On this week's episode of Inside Outside Innovation, we sit down to talk about new mental models for working with AI, the similarities between startups and middle school dating, and lessons learned from the robot Olympics. Let's get started.

Inside Outside Innovation is the podcast to help innovation leaders navigate what's next. Each week, we'll give you a front row seat into what it takes to grow and thrive in a world of hyper uncertainty and accelerating change. Join me, Brian Ardinger and Miles Zero’s Robyn Bolton as we discuss the latest tools, tactics, and trends for creating innovations with impact. Let's get started.

Interview Transcript with Brian Ardinger and Robyn Bolton

[00:00:40] Brian Ardinger: Welcome to another episode of Inside Outside Innovation. I'm your host, Brian Ardinger, and I have my co-host, Robyn Bolton. Welcome, Robyn.

[00:00:50] Robyn Bolton: Thank you. Great to be here as always. 

[00:00:52] Brian Ardinger: We are in a brand-new year 2026. Who would've thought? Exciting to start the year with you. Appreciate you coming on board. 

[00:00:58] Robyn Bolton: Yep. High point of the year so far. 

[00:01:00] Brian Ardinger: We've got a lot of things going on on the plate. Anything you want to talk about? 

[00:01:04] Robyn Bolton: Couple of new things I mentioned earlier, one of our stories from last year is back in the news, the Samsung AI fridge just voted worst in show at CES this year. People finally caught on to the fact that we may be overcomplicating the refrigerator.

Thought that was a funny callback, and I got to admit, I feel like you called it Brian and I echoed it of like we've gone too far. So, personally, professionally in my space, starting to do a lot more work in uncertainty and helping people figure out how to make decisions without the data they want or need, and how to help teams move through a world that is getting only more and more uncertain every day. So, it's exciting. 

[00:01:51] Brian Ardinger: Saw your newsletter this last week, and yeah, the new positioning, or you're talking about how it's not just about innovation, it's more about how do you deal with the fact that nothing that you expected to happen is going to happen, and how do you deal in probability and uncertainty. 

[00:02:06] Robyn Bolton: Great for innovators, because that's one thing that as the innovators, whether you're a startup founder, a consultant, a corporate innovator, every day you're dealing with uncertainty and trying to figure out how to move forward. Even though we've always called this innovation, it has much broader application these days. 

[00:02:23] Brian Ardinger: Absolutely. Let's get right into it.

We've got a couple of different articles we've been reading over the holiday season. The first article we want to talk about is called Six Mental Models for Working With AI. It's from Azeem Azhar. He's got a great Substack newsletter out there that publishes pretty much almost daily, I think it comes out.

 

But he was talking about the way he's been looking at AI over the past year and trying to come up with different models that are making it more effective. All these AI tools are brand new and that, and people are trying to figure out what works, what doesn't work, how to use them better, and I think it's sometimes interesting to take other people's perspectives and what has worked for them and discuss that.

So, in his article, he goes over a couple of different frameworks that he uses when he is either trying to understand better how to use a tool. One of the ones I was going to talk about is, he calls it the 50 x reframe, and he says, when he is dealing with a particular problem and trying to understand like, how can I automate it, how can I make it better, how can I make it faster and that he asked the question, what would I do if I had 50 people working on this problem. And asked the AI basically to help him think through the framework. Or if you know 50 people were working on this particular project, how could you automate it or what would change if you had 50 people to be able to dig into a particular area.

So, I thought that was a very interesting framework to think about it. And we oftentimes get constrained in like it's just me or just my team. But what if you just flipped the framework and said, what if I had 50 people on my team to work on it? How would that change what I'm doing? 

[00:03:46] Robyn Bolton: I loved that one. I mean that one, it's the first one listed in the article. And I'll admit, I started reading the article. It's a big skeptical when I started reading it because you know, his first sentence is the question of whether AI is good enough for serious knowledge work has been answered. And I was like. Yes, it's been answered. It's not. And then I kept reading. I'm like, oh, he has a different answer.

The 50 x reframe just stopped me in my tracks, was like, that's genius of shifting from how do I as one person do this better with AI's help to completely rethinking. I also loved his second idea, which was adversarial synthesis, which is basically to have multiple LLMs, Claude Chat, GPT, Gemini, working on the same problems, responding to the same prompts, and then going back and forth.

And that's something I actually have done, and it consistently results in a much, much better final project. Each LLM is tweaked for certain things, and the combination actually gets you to a much better answer. So that was another great, great tip he had. 

[00:04:54] Brian Ardinger: Yeah, and I think that's evolved. You know, I think early days when people were working with a lot of these LLMs, especially like in the writing sense, what they were asking to write a letter or write some copy and that I heard a lot of people using the different models that come up with different types of content. The same prompt with different types of content and then picking the best one. His framework is much more focused on how do you actually make them argue against each other? 

[00:05:16] Robyn Bolton: Yeah. 

[00:05:16] Brian Ardinger: To see which one, and that arguing back and forth between the different LLMs actually strengthens the argument and strengthens the output from it. So, something to keep in mind in 2026. Another framework to consider. 

[00:05:28] Robyn Bolton: Yeah. Get the bots arguing with each other. Fun for everybody. 

[00:05:33] Brian Ardinger: The second article that we want to talk about today is, I love this title. It's called Founders Can't Sell for the Same Reason Middle Schoolers Can't Flirt by Aaron Denon. Aaron's a professor at Duke. Teaches entrepreneurship who's constantly coming out with some great stuff around early-stage entrepreneurship.

And his article talks about the fact that what if the same fear that ruins dating lives is ruining your startup? And he talks a lot about the fact that startups is very much like dating, where at the early stages you're trying to gauge the situation, but you're not necessarily straightforward. You're not asking for the sale, you're not asking the person out, especially if you're a middle schooler, you're trying to dance around the edges.

I thought that was an interesting framework to think about when you are creating anything new, whether it's a startup or corporate innovation type of project, how direct are you about asking for what you need and really defining what you have to offer?

[00:06:28] Robyn Bolton: The positioning he had made me laugh, and it's so true. I am 100% guilty of not asking for the sale. Because, you know, I don't want to come off as salesy and pushy and founders are that way. You know, a lot of people have that hangup. We just, you know, when it comes to sales, we have the used car salesman, like, oh, I don't want to be that guy.

But reframing it as, oh no, it's middle school dating. And like you actually do at some point have to go ask your crush to the dance. They're just not going to magically sense it as like, oh yeah. That's the only way things are going to move forward, that we're going to get to an answer, whether it's yes or no, just getting a definite answer.

I thought it was a great reminder. I thought it was great positioning and actually going to, I still have a little middle schooler kind of angst and side of me, but like a call to get over that and ask for the sale.

[00:07:22] Brian Ardinger: I was thinking about this particular analogy as it applies to corporate innovation, and one of the things I was thinking about is a lot of these corporate innovation projects, you have a team with a new idea and the way they pitch it or whatever, is I have to be the prom king or queen.

I have to be going after the senior, and the project has got to be perfect. When a lot of times, you're never going to date the prom queen. No. But be a little bit more direct and understand what you're actually asking for and what you actually have to offer. 

[00:07:48] Robyn Bolton: A lot of times I also see, especially with corporate innovators, is they have a great idea, but kind of no plan, no rationale. They're just like, hey, I think we should do this. Which is kind of the equivalent of the asking your crush out on a date and then being like, I don't know. I hope everything comes together. Maybe I'll see you at the dance. So, you know, there's also that other thing. Innovation really should not be this similar to middle school dating, but now I'm starting to think it is.

[00:08:12] Brian Ardinger: It's all about uncertainty. 

[00:08:13] Robyn Bolton: All about uncertainty. 

[00:08:15] Brian Ardinger: Well, the last one we want to talk about today is from Physical Intelligence,and the name of the article is Moravec's Paradox and the Robot Olympics. You want to start with that one? 

[00:08:25] Robyn Bolton: So, I highly, highly recommend this one, especially for the videos because the idea behind it is that we've had ai. We've had computers doing amazing, brilliant things since the mid-nineties, you know, beating a world chess champion, beating a world champion in Go, which is a game. And yet there are some things that we think of as very, very simple that robots can't do. And this is that paradox of robots can do incredibly brilliant things but also can't do simple things.

This article contains videos of robots that this company has programmed around this kind of robot Olympics of different activities. So, like turning a sock inside out or peeling an orange. One of the ones that struck me, because every morning I have peanut butter toast for breakfast. So, one of the challenges was making a peanut butter sandwich.

And how incredibly difficult it is to program a robot to do any of those things. And that just kind of gap between the physical activities we do that we don't even think about but then translating that into a robotic environment. 

[00:09:40] Brian Ardinger: For me, one of the things that stood out in the article is again, I think we oftentimes think, well, hey, this LLM is so powerful and does amazing things or whatever. Why can't a robot do that? Or you know, because it seems so simple, but he was talking about the fact that it's almost a lack of data. Like we have data for every single book out there. We have data for every way to think about particular things, but we don't necessarily have data that a computer can use to know exactly what the nuance is for putting a key in a lock and turning it.

I mean, how much exact pressure do you use? Where do you turn the lock, all those kind of basic things that we as humans collect the data and process it. So, until we can get particular data, the computers, algorithms, the LLMs and that may not be able to actually replicate some of the things that we do from a human perspective. So, hey, we've got that going for us. 

[00:10:25] Robyn Bolton: Yeah, we can make peanut butter sandwiches. Take that robots. There is some really interesting work. It's still in universities, still kind of very, very much bench work around using videos as a source of data to train robots. So, taking a video of somebody peeling an orange, making a peanut butter sandwich, and using that as a data source for the robot to learn from.

So, there's some progress on it. Maybe one day we'll lose our superiority on orange peeling as well, but just really interesting to see the gap and to see, and something that feels so obvious to us is actually a lack of data problem for a robot. 

[00:11:04] Brian Ardinger: And the final thing I'll add to this, is the fact that after watching these videos, even though the robots are struggling in that it's pretty amazing that a machine can do some of this kind of stuff that was not possible mere weeks, months, years before.

[00:11:17] Robyn Bolton: Yeah, because they do eventually accomplish everything. It just takes a little bit longer. 

[00:11:21] Brian Ardinger: Excellent. Well, that ends up our articles today, and we're going to transition to our tactics to try. I've got one that I'm going to actually call you out, Robyn, in your last newsletter you have a tip in there and it's talking about next time you're working with somebody and they offer a point of view or opinion or anecdote one of the things that you can do to gauge the impact of that particular conversation is ask the question, how confident are you that you're right? 

And then ask, you know, what are you willing to bet? Are you willing to bet your annual salary? Are you willing to bet at dinner at a Michelin restaurant or a cup of coffee? And so, use that particular tactic to get a gauge for the amount of effort or insight that the person has and that can help make better decisions. 

[00:12:03] Robyn Bolton: Yes, so it's a tactic, a trick that I use all the time, especially when I'm working with teams to set up their plan for experimentation and kind of what assumptions are most critical. And I will never stop being amazed by how confident people are in their assumptions until you say, are you willing to bet your salary?

And then it's like, well, not my salary. I'm like, how about, you know, dinner to Michelin starred restaurant? And then it's like, some people will say yes, and some people are like, how about Applebee's? And I'm like, okay. Not as confident as we thought. Which isn't a bad thing. You just want to know. What are we looking at so that we're designing the right test? 

[00:12:45] Brian Ardinger: Well, and I think that's important too because I think a lot of times, especially in corporate environments, that people, they want to plan, they want to say, okay, is this new product going to get us $20 million in the next year? And a lot of times you don't know. So being able to say you don't know or to ask those particular clarifying questions, I think can go a long way to keeping the open transparency of what is actually being developed or looked at. 

[00:13:05] Robyn Bolton: Yeah, and much better plan than saying Absolutely yes. And then when you generate $10,000, losing your job.

[00:13:12] Brian Ardinger: Alright, well the only other thing I want to talk about is the fact that, hey, the IO2026 Summit is up and going. We are going to be announcing our first set of speakers probably in the next couple weeks. But in the interim, one of the great things about this conference is we are looking for the best and brightest startups, side hustles, corporate innovation initiatives, prototypes, anything that you've got going on out there.

So, if you are a startup out there listening, if you're a corporate innovator who wants to kind of showcase a project that you're working on, we have something called the Gallery Innovation. Go to the IO2026.com and go to the Gallery of Innovation and you can apply. It's free to apply. We ask a couple questions like, what's the project? Gimme a logo or something that I can put on a virtual showcase.

From that, we're going to try to mass a large number of interesting things that are being built. Out there. And then we're going to be giving away tickets to a certain number of those folks that have applied for the gallery. And then a certain number of those folks will also be given time on stage at this event to talk about what they're building, what they've learned, and give us a real insights and you know, of what is actually happening in the world of innovation, whether it's a startup or a corporate innovation initiative. 

So, check that out. The Gallery of Innovation at IO2026.com.

[00:14:21] Robyn Bolton: Definitely check it out. I'm excited to see the galleries. Definitely, definitely apply people. 

[00:14:26] Brian Ardinger: Fantastic. Well, that concludes another episode of Inside Outside Innovation. Thanks for coming out. We'll see you next time. 

[00:14:33] Robyn Bolton: See you next time.

[00:14:37] Brian Ardinger: That's it for another episode of Inside Outside Innovation. Today's episode was produced and engineered by Susan Stibal. If you want to learn more about our teams, our content, our services, check out insideoutside.io or if you want to connect with Robyn Bolton, go to MileZero.io, and until next time, go out and innovate.