, it's just hopping to jump to like, have it just be AI black box.

54:33Right. Cause they want to take away the humans completely. Cause if you have a human, then, oh, somebody can say that person did this thing and it wasn't according to the law and then you can get sued. But like, we don't have access to the algorithms. We don't have access to how, and even if you did, you don't really know on a individual by case basis how it works. So that is pretty scary. And, um, you know, I don't know what the solution is because the economic imperative to use these things is far too high and governments can come in and try to like mitigate harm. But like the status quo is like, don't constrain us because if you constrain us, China's win or, or whatever, whatever the rationale is. Right. And that seems to be winning. So, um, I don't know how we avoid this kind of Hollywood dystopia scenario where it's like, oh, I suddenly I just, I, you know, whatever movie you want to think about, oh, I can't get into the city today. I don't know why. Like it read my ID and it said, no, there's no followup. There's nobody to ask. It's not going to tell me why. You know what I mean? Like there's, I don't see how we're not getting into that kind of a place.

55:34Of course it's going to take a long time because people are going to fight against the have more scandals and I don't know when the big one will be, but this is why, of course, it's important for companies to focus on, uh, doing things the right way to avoid liability. But promise like, if everyone's kind of just like, oh, we're all agreeing that like, we can't have constrict, you know, constraints other than, you know, little window dressing on the side or a few million here, but like tens of billions here on like just getting up to the prize. Um, I mean, what the government's going to stop the field. It's not going to shut down like open AI and Google and Microsoft, you know, and, and, and meta, like it's, it's just, so I don't know. I think it's, um, you know, like with everything, um, people, people vote, people fight back, people protest, I don't know, whatever method that needs to happen. I think in order to have us not end up in a dystopia, it's going to be these types of social political movements that will have to just like they fought for minimum wage and, you know, worker protections and stuff.

56:38Uh, I think that that's, that's what's going to have to happen with regards to AI too. Um, cause the economic incentive is to just go towards this place where humans are just kind of cannon fodder. Like how do you, are you helping with training the model? Like otherwise, right. And we've seen how that works in, in Kenya with the, um, train with the, the fine tuning or, or, you know, in terms of like, um, open AI hiring, um, and, and using Sama source, um, you know, there've been this horrific stories of people and, you know, and it's been before with meta in terms of like, um, looking at, um, um, looking at, I guess, like horrible stuff that would be shown on, you know, on the, on the feed. And they would have to be like, no, that can't be seen though. No, that's not acceptable. So, um, anyway, I'm kind of rambling on there, but I feel like it's that, that like work needs to be done. People need to get involved. Um, and I'm hoping that, um, there are more organizations that have a voice that get to some sort of size to where governments listen to them.

57:38Um, you know, every time the governments have these, um, they've been doing these, these, these meetings, like the UK did them, uh, Congress, US Congress did them. Almost everyone is like, oh, it's an executive or a C it's a CEO from big tech. Right. It's like, these are the people who are in, in the rooms. Like it's not, it's not like somebody who got discriminated against and didn't get like the credit card offer that their husband got. Right. Those are people not, those are not the people in the room. So anyway, um, not to end on a downer there. I think that like, you know, one of the reasons why I got into this field, uh, in addition to like, oh, my job is my field is actually getting disrupted is, is that I think AI can be a catalyst for positive change. Uh, I don't think we're anywhere near that right now, unfortunately, but I think we can push for that. Um, you know, what does an economic system look like when you have AI agents running companies or helping to run companies? Um, a lot of things need to change in society and, um, how we build AI, how we integrate it into our society will make all the difference in terms of whether we end up in a dystopia or I don't think we're going to utopia, but whether we end up in a better place.

58:50Unknown: I

58:50Sacha: mean, that's a great, that's a great take. And I agree. And then one thing that, that you made me think about is I find it very interesting that a lot of, that it seems that there is more scrutiny around a technology than there is around, uh, humans, right? Because of course these LLMs and these models are black box and it's hard to tell how they took decisions and there is bias and, and there is a need for, for people to work on that and mitigate that. And, and I fully agree with that. However, humans are also full of biases, right? And sometimes we don't, we don't, uh, emphasize enough on human accountability as well. Like for, for example, for when people are hiring and stuff like that, you know, there is also a point with, uh, the example you said about self, uh, in the South of the US with people, you know, discriminating, uh, racially and everything. So I find it very, very interesting that there is more scrutiny around technology than there is around, uh, humans, but that's something to think about.

59:53Uh, Timothy, that was a really, really good chat and really appreciate it. Um, I want to give you like one last word if you want to, if you want to share something or say something to the audience, but it was a real pleasure to have you on today.

1:00:06Guest: Yeah. Really awesome to be on. Uh, thank you so much, Sash. Um, really love your, your, your interviewing, um, and, uh, looking forward to listening to your other interviews that you've done. Um, I guess, so I would just end on, I don't know what, what, what one to end on, but I guess, you know, I, I would just hope that like people who, um, are not in the field of AI, um, that they don't feel like it's not for them or that they don't know enough. Like, like this field is, is, is changing so rapidly and, um, it's becoming more accessible, but not just that. Like, even if you're like, I don't want to work on AI projects or I don't want to use AI tools, like we need to have people involved who have domain experience. We need to have people who have different perspectives. And, um, so it's, it's a way I think to get into this field and in whatever way that means for you.

1:01:07Um, it's a way I think to, to increase your potential impact in your life and your career. Um, I mean, I'd say definitely the only reason why I am doing what I'm doing today is because, uh, you know, AI as a field has, you know, this immense wind at its back, right? Um, there's no reason why I should have been able to, uh, be shared on social media by Yashu Bengio or to have him at, you know, our hackathon or, you know, I mean, or to have just any of the companies that we speak to that just even respond to my email. It's just because I'm working in this field, you know, and if you believe that AI is going to transform industries and society, um, then why not use it as like a, a leverage point to have the impact that you want to have? You know, like we've, we, we're seeing how it's, how it's transforming industries. And so the future of so many industries is going to be kind of AI first companies and not there'll be a give and take, but like, anyway, I guess that's just to say, like, um, I think that it's, it's a really great opportunity for people to get into this field and not just to like, uh, you know, better yourself, but to have the impact on the world that you want to have.

1:02:21Sacha: I fully agree. That was a beautiful last note. Thank you so much to my team.

1:02:26Guest: Thanks a lot, Sasha. I really appreciate it.