Why AI Scaffolding Matters More than Use Cases

Well, one of the things that we talked about before that I think would be interesting to get into today is this idea that the myth of this big sexy AI use case out in the world, maybe at NASA, and, yeah, right? They're both kind of a myth. And I think when people think of NASA and AI, their minds might go to some of the more space-bound

I both. I think both are a myth. Yeah.

implementations of AI, but what we were talking about before is like, you know, within an organization like NASA and any other organization, really like the best use cases for AI are often like quite boring. They're the things that help you lift tedium out of the workday. How did that kind of manifest in your time at NASA in terms of like preparing for AI readiness?

You mean like how did tedium manifest? Because I think we got that story already. It's like no slack. No way to communicate with anyone else. Isolate someone.

Or how did the identification of TD? Yeah.

no, no, like the real, the real answer is that the, the, the like emergence of AI and specifically, I, I just came up with the phrase of saying like consumer grade off the shelf AI. so chat GPT, whatever. I don't, I don't know. I remember what metas is called, but the one that's like built into like Instagram or whatever, like llama or whatever, Gemini.

Right.

I like that consumer grade. That's the right thing. Yeah.

Yeah, well, that was it. It was important to call that out because NASA, like, lots of places have had AI. Like we worked with people, like the, the, the deputy chief AI officer I worked with. She worked at IBM on Watson, like 10 years ago, like AI, we watched Watson like go up against Ken Jennings on Jeopardy. So AI is not like a new thing. And especially like NASA, if, if listeners, whatever, we got to put this in the notes, want to go read like Ames research center out in Mountain View.

Mm-hmm.

which is right next to the Googleplex has been doing AI for like 30 years and had like some of the first white papers on it, like literally in like the 1980s. And Ames is where they like assembled the curiosity Rover and like did that. And when they go into Mars and has a huge, huge, huge, wind tunnels. And so AI and machine learning and like NLP that's like, they're already doing that. And we have at NASA has like some of like the best world-class experts already working there.

Yeah.

That is not what we're talking about. That is not what AI is. We need like a different, we need like a different word for that.

No. I got a funny story. were in our early days as a company. We were our conference that we were invited to was the AWS space conference and the booth next to us was a moon rover you could purchase and then there was us next to that. Like that just that just makes your point. Like people are thinking

Mm-hmm.

Back then, this was some sort of alien technology.

Yeah. Yeah. Like there's, there's a huge difference. And there's also like, how many AI experts are there in any big science or technology and organization, like half a dozen, a dozen, versus I think there was something like 20,000 civil servants and then contractors and, partners. And so that consumer grade AI is like the chat GPT, the mid journey, the stuff that you see, um, on TV, right. And there's, commercials for, for Gemini, like on Hulu you can watch.

Yeah.

anything.

And so when you go and want to like, see where is this organization at and how are we going to take what's coming in from the outside world? You're really looking at what started as data landscape, right? And this is where we're like, we're going to like have to get kind of boring for a second before we make it more interesting is that AI it's fun to generate stuff all day. Like I make stuff with the like Dolly and like image generation tools and use chat to be all day, but I'm doing it.

you

extemporaneously, just, Hey, you're come up with a whatever. it does it. If you're trying to introduce this into an organization with existing data, petabytes of data spanning decades, like that was the start was what do you do with your, your terrain of all your data? Just think of any big company. And they're like, we want to start integrating this UI. It's like, it's not going to do you a lot of good. You can make funny images and you can like, like synthesize things by inputting text, but what are you going to do? And you want to like.

Mm-hmm.

Bye.

interrogate mountains and mountains and mountains of data with the AI and they're disconnected and they're silos. And I mean, you got stuff on like honest to God, reel to reel tape computers, like program with like Fortran and you're trying to like ingest it's like that, that you can't start there. So when you're looking at implementing that type of capability where you want your internal work to be as easy as opening up chat GPT on your phone, you have so much work to do of the organization is not even close to being ready.

Mm-hmm.

So

Mm.

to implement this as anything other than a toy. And so when we talk about the myth of the big sexy AI use case, I can tell you some of those sending an autonomous robot to an asteroid to look for precious metals and have an AI that does it. That's a big sexy use case. That is not what we're talking about. What we are talking about is what is the, day to day like usage of running your business or organization with this new capability to then start working towards.

Right.

You

use cases, and this is where I think there's a split between AI is the thing you want to sell in the product and the thing you want to integrate to like add value or put it on the robot or the set, whatever. And then AI as a tool set to enhance your existing capabilities and processes and workflows. And I think if we're talking like digital transformation or organizational maturity, you kind of have to put up the scaffolding first.

Mm-hmm.

and then climb on that to build the thing to deploy.

Hmm.

Yeah. And it's, think there's like a shift. There's something interesting about the NASA scenario, because if you go back in time and you have some opinions on this, but you go back in time and you look at, there was a space race, right? And that's, that drove a lot of innovation and a lot of, a lot of funding and activity. and a race to what? Like,

Mm-hmm.

who knows a race to like be ahead, right? To be ahead of someone else. And the objective was just to be first, right? At something. And we're kind of seeing that now, like with AI, you have people adopting in a race, but they don't even know a race for what? To adopt it faster. And I think it makes sense. Like adoption makes sense.

Mm-hmm.

And there is complacency in not being competitive as I think you've talked about in past conversations. Organizations can get complacent if they don't have that competitive component, but at the same time, it also still can be complacency into itself. Like racing towards what? And your point is like racing towards efficiency, right? means that you gotta be boring.

Kind of, yeah, like there's, I thought of this, I mean, I really did think of this yesterday as I was thinking about what we're going to talk about. And the example, which I think is a lot easier to understand. I always come back to what my dad and his brother, my uncle and grandfather and family does. They built houses like with their hands, like a very small crew. and so they would, they would do the whole, they would do the whole like.

Right.

Like talking like software development, the home building life cycle of home development. I'm so, mean, like from like digging footings to pouring concrete framing, everything to finish work, selling it. It was all just like a little small family business. So I, I know how to build a house from scratch. And what I'm looking at how AI works, I think, like, I just had this house in Maine that was 200 ish 150 years old. I'd put a new roof on. And so I hire the roofers and they come look at it.

Mm-hmm.

And one of the first things they do is a part of the, the bid is they say they're going to buy like scaffolding. And I was like, well, you don't, you don't have that. And he's like, it's, it's hard for like, every house is like, kind of like custom and we like attach it to the house. And we're going to be here for like two months. And, and so we're going to like screw it in or like size it it's going to be a part of it. So they go and buy it they attach it and then they get these huge tarps and they have to strip the shingles and they created this entire like.

Thanks

Hahaha!

Okay.

project of the scaffolding on the house and the place to like take all the old shingles and the scrape them off and the nails before they even got to doing the roofing. And now the roofing was a big sexy use case. was standing, standing seam, steel roofing. And so it was like size to the whole house. Like they would, it came off a big truck and they cut it it went on and like, it doesn't leak. lasts off for like a hundred years. So it was really high tech stuff, but you can't, you can't even put that on.

Wow.

until you spend like a week building the scaffolding and getting the harnesses and then like roof jacks or like these things you put on. you can, cause the roof is a, a 12, 12 pitch, which means it's 45 degrees. You can't walk on that. So you have to install steps and then they had to have like OSHA, I think like, harnesses and clips. So then they had to go and like tie off to the chimney and had like repelling gear. like, it took like a week just to get ready to get start and That's like the AI scaffolding. It's like, are trying to just climb up on these roofs with like our most high tech, you know, pneumatic nail gun. And we've got all these tools and stuff, and we haven't clipped off to our like belayer gear. We don't have the scaffolding set up. We don't have the tarps and the dumpsters to catch all the debris. We just want to get up there and be like, it's like, that's what everyone is doing. That is like the state of AI and transformation. The scaffolding is what we discovered and AI readiness is because when we pulled the organization,

Yes. Hahaha

Most of the people, what they want is simple day to day, help me do the job. I'm already doing use cases because the mental model of you're telling me I got to get up on that roof and use this amazing new nail gun. I'm like, can't, how am I supposed to get up there? don't even have ladders. What are you talking about? I was like,

Right, right, and then you show up.

Is that scaffolding then like a combination almost of cultural change and technology? it kind of a blend of those things?

I think. I think it's, I think this is, this is a kind of a, it's a weird, it's a weird moment in tech history where the technology occurred so fast. There was no way to be ready or even understand what would happen. There was no way to know that almost overnight you would go from seeing like funny memes of like GPT-3 trying to write Hallmark Christmas movie scripts. And it would try to do it and they were, they were hilarious. They were awful. Like it didn't understand it all what to do. I just remember one I read was about a script about a woman who was opening up a snow globe shop and like a man comes in and I think the meat cube and he wants to know if the snow globes are filled with more Christmas juice. It's like, what are you talking about? And then you like fast forward and now, within the course of like 18 months.

I'm

You know, GPT 4.5 or like, or four or whatever could totally nail it. And it could write a hallmark Christmas movie and you'd never know the difference. Like you can't like, like it took me over a year to get an ATO to install Figma on my computer. That's an authorize authorization to operate Figma. took a year to get that approved and installed and paid for. And you're telling me we're supposed to like, all of a sudden start using this extremely dangerous, powerful thing called like consumer grade. AI chat, GPT, Gemini, co-pilot, whatever it's like, it's impossible. And so it's part cultural, it's part like organizational calcification. And then in government, it's rules and regulations and safety and, you know, caution because you, you know, well, I mean, you used to not be able to just get in and do whatever you want. Very difficult.

Yeah.

Right.

So like most of the scaffolding in the world, it's somewhat unsightly, but it does the trick if you put it up right.

Yeah, like it's, it's, it's there to not only allow, you know, the action to take place, but it's also a safety, you know, mechanism to, you know, give the people who have to do the action a place to, you know, exist and like nowhere to go and like a consistent, you know, literal framework to navigate on. so that's where, when you ask people what they want to do, there are some of those.

me.

sexier use cases about like really cool science fiction stuff. But most of them are about help me do what I want to do now because we're standing on the ground. We're looking up where we need to go. And I can't get up there to do the cool thing until we do the boring thing, which is give me a way to get up there.

Yeah. And is it boring? Like we were talking IDEO, like, I don't know, now it's got to be a year or so. And they even more, they created something called Lotbot. I've mentioned it number of times on the show. And Lotbot was simple. was a bunch of cameras pointed at a parking lot and they would just, they called these cameras, cargo oils. And they, they would just tell you if there was a spot available or not. and, the interesting thing they did side note is each camera could talk in natural language. So think MCP. and so they could talk to each other. Hey, do you have a spot? don't know. Do you have a spot? but, but that was boring, right? Like, is, is there a spot available? But then we were just, you know, kind of thinking through the future of car-goyls

Mm-hmm.

And Lottbot and we're like, well, Lottbot could next secure you a desk. Like, Hey, is there a desk? Um, it could then coordinate when you come into the office with other team members, it could start to do other project management work for you. Um, and you could just see Lottbot slowly. mean, the idea was like, eventually it's running the company, right? It's, and we talked about like AI starting in the mail room, like just like those days and it's.

Mm-hmm.

You

Mm-hmm.

It's not unexciting, like when you put it in that respect, it's just maybe less dopaminergic.

Yeah, it's, is. We need, we need to come up with like a new viral, like graphic or something to replace like the two ketchup bottles of UX and UI with like AI readiness and like AI practicality and like show how most of, most of what I, what I think, and I have an example and I can, I can, I've been gone from into it for a number of years. I worked on all the QuickBooks products, but I.

you

I use QuickBooks to do my little tiny side business books. I just used it to send all the stuff to my account a couple of weeks ago. And there's just AI in everything, but it just, it just feels, it just feels kind of like added on and not well thought out. Yeah. And it's not yet, it's not integrated into the use case. It's just kind of like, it's replacing something.

you

What time is

No, it's not replacing something. It's just, it's enhancing something that already exists, which is a non AI use case and capability. Now kind of, you know, sped up to, add the capability, but it doesn't really add to the use case. It doesn't really change the use case. It doesn't really change how we're using it. And it doesn't seem like it is really easy to understand what value is it bringing other than just existing. And in fact, it could be even confusing because it's mixing two paradigms, which is a, you know, What, what, is it like a touring complete like assistant in to side the software product that should be able to answer questions in a human way and infer things and understand things that do things for me. But in the same time, I kind of have to, I kind of have to hold its hand. It reminds me of, Robin Williams and bicentennial man when he was the Android babysitter, he like just doesn't get stuff. Like just doesn't understand what's going on or like, remember like the star check next generation episodes where data would.

Yeah.

would get his mind erased and he's still data, but he like doesn't know. And he's just like this baby. it's this info, like that's what AI is like in these products. It's like, it's not actually helpful. We're helping AI. It's not helping us. And so I'm like, why did we invest so much time into putting it into this product instead of putting it into the tooling and the behind the scenes of the designers and the engineers and like product managers who are putting it together and use AI to help that process come up with a better way for AI.

Yeah.

Hahaha

No!

Mm-hmm.

Right.

Right.

to be implemented. And that's where you would start like a recursion loop of using AI to help figure out where it should go. Like robots, robots building other robots where it's like you send one robot manufacturing, you know, facility to the moon. And then once it's there, it then starts to make other things or like, that's, that's one of the things that is important about launching things to other planets. We want to do it from the moon because

Yeah.

Yeah.

Mm-hmm.

The gravity is one sixth, then there's no atmosphere. So it's a lot easier to do it from there, but making the stuff on earth and getting it to the moon and then using it there, that's really intensive. So what if you had just the basics, like think of like a 3d printer that could take materials that are already there and start manufacturing stuff. And then that manufacturing stuff, manufactures other stuff. And then pretty soon it's like a cyber colony building its own launch platform. And then it goes from there.

Mmm.

Right. Yeah.

Right. instead of a 3D printer building a rocket to take the stuff to the moon, just put the 3D printers on the moon and skip the whole. Yeah, I was thinking about that. Yeah, with the spreadsheet example, you're like, hey, why do I have a spreadsheet so I can put all my salespeople in it, look at their sales per month and figure out and sort the column so I can see who the top performers are. Right.

Yeah, and then you just get people there, like the cheapest way possible.

AI automatically sorts this and moves the data into the spreadsheet. But why not just ask it? Screw the spreadsheet. Who's my five top performers? Like, why do I have to do all this? You're helping me put data into a spreadsheet to answer a question that you could just answer.

Yeah.

Mm-hmm.

Yeah, we're, I mean, honestly, like literally, like this could be the metaphor for where we are right now. It's, Kramer on the Seinfeld episode where he accidentally, his number gets changed to the movie phone hotline. And so they start calling him up and he doesn't have the newspaper and they're like pushing buttons. And then eventually he's just like, why don't you just tell me the movie you want to see? Like that's what we're doing to AI. like, it's like that. It's like, it's, it's could have just told us, but we're implementing it into.

Yeah.

the paradigm of load the spreadsheet. Let me put all this stuff in. Now look at the spreadsheet instead of just asking it and cutting all of that out. This is where like we're post software.

Yeah. And it's because someone makes a spreadsheet. It's someone makes a spreadsheet to give it to someone else and they don't know what the other person does with it. So they're like, AI helps me make the spreadsheet, but they don't realize like you could cut that whole step out. It's someone in accounting makes a spreadsheet of the top performing sales folks and then sends the spreadsheet to the head of sales who then like it's just, we're automating what we shouldn't do.

Mm-hmm.

The whole thing has to be, the whole thing has to be looked at as, journey versus destination. And we kind of know the destination was say use the sales one. Like I want to know top sales performers, or I want to know like whatever we have this old journey to get there. And we're not looking at like, well, that journey was built upon a different model, a different paradigm, a different set of resources. If we still want to know this information, what would we do to get there? Like if we had a teleportation pad.

Yeah.

And it was like, yeah, I can use this teleportation pad and I can teleport from my living room into my garage and just walk right into my car, drive wherever I want. You'd be like, why don't you just like teleport where you wanted to go instead of driving. It's like, I've been driving for so long. even think of that. That's, that's it. Over and over and over again, we are like teleporting from our living room to our garage to get in our car. It's like, what? Like, why did anyone at Star Trek walk anywhere?

Alright.

Yeah. Yeah.

Yes.

And not just teleport. like, teleport me to 10 forward, please, Mr. LaForge. it's like, instead of taking the turbo lift, like it doesn't make sense. Cause in the show that wouldn't have been interesting because it would have like killed the reason to be on the ship. You can't just teleport everyone with AI. Like you kind of can or halfway can.

No.

Yeah, like teleport's a great example. I always think of that analogy too. Like we wouldn't have houses. we'd probably have, there'd be a building with a bunch of bathrooms, right? And then we'd live on a mountain top, like our bedroom would be on a mountain top. And when we had to go to the bathroom, we just teleport to the bathroom building into our personal bathroom. We wouldn't need a house. We'd have rooms in various places and just move around. It just changes the paradigm.

Yeah?

I'm raising my hand. I just really, really want the sound bite of don't automate what you're doing, automate what you should be doing. Like Rob, you described that very eloquently in the book, Josh, you as well. You guys are like circling around that. Can someone just say that?

What's up?

don't automate what you're doing. Automate what you should be doing.

It's like we're saying it's like we're talking about it but not but not saying it. Yeah.

Beep beep beep. Try again. That's ADR.

Well, it's funny about that. Intuit example is like, you know, I've used QuickBooks to do accounting for a small business and it was quite hellish. And, I can think of little small things they could do to make it better. Well, I mean, if, even if it could just like, tell me like where it thinks these accounts go, like where these deductions belong, like that, like these little boring things could start adding up. And if it gets to a point where all I have to do.

even do accounting? Accounting? What is accounting?

is upload some tax documents and then it presents me with like all my taxes filled out and we're just going through the empties fields and and working on that stuff together like that actually becomes very not boring. It's that's actually pretty exciting to a small business owner.

Well, and I can tell you, Rob, like the reason we do accounting, like we actually know the reason, like this is what goes on in these like the tax prep and accounting software world. It's just two things. It's first is to, because you have to, to file taxes, right? And if you change, if you change why you need to file taxes or how you file taxes or what the IRS wants to know, you have to change all the software. Like literally a lot of bookkeeping is simply to adhere to the laws of having to file taxes. It's a...

I'm sorry.

It's it's a, uh, Ouroboros snake where it only exists because some other law said it has to exist. But the other thing, this is the thing that actually matters. And this is why Intuit has these huge spikes in business and traffic in January, because nobody does, nobody does their bookkeeping all year. Nobody, nobody. They wait till December and January. They do not sit and do their books all year like they should. They want to do projections. That is it. That is the only reason to do bookkeeping is to do projections and understand cashflow.

Yeah.

Yes.

and

And so you could cut out the bookkeeping and just design something that lets you do projections and cashflow. And then if you have really sophisticated AI, it could probably start to do the projections better than people can because it's math and it's numbers and it's a probability. And that's all they really want out of it. And then black box away the tax prep stuff.

Right.

Alright.

but like the actual, like no business owner needs to know costs of goods sold. They don't need to know like depreciation. Like all of that is just to adhere to taxes. What they really want to know and what they always want to, was just what is actually coming in. How much do I have? How much am I spending? How much am I losing and gaining? And like, what can I expect in the future? So I could start to make plans. That could all be done in a completely different way. And then the tax part, the bookkeeping part could be just obfuscated. behind the scenes to some AI and it just is connected directly to the IRS, which we totally trust.

and

You

So they can just have your data and do it for you.

Yeah! Yeah, I just wonder. We talk about, you know, automating these tasks the way that we do them instead of the way that we should. And my question to you is like, where does, if you were to like, try to assign a group of individuals this like responsibility to figure this out, is it designers, is it developers? where, who, like once code just hundreds itself. What? What? Who? Who? Who's driving now? Who is it? Is it going to be the designers? Like, who should we be tapping into now to say, guys, like, drive this someone take over?

Yeah.

I really, don't, I mean, this is where we like, if we want to get esoteric and really think about the future, I really don't think it exists necessarily with the roles that we have. I do think it's going to be something that is going to be a lot closer to Hollywood and it's going to be a lot closer to like a producer, director, screenwriter. And I don't mean as metaphors. mean, literally those people and how they think. And how they do it because I like, this is like another soundbite. I want this, I want this clip to be a part of the promos. Okay. We're in a post software era, like designing the software and using the software. I think it's, it's, it's kind of like, books are still around and novels are still around, but the movie. Like kind of changed how, if you wanted to write a story, how it would be presented where before I had to give you text and you read it and you had to make the image in your mind. I now can put it up on a screen.

You

Thanks for watching.

And this is also like Marshall McLuhan, hot media versus cool media about like, like the level of investment you have to do to concentrate and focus. So software is the same way where designing it and coming up with it, that's kind of not the focus anymore of like what's going on, you know, on the app and on the screen. It's what are the humans doing in the choreography? And so how is the director directing? And so I think like who comes up with the ideas and like what to do. I really think we're going to have to tap into people who are more into like storytelling, playwrights.

No.

Yeah.

that are into knowing how to like block things out because again, we're designing things that very soon, if not already are like touring complete, like intelligences that can interact back and forth. And so who would you trust to mock up and role play a conversation between a business owner in the year 2028 and a conversational

in this.

bookkeeping voice that just appears. Okay. A designer, whether there's nothing about being a designer that would say, you know, how to do that or an engineer or a PM, but like a therapist, my, a psychologist, a wedding planner, a dance choreographer, a screenwriter, like all of these people can sit and think like, okay, how should this go? And I'm like, Hey, turbo books. Um, I'm getting ready to, I want to purchase a new bulldozer from an excavation company and they're about.

Thanks

Hmm.

$250,000 I found one at an auction down in Vegas. And I'm just curious, like given where we are today, how many more jobs or what do I need to do? Or like, how can we like fiddle with the numbers to be able to have that money on hand? Cause the bank said it'll loan me 80%, but I got to come up with 20%. And so like, wait, what is a fifth of 250, $50,000. Oh, 50,000. Okay. So yeah, I can come up that. Like, what do you think about this? And also can you look around at other auctions and see if there's any other bulldozers? Cause if we already get the funding, maybe we could go buy it somewhere else and like shipping it here.

Mm-hmm.

would actually save us ahead of time. And then it replies, what's going on behind the scenes? Like who cares? The AI is handling that. What was the interface there? There was no interface. It was me just talking, you know, into the speakerphone of my Ford F-150 out on the site. And so it's like designers, engineers, PMs, none of that, career path necessarily sets you up for that. There might be people who are good at that and can easily pivot into that, but I think we're going to have to take a step back and,

Yeah.

Right.

You know, like what was the thing like poets as Kings or whatever, start to think about this from a purely human interaction, choreography standpoint, decide how we want it to play out and then start applying the technology because soft software hasn't evolved since like DOS.

Thank

Yeah.

Yeah.

You just said something super, you connected a couple of dots that I want to dive into your McLuhanism stuff. One of the things he talks about among many things like reshape our tools, our tools shape us. That wasn't him, but he popularized it. But the next thing, was when these, so what you're saying is in essence, the way I'm taking it is. Software UIs are media. And if we look at software UIs as the media, one of the things he says is the things that the new media that comes out always envelops the old media. Like it's, it containerizes it, right? Like, so, so we might have, you know, movies and then movies, then we get television and television plays movies. And then, and then television shows. get enveloped by streaming and then we put TV shows on the internet, right? And so we're constantly enveloping our old media by our new media. And I was just thinking that as we see that conversation and we're going to have these micro UIs, right? We're going to have these conversations and they're going to have these little tiny UIs.

Mm-hmm.

Like, hey, let me to demonstrate this point, I'm going to show you a pie chart or, and, so it's kind of not thinking about the UI as something that's going to go away, but something that's going to get enveloped by a new media, which is AI, the non UI, right. And it's still going to exist inside of it as it evolves because that's

Mm-hmm.

That's how we work. We don't let go of these things, but we kind of reformulate them inside of the new media. it's interesting to think about, I mean, maybe a lot of people are already here, but I haven't really thought about the UI as a media, just like television or anything else, but it really does stand to reason it is. It's just a communication, right? It is a media.

Mm-hmm.

Mm-hmm. This was...

I feel like there's like a euros burros elements of that too. And that if, this new media is primarily conversational, I mean, I guess micro UIs aside, it's almost like back to how we originally started communicating as people. Yeah.

Yeah, there's a loop here. There's like a, like we're circling back to the original media, like around the fireplace storytelling. It's like, instead of a linear path on media where it ends somewhere in telepathy, it's like looping just back to the storytelling by the fire.

Yeah.

It is, and it's, it's becoming, it's becoming like a, just a huge fuzzy logic cloud of communication and information, which is how, you know, humans like pre-technology just had to, you know, use like oral history and like folklore and memory and song and rhyme, you know, to share stories of past information along. It's, it's, but like our brains. We're still basically as advanced then as they are now. And so we have these super organic computers we call the human brain. And yet all we have is the ability to communicate via language and metaphor and unpredictability. like language is an invention, right? Like we're not born with that. We, we learn it. you hear about that kid who was raised by wolves. Like I forget, but they call it like the real life mog or who, the kid from the jungle book. I forget what they, what his name was, but yeah.

Mm-hmm.

Thank

Mugwai is it Mugwai or Mowgli Mowgli? Yeah, yeah, not the gremlin. Mugwai

And the mogwai's are that mogli mogwai's are gremlins mogli. But there was a kid where he grew up with like wild animals and like you get too old to learn language. He never was able to learn language because it's an invention. We learn young. so anthropologically we're back telling these stories and conveying things that are unpredictable. You don't know if what you're saying is going to be interpreted by the other party correctly. you don't know what they're seeing in their heads. You just have to overload it with all these different.

Now that's the blast of the Mohicans.

Hmm.

scaffoldings again, to make sure that it was, understood. It's why they ring around the rosies was like a thing to remind you at a deal with the black plague, you know, and they also, if you go and like Google the like Yucca mountain, nuclear waste storage facility, was this idea to in Nevada and Yucca mountain to create this place, to put all this nuclear waste extremely dangerous and radioactive. And they wanted to create signage that would last for 10,000 years. And they're like.

you

Bye

10,000 years, we don't even know what language we'll speak. Like none of the languages we have now existed 10,000 years ago. Like how would you prevent that? Cause the, it'll still be very deadly radioactive in 10,000 years. And so they had this contest. come up with all these ideas to say like, how would we make a way for Yucca mountain for people to stay awake? Cause we're to put all of our nuclear waste there and it could literally destroy the planet. And they're like, well, what if we put like skulls and crossbones and dangerous stuff and skeletons? It's like, my God, that's just going to make people want to go there more. Like you cannot do that. And so they thought about it.

Okay.

business.

Yeah.

No.

And the idea they came up with was to invent a folk song that they would teach the people and you would learn to sing it about how if like, if the animals around this mountain start to glow green and radioactive, get away, cause it'll kill you. like, that's the only thing they could think of that lasted over centuries was folklore.

Bye!

They just need to take Hallelujah the song and they just change the lyrics.

Yeah, to get the flip away from this mountain.

Yeah.

But that's, that's where we, that's where we are now is like, we can't predict it. AI is like that Yucca mountain problem where it's like when you're talking to something or not, sorry, when you have a customer or a person who's going to use your AI powered thing, you don't know what they're going to say. You don't know how they're going to interact. You can't guard rail them with like interface design and you can't like have buttons and scroll bars and form fields and whatever. Because if like, like how often, is Siri in the room? Okay. Siri's not in the room. How often does Siri just have no idea what you're talking about? I have Google.

Mmm.

You

It's worthless. It's worthless because I will say something and it doesn't understand what I'm asking. But it's like, why isn't it looping back around and trying to understand or like reprocessing? It's like, well, because they didn't sit and think through, open ended questions that you thought through. Well, they could ask about the weather. You can ask about the news. You can ask it to play music, but I want to, what if I want to ask it something that isn't a part of the predetermined like set of inputs? It doesn't know what to, it doesn't know what to do. That is the future.

Ha

And it's like, I don't, I hate using my voice to control things. I never use voice commands with the Alexa or with the Siri or any of that. just, just, I just, I can't, I can't stand it. I like, I think there's ways you could do it with a phone with gestures. We were working with Meta. We were inventing, my, my work partner, Paul, who was a kind of like an XR computer scientist. We were even coming up with ways to create a open source like

and

bootstrap style VR, AR language based on American Sign Language. Because it's like, why are like, it can see my hands, it knows what I'm doing. Why do I have to say anything? Why can't I just hurry and do some like sign language type gestures? And then it picks up and knows what to do. And then when we started giving advice back to Meta on new like, like interface paradigms, it became like skeuomorphic again, where it was like, is there something you want on your wrist or is it on your back or I have an idea, I pull it out of my forehead and all of these ways that don't adhere at all.

and then

to the software paradigm, because most, and we did it because most, XR, VR apps and software, all they would do is take your 2d computer screen, or it's like UI and rip it out and then put it on like a wrist pad. Cause 50 years ago when we were doing lost in space, all we could think of was like, what if the computer was on your arm? It's like, that is the dumbest idea ever. The computer is everywhere. It should just know that I can sign or I can do gestures and it's watching me at all times. So we're like, stop putting interfaces that I have to touch with my fingers.

I don't think I'm.

Thank you for watching.

The whole world is an interface. My body is an interface. Like, when I go like this to like that, why doesn't it know I just pointed at my eyes and it shows you? I circle something and retina tracking knows, and then you get a thing and a pin drops where I was looking and like all of that can just be heuristically integrated into what I'm doing. The problem is you don't know what I'm going to do. So the next step would be, can you figure out what I mean when I go like this? How hard is it when I point at my eyes and then point at something else? I pointed my eyes, I pointed you and I point somewhere else.

Yeah.

Thank

It's so easy to figure out as people, because we're used to being so hard to understand. We're constantly having to monitor the other person. That's what designers need to be thinking of is how can I make something that can understand a hand gesture that I didn't pre-program, but I'm going to figure it out via the context. Like that's the future of software, which is why we need.

Yeah.

Yes.

We need people who aren't a part of like the techno sphere right now to start thinking of new ways to do it. Otherwise we're just going to keep getting more of what we have. And this is, we talked about this when we did the pre-call about like Conway's law is that we're constrained to developing systems that mimic the communication structure of the organization building the systems. And so this is where we go back to AI readiness. You can't build the future of technology in the future of AI.

Right.

Yeah.

Anything else?

from an organization that itself is not evolving to be a part of the future of how AI enabled organizations work. And we don't know how those work because they're just barely a year old. And like that's where we're stuck.

Mm hmm. So what if the, yeah, so what if the opposite of, of, of being ready for AI is productivity and measuring yourself based on productivity and it's creativity, right? As Elon Musk is sitting there saying the future is, we'll have more than anyone needs, which I think is for a lot of people already true. and, And we just go, okay, if the future is creativity and measuring creativity is difficult, productivity is easy to measure as compared to creativity. And being ready is to say, you have to stop defining yourself and your value in the world and in your job as what you produce and start valuing yourself at how creative you are. then is AI readiness about changing, like fundamentally, like you said, therapists, right? Is it therapy? Are you going, is AI readiness more about therapy because of how fast AI is just completely changing people's lives? Like it preparedness therapy, because you can't prepare people. And by the time you get to them, it's passed them up and you're just trying to, you know, pacify their.

Yeah.

You

It already had, mean, we're already like, got unlovable. Right. I heard about that last week. I'm like, what is lovable? Is I get on it's like a vibe coding app generator thing. And so I, I always with these AI's I try to be as like, I try to use it like the worst prompts in as little detail and make it as conversational as possible. Cause I, mean, I see everybody's LinkedIn posts about how to be like a

Hahaha.

Mm-hmm.

A prompt guru ninja, like how to make these huge prompts. give all this bit and they get all the results. I'm like, that is, that is so like anti pick any tech philosopher. Like, like we should not be telling you better at prom. Do we should be able to go and just spit out whatever. And it figures it out. So I go, I just talked to it. Like I'm the dumbest person who's ever lived. And I just asked for things that I want to see. If how much like, I should not be bridging the gap between me and AI. It should be doing it for me. And so I go to lovable and I'm like, I'm just going to build an app really quick. And I'm like,

Yeah.

Mm-hmm.

Hey, I'm starting to notice like, get like these weird mood, emotional feelings and I don't know why I'm just like tired and whatever. And then I have a migraine like the next day. And I was, I was in my forties before I learned that migraine prodromes can occur like two or three days ahead of the pain. No doctor ever told me that. And so it's like, I want to like, I want to have like some sort of app where you just ask me every day.

Mm.

and

Were you oddly cranky? Did, were you tired for no reason? Did you have a hard time concentrating? Just ask me. And, and then like, when we have a headache, we'll like, I'll tell you, and then we're going to go back and see if any of that makes sense. Or like today it's overcast in Maine. Maybe that was why I had a headache. So I just tell that into lovable. I don't give it anything. It comes back. Like, I don't know, less than a minute later with a fully working migraine tracker app. And it asked me these questions. I'm like, oh, this is good. But like, can you also give me a button where it's like, I took a medication.

Mmm... Yep.

So I can remember and it comes back and does it again. I mean, less than five minutes and I saved the prompts, three prompts, okay. A handful of sentences. had a functional, working migraine prediction tracking app. And of course they got this from other places and training data. And like, I'm like, I'm not going to get into copyright and all that stuff. Like we all, we like know how problematic that is, but just, it doesn't matter. The fact was I was able to just type three.

Thank

Mm-hmm.

really poorly worded asks into this app building prompt. And I have a working migraine prediction app that now I'm like, well, this really works. I'm like, Hey, can you, uh, can we like save this data somewhere? Like, where would you store this? And like, whatever. And I started to keep doing it. And like, I'm thinking about just keep going until I can get like a native, um, coded like iOS, like installable and then have that and then start using it and like add notifications. And I'm like, I don't know how to code. I'm just telling it what to do. Like it's already there.

Hehehehehe

Like I'm already asking this thing to build this thing and it worked. It worked like three days ago. And like what's next? What happens when I can hook this to a database? What happens when I can maybe distribute this? What happens when I can monetize it? What happens when I can start building tools and products ad hoc in the moment? So now think post post software. Think about we're going to go back to Star Trek the next generation. I don't know why it just works.

Mm-hmm.

Mm-hmm.

Think about the holodeck. How many episodes relied on the bridge crew going into the holodeck and just vibe coding their way to a solution in the holodeck and just coming up with stuff and the holodeck had to know how to build something that's going to be softer. Like, why would I even use bookkeeping software? I'll just go to lovable or whatever the next vibe coding, coding app would be and be like, Hey, can you hurry and build me a bookkeeping tool that like works the way I work? And then five minutes later it's done and it does my books and sends it off. So then at that point it's like,

Mm-hmm.

Mm-hmm.

Or it's like the star check replicator. It's like, need to get good at taking individual atoms and assembling the food that you were asked to assemble, not getting better at making food. We're getting to the point to where literally software is just going to be a bunch of atoms. And I say, I want you to spin up and then it does it. And your phone is just full of custom bespoke apps that meet your exact need in the way you need it. And so now.

Yeah.

Yeah.

Like who's building that designers, coders, therapists, screenwriters, directors, who knows artists? I mean, it's completely unpredictable because we've never, we've never had access to a non-human intelligence that can think faster than we can until now.

I'm picturing it 3D printing big tufts of cotton candy out of thin air. I think that's wrong, but that's what I was seeing momentarily.

Yeah. I think there's like one of the design questions I have. This is sort of a hypothetical, but a friend of mine started a design company called lousy design. And I was thinking like AI is going to get good enough at design that all design is going to be good. And so it's going to take a human to do bad design because they won't know how to do bad design. And so now we're going to hire humans. to do lousy design and that's what we're gonna sort of appreciate. And the question in my mind is, or will AI get so smart that it'll know that it's time for lousy design and actually start designing lousy apps and lousy designed app or things? Or is that always gonna be our new job to be flawed? Like is that what our place in society is going to be is to be flawed. Just like the Olympics, right? I don't want to see robots running. Like, is our new job to be flawed in unique and interesting ways?

I mean, this is so beyond, I think, anybody's ability to predict. I had an idea for a sci-fi book I wanted to write about when AI can produce things for you, what becomes the bougiest, the expensive thing is to have storytellers, let's just say playwrights, verify. that they're going to write for you, the wealthy person or powerful person or whatever, they're going to write a play just for you to be performed performed by people just for you. That isn't digital. That wasn't generated because it's the only place you can actually get something that didn't come out of the machine. And that that will become the new like luxury item, which is like human built stuff. And it all circles all the way back around to when, know, um,

Right.

the powerful people of, you know, historic times, emperors and Kings and Queens and like Pharaohs, whatever were the only ones who could afford to like commission like art or to have a play done like just for them. And then imagine what it would be like for some reason I have like Elton John in my head or Elton John, like we'll write a song just for you in the future and only perform it for you. And you invite people to hear it and you make sure there's no recording devices. And it's like.

Thank

Mm.

Cause you like, I, if you can't like reproduce an Alton John song with AI right now, I would be surprised. Like by the end of the year, for sure, we're going to add audio reproduction, like how Sora AI is doing like video reproduction. Right? So imagine the elite status of being able to hire performers or the, what you're paying for is the verification that it is done by the actual artist or performer or musician or whatever.

listening.

That's it.

And not the machine. Like that's kind of like all that's left. Cause you'll be able to replicate the real thing. And we almost are going back. Like it's weird. We're going back like hundreds of years to where people will want to go out and see a real play. Because you are so overloaded with digital stuff. Like the ultimate experience will be to go out and like have an analog night.

Mm-hmm.

Yeah, and see a custom play if you can afford it.

Yeah. And by the way, you're not tied to your desk and computer screen so you can because there's nothing for you to do in front of a computer screen anymore that a computer can't do on its own. like, yeah. Super cool. I've enjoyed this conversation. I think

Yeah.

The big takeaway for me in terms of our journey was the, this McCloughan thing, it really hits me that, that this is, UIs are media and that the new media will envelop the old. It will be, the old will be part of the new. It's not just going to be words. It's, it's, it's going to be little pieces of UIs. and a little bits of what was valuable in that kind of pulled out and improved, but that the new, the new UI will encase the old. that's kind of an interesting way to think about it. And it's kind of what's happened with, with every channel that's, that's come out. interesting. This is great.

Yeah. It reminds me like one final thing. We'll pull this back to like pandemic era. Does anyone remember Quibi? Good. So Quibi was a very short-lived micro content TV, where the episodes were like five minutes long and they hired all these A-list actors and they were only available on your phone. And then they were meant to consume like, like when you were,

No.

Stuck places or doing things really had five minutes. So the episodes were all. You know, created to have like five minutes story arcs. They'd release a whole bunch of them. So you could watch a bunch in order. And they would also do it where they were filmed and framed, like the cinematography to be looked at. What do call this portrait mode? But then if you turned your phone landscape, it would actually turn to a different framing where I don't know if they shot it twice or had two cameras, but it was just like the idea that even, the idea of. A television show was changed again. And it could be failed very, very quickly. mean, I really only think it was out for like a month or two, but they tried to say, okay, we have this idea that people want to watch shows and stories. What's the next thing? How is this pandemic has changed like media and content and like binging and like the amount of time people have, and you're wearing a mask waiting in line, social distancing at the post office. What if you could pull out your phone and watch a five minute thing, whatever. And so then they changed how the stories were told. So now.

Mm.

picture like Quibi 2.0, you can keep that idea free where you could just tell it what you want to happen in the story next. And so you get to the part and there was one where, um, there was one where I don't remember what was. It an anachyndric and like, like her, her boyfriend got like a, um, a girlfriend robot Android that somehow like came to life and like talked back and forth to her. Imagine if you could just tell the show it's like, Hey, you know what, for the next episode, I actually want this to happen.

Yeah.

Hmm.

Or like change it up. Can you actually make this person and this person fall in love? And then it goes, and then it shows you the show there. And you say, I only have, I have eight minutes until this train gets here. Can you make the eight minute episode and then it does it, you know, now it's like we go from movies to radio to movies, to TV, to TV on your phone, to streaming to now things that are just generated for you ad hoc. And it's like, at what point, at what point do we eclipse?

You

Mm-hmm.

Like the reason for creativity, which was take what's in my imagination and somehow bring it into the world so that you can experience it. Well, what happens when it can just be created and when you want like a movie to entertain you in a certain way, you can just ask for it and it will create it. And it's like, is it disempowering creators? Yeah, but it's also like, it's empowering people to take what is in their imagination.

and some.

That's it.

Yeah.

And make it a reality without having to have an intermediary and like that a post scarcity, post capitalism society, like who knows, but it's coming and it's already here in a lot of ways.

Yeah.

Yeah. Yeah, it's the it's I think it's the age of the endless scroll. Like there's no you just go tap into your favorite show that never ends or read your book that doesn't have an ending. It's just the endless scroll. And and do we want long form or is it just going to be these little snippets like TikTok videos that never ends because it just gets better and better at hypnotizing us and watching it at the end.

Mm-hmm.

Yeah.

I'm an optimist, so I'm a firm believer that, you know, you say Star Trek, one of the things I used to think when I was a kid is they go to these planets and then people would like have a garden and I just would not understand. Like, why would you have a garden? Like, look what they have, right? They can beam stuff around. Why would you need a garden? And now it's way less ridiculous to me because I feel like...

You

If this stuff is taken off our plate, maybe that's what we want to do is grow our own food, not because it's faster or easier, but because that's what we would rather do than sit in front of a computer is go to our backyard and grow our own food. And now I kind of get it. maybe that's life is circling back. Maybe this is a loop. You know, we're all on our way to being more Mennonites.

Yeah.

Maybe. What if technology, well, what if technology was just a temporary thing for humanity that was meant to erase itself back into the background? Because we haven't evolved biologically in hundreds of thousands of years. And so is really the point of the last couple hundred years of the industrial revolution for it to like reintegrate itself back into where we do have gardens and walk around. And you wouldn't even notice that technology, because at the end of the day, we're still wired.

Right.

for a pre-agrarian lifestyle as hunter gatherer, like primates. And so is this all just a way to get us back to that where we were probably, we didn't live as long, but we probably had a much different type of life and a much different palette of emotions. Cause I don't know about you, but I don't remember the last time that I felt like the modern world and civilization was doing more than just introducing more.

That's it

Yeah.

Stress and more doom and gloom and more uncertainty. It's like we're living the future. I don't remember who says, but like we live in the future and everything sucks. It's like, I don't like all as I dream about is being able to go and have a homestead or a farm or a cabin just to get away from everything because I've been hooked to the internet and technology for, mean, literally 30 years now. I don't feel any better because of it. I just, I just have a job.

Yeah.

Yeah.

Yeah.

No.

That is about that, but I don't feel like it's actually made me any happier. Brought me any more joy than any other time period. And I have a pastime. You'll edit this out. Don't publish this. Just kidding. Do publish this. watch, I watched 10 hour long YouTube compilations at night, just to zone out of TV commercials from the seventies and the eighties, just endless TV commercials. Cause it reminds me of a time before we were being crushed by the internet and the

No.

sorry.

The chronically online lifestyle where everything was just digital and everything had to do with like the web. And just to think back to where it's like, all's I would like love is a way to disconnect. I can't do it. I'll never do it, but maybe we evolved to the point where it becomes kind of like a background noise instead of this thing that's at the forefront. And I mean, if I see, if I see another commercial advertising, you know, an AI while I'm watching TV or whatever.

Mm-hmm.

And the commercials are basically saying like, use AI to stop making the handcrafted signs that your grandfather made for generations and just use this to like replace whatever. It's like, my God. Like we have now become a parody of what like Chuck Palahniuk wrote about in Fight Club 25 years ago to where we are now going and buying the glasses that have the little imperfections and chips. So, you know, that they're handmade by the people. And it's just like.

Yes, yes, yes, yes, yes, yes, yes,

Yes.

This has all been predicted and it's really cool and it's fun and it's interesting, but I have yet to see it improve anyone's life.

Yeah, it's like the new AI that can go on vacation for you. Buy it here.

Polish on the brass on the Titanic, right?

Yeah.

While you work in the mines of Amazon coal making, you know, they say there's going to be two jobs in the future. You're making $400,000 a year as an Amazon programmer, or you're making $40,000 a year as a miner in the Amazon coal mines that fuels the reactors that power. But that was like pre AI. Now there wouldn't even be the Amazon program. We'd all be working in the coal mines, fueling the reactors that power the AI engines. Just go read Harlan Ellison. have no mouth and I must scream. That is the point.

Yeah.

You

of the whole story is that the AI took over the entire planet.

Yes. Well, great guys. This is awesome.

We'll have lots of content to watch.

catch up list on this episode is going to be all these links of books and movies and things you have to go watch.

I get to read Fight Club again.