Entrepreneurial Thought Leaders (ETL)
00:35:31 · Michael Bernstein
[00:35:31] S2: Yeah. I don't think there's any magic here. Like, I think what I'll probably say is gonna echo what a lot of people would say. Something that's good for everything may simultaneously be good for nothing. Right? You have to find use cases that really stick, and then you can expand out from there. Now at the beginning when we were talking to people, everyone was like, oh, this is gonna be used just for message testing of market research, and I was privately hoping, like, I really hope this is wrong because while that's important, it's not the thing that gets me up. We're much more interested in societal simulations, simulation of really tough counterfactuals you know, around health and and public policy and things like this. But people don't know how to start there. Right? Like, we're, like, ask it anything, people are like, oh, right? They get completely stuck. But if someone knows, like they... They're already asking questions. They know how to uninstall, so to speak, the thing they're doing, or like install this next to it and say, oh, I get it. And then they can sort of explore the the new uses. So we started by, you know, as you even saw in the demo there, we try to guide people toward kind of use cases that we know work well, that are common use cases that people want to ask about. We try to help them not only land there, but also have conviction that this is the right place to ask that question. Mhmm. And then over time, people start... Like, once they do that, they're like, oh, well, I can also do this, and
00:36:53 · Ravi Belani, Michael Bernstein
[00:36:53] S1: they start expanding. Fantastic. I'm gonna start to ask the signature questions, but if you have a question, start thinking about queuing up if you want to. So Michael, if you're 20 years old today and starting and starting this company... [00:37:05] S2: I am 20. What are you talking about? [00:37:06] S1: Well, you're 20 for the twenty third time, it sounds like. Oh. Oh, 20... Well, 20... The first. That's okay. Keep going. I think it's twenty third. Yeah. Like when you're 20 the first time, that's 20 for the first time. It's not 21. [00:37:17] S2: But we can Okay. Okay. Fine. Fine. Okay. [00:37:19] S1: If you're 20 years old, what would you spend the next five years learning or building?
00:37:19 · Ravi Belani, Michael Bernstein
[00:37:19] S1: If you're 20 years old, what would you spend the next five years learning or building? [00:37:24] S2: Well, I'm teaching here, so I feel like I gotta say HCI. Look, I think if you buy the Myspace hypothesis, the big gap is what problems do we need to solve? And I think that something around human centered design is what teaches us to look for these opportunities. It's this thing called Pasteur's quadrant. It's like solve real world problems that have like very deep and horizontal application. Like Pasteur, like famously... Okay. What did he invent? Vaccine. Right? Because there was a thing in front of him. He did fundamental research that was targeting a very... A problem that was right in front of him. I think this is where HCI likes to live. It's like, what are real human needs where there's real technological opportunity? That's where I like to live, and I think that Simile just grew out of that. We saw the problems in social media trying to predict what's going to go wrong or trying to understand the effects of policies, And I think that training yourself to understand how do I connect needs with technological opportunity, that is what HCI does. And so I would be taking exactly the thing that I'm teaching. Easy answer, I guess. [00:38:38] S1: What's the best mistake you've ever made?
00:38:38 · Ravi Belani, Michael Bernstein
[00:38:38] S1: What's the best mistake you've ever made? [00:38:42] S2: Applying by accident to a PhD. So I was here as an undergrad. My undergrad advisor suggested that I I wanted to apply to a PhD, but I wasn't sure if I had enough experience. He said, Okay, you should apply for master's programs first. So I did. Turns out... MIT has this dropdown. Are you applying for the SM? Because they do the Latin thing where they reflect MS to SM. SM or PhD? Turns out they ignore that drop down, and so I applied to the PhD without realizing it. And then I got in. I said, oh, well, guess guess I'm gonna do that then. So, that was my best mistake. Set me off on my path. [00:39:23] S1: And I know there's a long pipeline, so I'm gonna discuss one more signature question. We'll move to the students, which is what have you refused to compromise on and what has that cost you?
00:39:23 · Ravi Belani, Michael Bernstein, James
[00:39:23] S1: And I know there's a long pipeline, so I'm gonna discuss one more signature question. We'll move to the students, which is what have you refused to compromise on and what has that cost you? [00:39:31] S2: I think throughout my career and even since starting Simile, there is a real temptation to get pulled toward things that can be sort of narrowly defined technically. Are you, you know, benchmarking on this particular benchmark that you care about? You know, can we solve math proofs? Can we like... I've been told so many times early in my career especially, oh, what you're doing, that's not computer science. I like... I think that people have these narrow blinders on around what problems matter, and I think as soon as you take those off but you still have the technical depth to be able to actually tackle it, that is the thing that, like, I just... I have not compromised on. Like, it is wild to me that when I was an undergraduate, I was told that what I was doing was not science, and here I am now a professor of computer science. That's wild to me, but don't compromise on it. [00:40:27] S4: Don't comp... That's great. Thanks a lot for your talk. I'm James. Hi. My classmate Michael and I are building in the behavioral simulation space, and we have a question for [00:40:35] S1: And, Ging, if you [00:40:35] S4: can keep... How do you ensure that contrarian or abnormal or pathological human behaviors, which can sometimes cause, like, BACS one zero one events, are accurately modeled in your, like, data?
00:40:35 · James, Michael Bernstein
[00:40:35] S4: can keep... How do you ensure that contrarian or abnormal or pathological human behaviors, which can sometimes cause, like, BACS one zero one events, are accurately modeled in your, like, data? [00:40:45] S2: Yeah. So how do you get these long tail events modeled? Like, the post training procedures behind most language models really... Like whatever the distribution would be in this... In society, they push the the model toward very high conviction at, like, sort of the mean. This is exactly what you don't want for a simulation engine. But there's no... Nothing written that says it has to work that way. Essentially, what it requires is that you have training data, again, this sort of like dark matter that I'm describing, that teaches the model what the actual population distribution is. It should know that there are outliers who are whatever you're saying, trolls or people with unusual opinions. Actually, can give you an example. We spoke to one of these early, early on in Simile's days. We spoke to someone who was like head of data science at a really high end fashion firm and they didn't care about what the average person... They're like, show me the tails. Because we don't... They're like, we don't care what people think, but we're interested in what people are doing. That's like the unusual stuff because that's where we get inspiration. So if you want have simulation work, you have to focus... I think it's easy to focus on the mean, but it's the variance that actually, in some sense, really matters.