AI Leadership Insights: MIT Sloan AI and ML Conference

Mike Vizard: Hello and welcome to the latest edition of the Techstrong AI Leadership Insights series. I’m your host, Mike Vizard. Today we’re going to do something a little bit different. We’re talking with some of the leaders of a AI event being hosted by MIT, and the first thing that will leap out at you is, well, how young they are, because it turns out that the kids are onto something and maybe we can all learn something from them. So stay tuned. Hey guys. Thanks for the throw. We’re here with the leaders of a virtual plus physical event. The physical event is already sold out about AI and ML that’s being held over at MIT this weekend, and we’re encouraging folks who are students and maybe even not so many students to come and sign up for this event because, well, sometimes you got to know what the kids are thinking because they drive a lot of this stuff. Folks, welcome to the show.

Tim Valicenti: Yeah, thanks for having us.

Eunhae Lee: Thanks for having us.

Ashwin Bhide: Thanks for having us.

Mike Vizard: All right. So how did this show come about, this event and what’s the ultimate goal here and who exactly are you hoping will attend?

Eunhae Lee: Sure, I can jump into that. So we are the MIT, we’re running the MIT Sloan AI ML conference and we’re the largest and only student run AI ML conference at MIT. So our leadership team is completely comprised of students. This is our second annual conference and this year’s theme is Leading The W(AI)ve with A being replaced by AI. So it’s like W(AI)ve, and our goal really is to empower the next leaders in AI. We are inviting speakers from top tier AI organizations such as Google, Microsoft, OpenAI, Anthropic, and even the White House, and we have a lot of exciting startups joining us. We have about 400 in-person attendees and we’ll have live-streaming available as well for those who are watching this and want to attend, and yeah, I’ll leave it there.

Mike Vizard: What’s your assessment of the current state of AI? I mean, sometimes I will talk to business people and there’s a certain amount of irrational exuberance that goes with this and they think all this stuff is just magically going to happen and there’s other folks who are, well, the IT systems that we have in place today are something of a mess and it’ll take us a long time to get them AI ready and we need to think about this longer term. I guess we’ll throw that one over to Tim, but what’s your sense of where are we on this journey? And then we’ll come around to Ashwin.

Tim Valicenti: Yeah, absolutely. I talk with a lot of industry leaders and some of the executive MBAs that are kind of in the weeds of this. I think we’re still at 1998, 1997 version of this. I think a lot of founders are also confused on kind of where to build. Are they building the chat bot tools? Are they building kind of where the puck is going? Those are some of the insights that we’re planning to kind of extract from the conference. We have a lot of founders that are going to be talking about what moats they’re building within their companies to compete with the big guys. We have some of the big guys like Google and Microsoft that will be talking about how some of their enterprise clients are using these tools today. And it’ll probably be, I mean the answer is it’s going to be another very like answer. It depends. And kind of we’ll have to wait and see.
I think many of the big tech companies are definitely leading the way and they have the infrastructure. And then on the other side of the spectrum, it’s a whole different ball game and people are still kind of getting caught up on other internet technologies and layering AI on top of that, maybe overwhelming. And the technology piece is just one part. I think the behavior change and waiting on consumers and business stakeholders that will be using this day to day is a whole other piece that we need to wait and see. But again, all topics that we’ll be talking about at the conference throughout our panels. I think we have 12 or 13 panels or so, so.

Mike Vizard: Ashwin, you want to add to that and maybe a little perspective on, we’ve been dealing with this divide between business people and IT for as long as I can remember. Are we finally at some point where maybe AI will bring us all together?

Ashwin Bhide: Sure. Yeah. So I think it’s important to kind of underscore that business people and IT are going to be working with one another to actually integrate IT into the business. It’s not as easy as there’s going to be a potential divide between the two for this case. But that said, I also want to make sure that we kind of underscore the importance of speaking about these kind of events with, or these kind of topics at the conference. So we do have a bunch of different panels that are going to be talking about this such as AI and the enterprise. We have a panel on open source versus closed source. And yeah, I think a conference is probably the best venue for listening to both enterprise companies as well as smaller startups to kind of understand where this is going to play out and how this is going to look five years from now, 10 years down the line.

Mike Vizard: Maybe one of you can jump in on this one. What was the hardest part about putting the conference together and what do you know about this topic now that you wish you knew when you first started this whole thing?

Eunhae Lee: Yeah. Well, I think the biggest challenge of organizing the conference, I think it’s the same with any conference. We do have a lot of moving pieces around, but at the same time this year we actually have a pretty robust team in operations, content, in marketing, employee engagement, community and so on. And so with fantastic work of all of our team, we’ve been able to pull together a great conference. We’re super excited about this Saturday.

Mike Vizard: All right. So you didn’t just bootstrap this thing, you actually ran it like a business? Is that what you’re telling me?

Eunhae Lee: Yeah, actually, well, this is our second year. Last year we were more bootstrapping because it was like a startup-y vibe. This year we were able to build a larger team and try to put some structure around it, and I think it’s been going really well so far.

Tim Valicenti: If you wanted one thing of what we wish we knew, one area is like, we knew there’d be high in demand, but we actually have to use a second auditorium now, and we’re starting to plan events that were going to have two sessions in parallel. I think had we known that the demand was even much higher ahead of time, we would’ve maybe made this a two-day conference. We might have to go that direction next year, we’ll see, but we’re already at the point where our main auditorium was full of events and we’ve had to move into parallel spaces now.

Mike Vizard: Ashwin, do you think that there’s a generational issue here? And as a boomer, I’m asking you this question, so. Do the old guys get it or are they just a little slow off the mark?

Ashwin Bhide: No, I mean, I think everyone gets it. I think just my personal experience, I’ve seen that my parents, my grandparents, they’ve started kind of playing around with it as well, like the easy to use AI tools like ChatGPT. I think ChatGPT has done such a good job. I mean, OpenAI has done such a good job with ChatGPT making it so easy to use, that seamless interface, easy to chat with. The AI takes on different personas. I think it’s done such a good job kind of personalizing the experience to anyone who actually use it, that I feel like folks are kind of getting the application of it rather than understanding how it works internally, but I think that’s going to unlock a lot more productivity for everyone, not just the younger generation. Every single generation is going to have an increase in productivity.

Tim Valicenti: I think that says a lot as well. And if other generations are seeing value immediately, you give it to your parents, your grandparents, and they say, oh, I see how I could use this for X, Y, and Z. It’s not like all of them have a blockchain running on their computer. Blockchain has its use cases, but the fact that non-techies can see value here is maybe a good sign that this is here to stay.

Mike Vizard: Eunhae, do we just not know enough about AI? And I’m asking the question because things you don’t understand you’re naturally afraid of, and we get a lot of that fear and loathing kind of discussion going, but is part of the exercise here to kind of take the mystery off of AI, and this may be just ultimately an exercise in math that we can all understand?

Eunhae Lee: Yeah, no, I think we honestly, even the people who are in AI could probably empathize with the fact that the world is evolving so fast and we constantly need to learn new things. I think it’s just a phenomenon that happens for everyone. I think actually, so the focus of the conference is to really provide practical insights for everyone, especially for those who want to build new products or build new businesses around this. So we really want to bridge the technology and innovation and also the market side together.

Mike Vizard: Of course, there’s a lot of concern or as [inaudible 00:09:43] said it, “It’s one thing to be wrong, it’s another thing to be wrong at scale,” and that’s his fear with AI. But Tim, what is your sense of what’s the assessment of the safety that we have in place, the guardrails? Are we putting enough focus on it?

Tim Valicenti: Yeah, always a loaded question, right? We have a whole panel at the conference talking about policy and kind of safe use of AI and whatnot, and I think that is something that students want to hear about and even non students are attending. It’s mostly for students, but folks externally want to see this as well. It’s tough. Part of this, we might look back in time and say, we didn’t know what we didn’t know at the time. I think you could have said that with the internet. You could say that with almost every technology shift. People have the same view on jobs and job displacement. I think these tend to create new jobs and the fear of this displacing everything is normally not validated when you look back five, 10 years, and there’s a lot of different ways that we’re going about this. Right. The OpenAIs of the world believed originally that open source was the way to go to make this safe.
They have an entire team dedicated to alignment of their AI tools with what is best for society, and now they’ve changed to closed source, and maybe that’s the answer. We have all these open source companies like Hugging Face and Stability, and they’re going to be, with Stability, we have some high representation on the open source side at our conference and they have different views. I think the good thing is that we have people building in both of those camps, and that’ll be more robust for society to reach kind of that better answer, but I think most people want to use this for good.

Eunhae Lee: Yeah.

Mike Vizard: Go ahead, sorry.

Eunhae Lee: Oh, sorry. I just wanted to add that I think that’s why I think we have a very unique position as a conference that’s situated in an academic setting because we’re able to really have an open dialogue about those things, including those challenging questions and really hear from different perspectives.

Mike Vizard: All right. Ashwin, let me ask you something. The perception today is that AI requires data scientists who are borderline rocket scientists, and then I got to get a small army of data engineers together. Do you envision, or maybe it already has and we just haven’t quite discussed it, but is this going to be democratized where just about everybody and anybody can use it to do something and maybe that’s kind of where we need to get, rather than having it be this, what still feels a little rarefied?

Ashwin Bhide: Yeah, I think the latest developments in AI, especially with these new models that have come out, I think it’s really democratized AI and ML. I feel like, for example, I think Perplexity came out and said that the first version of their app was basically, it was like a Q&A service on Twitter, so how many followers does X have, how many followers does Y have and whatnot. And what it basically did was it used I think, OpenAI or another other LMSs APIs behind the scenes. And it converted that natural language into text and it ran a query, sorry, it converted the natural language into SQL and it ran a query to actually pull that data back.
So with such ease in actually retrieving data and playing around with data, I think it is democratizing the process and making it easier for folks to become data scientists or just analyzing these massive data sets. I think the one thing that it might still be a challenge with is with super large data sets where you might have to run distributed compute over many different servers. But for small data sets, I think it’s pretty easy to play around with.

Mike Vizard: All right. I’m going to go a round Robin here on this last question starting with Eunhae. What’s your best advice to folks, because I think people are struggling with where to get started. So do I just dive in on the deep end of the pool or is there a rational way to go after this thing?

Eunhae Lee: To me, I think the most important thing is the mindset of, so I think a lot of people are familiar with the concept of growth mindset. I think it’s very easy to feel that this whole new thing, it’s very daunting, especially people who don’t have a technology background, but everyone starts somewhere. And I guess a little personal anecdote is that I also started from a non-technical background and an example of how I continue to started learning and then now I’m in this field. So I think it’s a field that now is open to a lot more. There are a lot more opportunities in this field. It’s definitely not just for technical deep tech people and yeah, starting somewhere I think is the best way to go.

Mike Vizard: Tim, thoughts?

Tim Valicenti: Yeah, I mean at the conference we have a hands-on workshop with open AI where you don’t need to be technical, but you can get your hands wet or dirty to actually build a kind of version of a GPT that does a task for you, maybe looks at some documents and how it can be used out in the world, more for demo purposes, but just a way to get your hands dirty. I think with anything, I think this is especially true or especially a mindset that’s apparent at MIT, it’s just starting to build and being okay with it not being perfect or being okay not knowing how to do it. If anything, building is easier now than ever before because you have the world’s most patient tutor on your desktop or your phone via ChatGPT. You can use it to learn things very quickly and just yeah, definitely advise everyone to start building and it doesn’t matter what you don’t know because you can use these tools to help you learn that along the way.

Mike Vizard: All right. Ashwin you will have the last word, which means it’s also your responsibility to tell everybody the URL where they can find this conference.

Ashwin Bhide: Sure, sure thing. Yeah, I think just to add to Tim’s answer a bit, we spoke a little bit about how AI is advancing and whether it’s doing good or bad, but I think this is actually a super useful case. Folks who aren’t from a technical background or traditionally technical background, they can now play around with ChatGPT to write small snippets of code for them, and that kind of helps them gain more confidence as they try to develop apps or just write code. Additionally, you have stuff like Microsoft’s Copilot as well as [inaudible 00:16:17] that actually helps you break into these more technical fields.
So I think this is actually a very positive thing for folks who are interested in learning more about it as well as trying to get into the space. And yeah, just for the final word over here, I wanted to, well, first of all, thank you for having us on this. It’s been amazing to chat with you. And for folks who are coming to the conference on Saturday, we really look forward to welcoming you. And if you aren’t able to join us in person, you can visit mitaimlconference.com to get more details about the conference, as well as get access to our live stream for the event.

Mike Vizard: All right, folks, you heard it here. That’s where to check it all out. And chances are, wherever you are right now, it’s snowing, cold, rainy, and this is a good thing to do on a Saturday. Hey guys, thanks for being with us.

Tim Valicenti: Thanks, Michael.

Eunhae Lee: Thank you.

Ashwin Bhide: Thanks so much, Michael.