Introduction

My name's Doug Williams. I'm with the MIT, or the Martin Trust Center for MIT Entrepreneurship.

Martin Trust was a mild mannered manufacturing type whose manufacturing prowess caused him to own Victoria's Secret and the Limited. And he believes in entrepreneurship as a door opener, so he helped endow the center. That's where that comes from.

And what I'm going to show today is our orbit tools. There was a request in the presentation to have one of them be a workshop. So there'll be no slides. I'll be going through it.

I'm either going to use Anode as an example or possibly Screens. Can I use Screens as an example? Yeah.

OK, so one of the things I'll show probably early on is how to get to our wait list, because you all can sign up whether you're a student or not. And then another question, students and non-students, how many of you all are involved in startups?

OK, kind of about half. OK, that's a good mix.

We use this tool and this methodology. It's focused on entrepreneurship, but it works for corporate innovation, too. So MITRE attended one of our executive sessions and asked us to come out as a for instance.

A lot of the same principles apply.

Workshop Overview

So maybe one thing I can do is I'll start with how to sign up for the wait list, which is implying you want to sign up. So you just go to Orbit, log in. In fact, you just go to Orbit when you go to log in. And at the bottom of the page is a wait list. So just sign up there.

And we've got about 6,000 people at MIT and about another 6,000 outside MIT and a mix of entrepreneurs. We're really seeding pilots, collaboration with other universities.

So the basic orbit itself is our one-stop shop for MIT entrepreneurship. So we pull together our key application dates. Today, if you were to sign up, you'd either be in an entrepreneurial cohort or a cohort for your schools.

But we're going to begin pulling the cohorts together. We didn't want to muddle them. So it's very clean to have everyone in Northeastern together. But at the same time, the Northeastern folks want to meet the MIT folks. So we'll begin doing that.

And we wanted to kind of up our game, including labeling everybody at MIT with MIT. That was done this weekend, because when we put everyone together, since we started MIT, we didn't have that label.

We post startup opportunities. If you have a startup, you can post your opportunity. So these are all startups.

We have events. So we scrape primarily today the MIT calendars. I think we'll start to go wider. But you can also send an event in.

resources at MIT. So it's been MIT-focused. We're going to slowly kind of open it up.

And the Jetpacks is the main thing that people are joining to get it. Over time, we want to use up our game on meeting co-founders, meeting people that can help you, meeting mentors, and do it in a useful way, not just a jumble.

AI Jetpack

So that's Orbit, and then I'm going to start with the AI Jetpack. production account, I'm doing a lot of these demos. So I've got lots of different ideas. But when you first log in yourself, you'll have this.

And 1what we decided to automate first was our core entrepreneurship framework, so discipline entrepreneurship and startup tactics.

And before I get too deep, who has seen this demo before? You have, yeah, a couple. OK, so not too bad, three.

Bill Ouellette is the head of our center. And what he did was pull all the best practices together.

So I've done a number of startups, Zipcar, RunKeeper, just did a health care one. And these fundamentals hold together. But I used to use lean startup design thinking as I was kind of bringing the teams up to speed.

And what Bill did is pulled them all together into one framework. And then that was about 10 years ago that we just reissued the book.

And then Startup Tactics from Paul Cheek is as we're working with MIT startups, what were the questions people were asking?

What do you do about legal? What do you do about goals? What do you do about market research?

So startup tactics is a companion that helps you get up and running.

And again, it's true for anything. So I think I will do screens, because there's a couple of you here. So screens, extend.

What's the URL for you? The URL is Screen's S-K-R-E-E-N-S dot com. S-K-R-E-E-N-S? Yeah. Dot com to AI enable video displays.

Sounds good. Okay. So that's what they're doing. If you're interested in it, they're in the back.

So you submit your idea. Oh, and the thing I wanted to show is Bill pulled it together, and what he did was try to normalize it. So not use specialized language, keep a sense of humor.

Because if you're doing a startup or if you're doing AI-enabling screens, you've got to have a sense of humor. You don't know exactly what you should do, so you've got to plunge in with all your might, but keep your sense of humor. So I really like that about the framework.

Entrepreneurship Framework

So I put my idea in and click it. And then these are the 24 steps. So this is the core framework at MIT that MIT uses. And it's really built off of the other frameworks.

And what I find most interesting is We're getting people from the labs. Like, I did a seminar like this with 40 postdocs. They are deep into the technology. But with all those people, we start them with, who's your customer?

So that's the whole game. So whatever you're doing, you have an idea, A note. Where are the A note? Yeah. A note, you're going to really need to focus on who you're going to be great for, because good is not good enough.

I use an analogy at RunKeeper. We had grown to about 10 million users, and we were getting encouraged by our investors to see how fast we could grow. And it was a kind of heyday of the mobile industry. So we were adding additional sports, but we added biking and walking and other areas of fitness.

And we felt like we had kind of lost our way. We were kind of gone from a pretty good running app to a OK everything app. And Apple didn't like us doing that. So we wanted to refocus on running.

And we thought, OK, focus on running. And Jason said, OK, what do we do next? Product people said, I was head of technology, said, we need to go talk to our customers. And he couldn't believe it.

We have 10 million customers. We need to talk to our customers. But to win, we had to figure out which customers.

So a disproportionate number of RunKeeper employees were doing marathons, but our users were beginning to intermediate runners. If you're gonna carry a phone, and it was really about, so their problems were different, and our marathon training programs made them feel bad.

You'd start and fail. So for beginning to intermediate, it was fitting it in your day and thinking about those. So we became great for beginning to intermediate runners and grew to 50 million users and sold to ASIC.

Target Market

So I've been where you're kind of exploring, but ultimately you want to focus on a set of users and build from there. So that's really what this builds from. So the same thing applies for screens. For screens, we want to think about what are your target markets.

And you probably already have some areas you work in. But this helps, whatever your startup idea is, helps you think about it. And my friends and I have always had different ideas. I think everybody here, if you're in Boston, you're kicking around ideas.

And you just sort of kick them around with each other. And I became really good at saying, no, that's not a good idea to a lot of people, which didn't make me feel good. Now I can say, segment your market.

And I couldn't say, segment your market. That's no fun. I have a startup. I want to do all this work.

Um, so having this available really, um, is a game changer because this is a means to an end. It's not the end. And the critical thing is Bill and the team have worked on this criteria. So when I'm doing market segmentation, I could have God knows what's every startup I've done has been about market segmentation.

Um, but here, uh, it's, it's organized, uh, and you can keep it. So, um,

Do those market segments look reasonable? No. Not for a year after? No. OK.

What do you sell to? So we sell platforms to service providers like cable and telco who are sending video down to the home.

So are you on the, so who, I don't, what's your product then? So our product is a personal AI agent to help personalize your video so you can put up multiple screens.

Ah, so it's not the display itself. Display meaning that, yeah. Ah, ableize what is displayed by cable and other providers? Yep.

On the big screen. So what we like to say is this is like a group of freshmen and sophomore working on your behalf. But you're ultimately driving it. And so update idea is a steering wheel. And so that was a good example.

And for the purposes of time, we may need to work it. But that shows that you need to keep iterating until it's right. And you can actually put the segments, if you disagree with them, you can put the segments in yourself. So what we're really emphasizing is you should be iterating and getting narrower and narrower.

Product Segmentation

And if you come out with a new product, you want to figure out kind of what's your initial launch target. which you all are mature enough, or you've been up and running. And then we keep the versions of it. And so you can name the versions as you iterate.

So if you're early on, you want to go check the market. And you might even put in multiple ideas. I want to do this for cable providers and make that an idea in itself. And I want to do it for event management. And kind of work your way through it.

But in the interest of time, we'll go through and then start thinking about, if those are the right segments, how do I pick which one?

And I may have some intuition. So our health care provider, there are new standards for health care data, FHIR, FHIR.

And we could sell to hospitals. We could sell to insurance companies. We could sell to digital health companies and pharmaceutical.

And we had early customers in all of them. And we wanted to do digital health. Ricky and I really wanted to do digital health.

even though we knew that wasn't necessarily where the money was. But it turned out, as we did these, all the money was in health insurance for us.

Epic did all the standards work for the hospitals. And the health insurance had old Kluge providers. And 20 of us could go and win all those contracts. So we pivoted the whole company.

And then Ricky and I agreed, after we raised our C round, we would start looking at the next thing. So I went to MIT. He started genhealth.ai.

But we knew we were getting into. So we picked one. If we had just stayed in digital health, where we liked, the company wouldn't have been successful.

So you've got to figure out kind of where you are. and decide where it fits with you. So is it well-funded? Can you sell to it?

So it might be well-funded, but it's impossible to sell to. Do they have a reason to buy?

Market Focus

So again, it's kind of helping you think through your market. And it's an assistant.

So you may be up and running already. Wes, you're up and running. It doesn't hurt to have this. And I would have used it all the time at any of my startups.

And then I like to show, is it consistent? How does a computer know? This is what I have to answer if I'm doing a startup. And if you're going to do a startup or you're going to do a new innovation at your company, because none of us are here because we want just status quo or believe that status quo is going to be OK.

So we want to lean in. So you could put this in.

Who's here at a corporation that's willing to talk about it? Anybody? Did you raise your hand?

Yeah, you're at a corporation. And what's the company? Cloud File Storage. And do you have any new innovations in Cloud File Storage?

Yeah. Yeah, so the same thing applies. You have to decide if it's your current target market, your new one, kind of where are you with that? And are you passionate about it?

And I've reached different points, I think all of us, where my passion, I realized, was lower. So it was kind of like finish this project and figure out what's next. So inside the company, because you're obviously valuable, or elsewhere. So it kind of guides you through the process.

can guide who's your profile. You want to get really deep.

What's the beachhead total addressable market? So for people that are on the engineering side, it's kind of a recipe book for thinking about it, which is super valuable. So the engineers.

The best teams here at MIT are a combination of engineers and the business side. But you want both sides now. With AI, the non-engineers can be doing coding and should.

You know tools like lovable and such you can do your own your own software And then the other way around the engineers should have a sense of it and this kind of guides it So if you don't know what the how to compute total addressable market you could ask chat GPT it knows but this keeps in one place and what's really valuable is you can think through what are the really important annualized revenue per year is really important. Number of end users is really important. And you can kind of work the top levels.

So we had a startup doing nutritional supplements for pets. And they knew the real numbers. And they were kind of ballpark. But what was more important was getting the real numbers in.

So you want to look at these numbers, and the answer isn't orbit told me. The answer is I did all this research. And so you work with this in conjunction with Perplexity, which does a great job.

We're using OpenAI Omni Mini because it's fast and cheap. But you can use it with all the other models.

Summary and Export

So that's the basic discipline entrepreneurship. You can export your summary of what you did into a PDF and drop it into the other models. I do that a lot, depending on how much information you've shared.

Startup Tactics

And then the other framework is startup tactics.

So this is helping you get up and going. So you all are in the back, lazing around a 7, and you're watching me.

So it's important that you think through what your goals, because Tomorrow, you're going to have to be in the office. So just thinking through what are the goals.

And these may or may not be the goals. It's certainly not going to be a week for this. We left it kind of intentional.

But the point is you do want to set goals, and it kind of sets it in place. And you can interact with it. So it's like, please change. Go one, two, four weeks. So you can kind of interact with it.

Let's see if it took it. Yeah, sometimes you have to tell it again.

And then things like market research, which you already have products. Yeah. So you're already kind of in contact.

A lot of startups may or may not be. But just getting you to think about what to ask for. So it's kind of kickstart.

So that's the basic framework. We're getting ready to add climate and energy ventures.

Future Enhancements

And then this produces a lot of output, so we're also adding canvas approach, where it's forest and the trees, which is also important.

So if you go through all the steps, you might get 40 pages of output information, not insights necessarily. And the canvas gives you kind of the high level view.

adding those kinds of things, and then also beginning to, now that we've crossed 10,000 users, we want to start making it more useful to each other. I know people in healthcare doing startups in women's health and various, and we want to try to make that, support that through the user profiles we're getting. So you'll see more of that.

Q&A Session

So that's it.

I guess we have a few minutes for questions. But any questions? I do have a question.

In terms of the kind of specific business idea that someone is trying to explore, particularly if that business idea is more related to some of the more in-depth scientific knowledge. For example, I want to do a project based upon a particular disease, See whether that makes sense. I'm wondering, this system, in that aspect, how could that work?

So the question was, if you've got a really deep scientific idea, how it would work with it. What we've seen thus far is, because OpenAI in particular, but all of them are scraping the web, Anything that's public, it's aware of. So we're doing a tech transfer program here at MIT and also Northwestern.

We did one with Raman microscopy, and a woman's parallelizing it. It kind of knows the underlying part. But it doesn't necessarily know all the business side. The insights are with the customer.

But from the technology, it has a basic understanding. But you really have to steer it. So it'll often get you on the right track.

The other thing we're adding in life sciences and health care, there are certain regulatory things like climate as well. Each of those, we're going to add modules to kind of help you. What are the considerations to get for medical devices? We want to take that knowledge and make it more readily available.

And MIT's mission is to share best practices. So our alumni are really pushing us. Basically, they've said to me, if this is such a good tool, This is two years ago. Why are you keeping it MIT?

And first, we had to make it a tool worth sharing, and then part of the mission to share it. Great.

I saw that you had adjusted the goal, the timelines of the goals, at least for that first one. Yeah, yeah. Is it possible to adjust all of them kind of accordingly? Like, for instance, if you change the timeline of the first goal, can you set it to the other one?

You can instruct it to do that. So you might say, change first goal and then give it instructions. You're really talking to a chat bot. What we've done is put the criteria and break it down into steps.

So it does a great job. I think this will be the last question, then we'll get to the other speakers.

So if you're like an academic lab and use a cache of un-public research, can you make that available to it to see? Really good question. Today, we haven't allowed you to upload your own information. We've had two requests.

That's one of them, is upload your own. And I'll explain where the data goes. The other one is the output. So you can do like Y combinator outputs and things. And so those are both things we're going to work on next.

You also would have, I think for you, if you had a lab, this is an academic platform, but we use Stack AI, which is shown here, as kind of our middleware. So the data's indexed locally. So I think as long as you're comfortable with an academic platform, I think you could share it. But we haven't added that facility yet.

All right, I'll be around after, and we'll move to the next. Thank you. So another round of applause, please.