The Danger of "Intelligence Washing" in Startups (w/ Skylar Shaw)

That's why I say AI is so wonderful. I've been using it not only for work, but also in creative pursuits, my hobbies, songwriting, and playing little garage basement bands, you know, I've been doing that for years. Yeah. So, I met someone who had a certain personality, which was very chill and vibey and creative and awesome, but then I received an email from him. It was nothing like him, but I said, "You didn't happen to use AI? " He's like, "How did you know? " Do you think the the legal system's being upturned by AI? Do you think it's affecting negatively? Is AI hype or help? Good day and welcome to the show. I'm your host Pete Cooper, and our topic is AI hype or help.

Today's guest is Skyler Scholl. Skyler is a consultant with a legal background. He helps startups and other small businesses grow as a consultant, but also with his specialty in legal. He's got some really good insights into AI. So, without further ado, let's get into it. Well, welcome Skyler. Welcome to the show. I believe you are in the field of the legal field, but also very knees deep into AI. So, I'm really looking forward to getting into some of the changes that are happening with AI, but before we do that, what do you do, Skyler? Hi Pete. Yeah, it's great to be on. Thank you for having me. So, the the the quickest answer is I'm DC-based. I'm an attorney.

I have a relatively diverse background that includes business, finance, investment fund executive, in undergrad among other things, you know, I studied design and product and industrial. So, I have a fairly holistic um, you know, approach to being a lawyer, I suppose, but still all the the traditional goodness. I primarily help small clients who are either startups or small organizations to, you know, start, grow, change, and as I like to say, stay out of trouble if they listen to me. And mostly they do. Um so that's kind of the quick answer.

I'm happy to expound upon that, but I uh you know, my my clients are often small businesses or startups involved with tech, and one of my favorites is the intersection of technology, creative endeavors or professions, and or great causes. And on the Venn diagram, uh some some of my clients are one of those, a few of my clients are none of those, but some of them are two of those things, and occasionally I have what I the beautiful unicorn trifecta of all three. So. So uh I mean, having been a founder myself uh of of a startup, it's uh definitely fraught with uh ways to make mistakes, right? And very very easy. Um do you have any sort of do you work on the business side as well?

Or is it mainly just focus on the legal side, trying to keep them out of trouble legally, across the board? How do you fit there? Great question. So primarily I'm brought on or hired as legal counsel, but what's happened over the years is on the side I have been a founder, a principal myself, and then occasionally what happens is if we mesh and gel and things are going well and the startup survives any amount of time, occasionally I'm pulled in as a principal or co-owner, co-founder, who, you know, obviously takes the charge on the legal issues, and in that respect I can become in-house. So it just depends.

Sometimes what happens is because my background is a little more diverse than some attorneys, and I very much love the creative in the business, the strategy, etc. Like I said, sometimes I get pulled in to being um well, in-house, you know.

Mhm. Mhm. So um I've got many many stories about things that I have done and failed, and I I'll try to think of some good examples, but I think one of the big ones is not going after the problem. Um I've been accused of that like uh building some technology and trying to find a market for or a fit for it. This is a great idea. We can do this, but do I really have a customer problem? Do you see that a lot? Absolutely.

There are some really talented people. You see it obviously in Silicon Valley the most, but it can happen anywhere where someone builds something really cool. And they they basically have a solution and then they go search for a problem. And it just so often isn't a problem that people really wrestle with. Now to be fair, they could pivot into something that is. And as I always tell people, you know, don't expect your first time up to bat, you know, to use that sports analogy, to hit a home run. But it can get you to meet people. It can get you to network. You can pivot it.

In fact, the last startup I was brought into, not just to advise, I unlike my earlier days had enough sense to know, you know, it I don't have a particular reason to think this startup is necessarily going to succeed in a traditional fancy showy way, but I like these people. The networking's there. Um I enjoy the camaraderie and whatnot. So for me, it wasn't so much thinking and saying will this blow up or be acquired? Because, you know, when you start talking IPO, that almost marks you as a bit of a a newbie or green. But, you know, reasonably speaking, a successful business lifestyle is not a dirty word, lifestyle business or an acquisition Okay. That's reasonable, you know, nothing crazy.

I said, "That may not happen at all, but I like these people. I'm interested and I have something to learn and something to give. Um and that was enough. " And I think the problem is it's so hard to manage expectations. When I was much younger and greener, it was like, "This is the most exciting thing in the world. " And then you're crushed if it doesn't happen, right? Yeah. Yeah. If you're too cool and too aloof, you just don't care. And you can't not care and get into something involved in something speculative. Mhm. The trick is through time and wisdom experience, I'm starting, as I grow up, to learn about how to live in that beautiful middle area where you care, but you're not crushed.

And you think of it all as a portfolio model, right? Mhm. Mhm. You're just going to whiff. And then if one brings you something of value, awesome. Yeah, I love that model, Scott. I think lived lived this life for 12 years. I've always got multiple things on the go. And it's a question of sort of where I need to press and where I need to manage, you know, whether it's cash in the door or working on a big opportunity that I see. And that idea of a portfolio definitely resonates with me. And it helps you I think it helps me at least care about my portfolio and each program, but not be crushed when one of them's had a setback.

You know, I can say, "Okay, well, you know, maybe I just emotionally need to just take a step back and go over here a little bit and spend some more time over here. " Unlike if you're in a job where you get fired from your job, you know, it just comes to an end. You lose all your relationships, not only financial, but your background, everything just just stops and you hit a cliff and you fall off, you know, so nice if you get picked up again, but and straight back in, but that doesn't always happen to everyone. Yeah, yeah.

And you know, one of the things I can do through time and experience and honestly learning a million I say to people they say, "Well, you've been in startups, you know, is is your power what you bring to the table that you've hit it huge, you know, you're the next Mark Cuban or whoever. " I said, "Absolutely not. But I I can show you 101 plus ways to not do it because between me and my co-founders, especially in my younger days, we've done it wrong. Mhm. You know, and and that can teach you a lot. Well, Scott, after this, maybe we should talk about some of the stuff that I've been working on. Um Sounds good.

systems, AI stuff around in the healthcare, and then I've also spent last 2 years in in construction. Um seems to be lots of opportunity over there. Seems to be construction is a little late to the uh not the AI. I mean, I think it's starting to happen in AI, but just technology in general has been it's been slow to uptake. A general resistance to technology. I think people go into construction to try and avoid technology. So, when it starts to come back in, there's like, "I didn't come here for this. " You know, it's I wanted to do the hands-on thing. So, Yeah.

No, I've never seen with my uh dealings with my real estate clients and such out there it's uh it's not that it's not there, but it does seem to be not as adapted as quickly as the technology in creative professions. Yeah. Yeah. All right, cool. So, um what's happening? What's happening right now with with you and I guess your one of the key focuses is the legal side of things. What's going on right now in the industry um when it comes to your legal lens? Oh, goodness. That is an amazing question, Huge. We'll We'll just start somewhere and try to pull a thread on that sweater. Um you know, I mean, one thing to keep in mind is that the law always lags technology. And that's not entirely bad.

It's certainly foreseeable. It's It's really always been the case pretty much. Um and it's natural. Law is uh often a little more reactive than than forward-looking. Uh at least in our legal system. You know, in civil law countries, maybe slightly less so. Um yeah, there's there's so little regulation. There's obviously a lot of talk about it, but there's precious little. There's There's relatively little case law even especially that's practical. Mhm. Um and you know, my focus is on small clients and being in the trenches.

You know, obviously if you want a really keen day-to-day perspective on the Fortune 100 or 500, I'm not, you know, that's not so much me, although I keep track of the news. Um but I see a lot of hardworking entrepreneurs, little businesses, and startups uh struggling with the promises, the hype, and and the intelligence washing as I call it. We've heard of whitewashing, greenwashing. I think there's a lot of intelligence washing, and by that I just mean everyone's product is AI. Uh and you know, we know a lot of that is layers built on top of things like ChatGPT-4 plus and uh you know, what, five plus now, and and Gemini.

In other words, you know, and even some of the clients I currently have, they're doing things to repackage and re-front end that cheaper, more or less free, pretty reliable data that AI is parsing and things. So, a lot of that's going on. Um advisors like myself in legal, but not just, also are being pushed up the value curve. Certainly in perception, but to some extent in reality. Um and by that I mean increasingly a really savvy person with AI might, I stress might, might be able to come up with a draft or legal advice for something incredibly simple with AI alone. Mhm. And I say might.

Um but the problem I see is it is instilling a huge amount of confidence that may not be entirely warranted in a person that doesn't do this. And the smarter the person is, sometimes the worse it is. Um you know, I think it's always a good thing to assume that everybody's job's hard, you know, I don't assume to know more about the arcane aspects of tax than my accountant. I certainly don't question my doctor or dentist too much. I mean, I'm not saying I'm not engaged, but so I think you have to be careful that I think it can be good for issue spotting. It can be good for education. It can be good to tee up more concise conversations if your advisor's hourly. Mhm.

But it can give you a false sense of security. And I'll give you two examples that I think are just perfect. One is in less of a business context. So in the context of estate planning and this has already been true of some of the the document providers and sort of the you know, whoever picks up the phone, some of the legal services. You know, there's not really necessarily deep expertise or a relationship. So in the estate planning context, when you're providing for your children, you know, you you may think automatically of who can be good for them day-to-day, right? Get them off to school and such. But another big aspect is who can handle their money. Often some money's left in trust.

Now, someone that does this for a living can tell you that it's often not the same person, right? Doesn't need to be. But many of these forms and things I see or what AI can cough up is a single blank in a form for first for one person. Now, the problem is if if you don't do that for a living or you haven't been through that, you don't even know to ask the question. You don't even know what you're missing, right? Mhm. A more germane example I see for business is, you know, a proper buyout or buy-sell agreement for two or more co-owners can easily be the streamline version I use is about 18 pages and that's keeping it as plain English as possible because there's a lot to cover.

And if you do it well, there's not a lot to fight about. Mhm. But but is some of the forms and what I've seen AI put out too doesn't even have clear restrictions and certainly no detail about transferring ownership, right? That is a massively important that that takes some ink to do right. And if not, you you might be in business with your your especially for a young founder your co-founder's roommate or ex-boyfriend or girlfriend or something. It's so fundamental, but the thing is you wouldn't know to ask if you don't do this for a living or if you haven't at least been around the block a time or two. So that's why I say AI is so wonderful.

I've been using it not only for work but also in creative pursuits, my hobbies songwriting and playing little garage basement bands, you know, I've been doing that for years, yeah. Yeah, yeah, I love it. Well, you can see the instruments on my back wall. I'm a I'm I'm a fan. But so it can instill a false sense of confidence. I It helps me a lot in my job. It's almost like because I have pretty lean on staffing. I'm not a huge shop, you know, it's almost like a super smart, super fast, eager junior person. But the reason I can use it well and I assume it's true for other advisors and professionals is because I already have 20 years of experience to filter the output through.

I know what I'm looking for and often time I'm I'm going in not I have my trusted drafts that I update, you know, my templates that I've had for years. It's about some very arcane thing I want to dive into and a starting point for research just like Wikipedia even can be a starting point for research. But I think if you let it be your ending point, you may be asking for trouble at least as it is now. Good. Yeah, I think you know, I'm going to make a simple example where you know, I use a spell checker.

Now back in the day when I was still at college I I had to write up my my thesis my undergraduate thesis and the idea of doing it with pen and paper which seemed so painful to me that I went and bought a computer with a word presser presser and that was the best thing I ever bought cuz then I could edit what I was writing. I still never became any good at writing and I would make a lot of errors and then kind of, you know, tools came in to help me with the spelling. And that was a big breakthrough, and that's only been in the last few years. Now, when I type an email, the spelling grammar's all fixed up for me, but it sometimes changes the meaning.

All right, if I let it go, and it'll change, uh, you know, the cat sat on the mat to the dog sat on the mat. If I don't read it and and accept it, then it's going to go out wrong. And then that becomes even worse with with the, you know, ChatGPT's a good example, any of those things where if you don't review what it says critically, it can be completely wrong. And and I've had now my team, I encourage them to use the tools, but I say, "Do not give it to me unless you understand every single word. " And the same thing with coding, unless you've coded, you understand every single line, you know, you don't know it. And so, initially when they come on, they they realize I'm reading.

I say, "What does this mean? What does that mean? " They don't know. So, now they have to know. Certainly, to me, it helps me write faster, research faster, but it can't And it it temptation, I think is what you said, smart people tend to sort of be more sad. You think, "Well, you know, I've created it, it must be great. " Even if I'm using the tools, um, without sort of just hang on, maybe this is all wrong. What is this thing? Can I check fact-check that, you know?

Can lead you down a very path of reinforcing your as a founder, you've got that natural positivity to see the things that, um, encourage you to keep you optimistic about the future, and completely can miss something really important. Oh, absolutely. And what I analogize to it, I remember way back to undergrad when I took a intro to stats course, the professor said, "Look, the first time you do these basic stat statistic, um, equations and work this out, I'm going to have you do it by hand. He said, you're going to hate me for it. And after that, I'm going to tell you to use Excel for the rest of your life, and I'm going to show you how to do that.

But, if you don't have some sense of what it was like to do by hand once, you're not going to understand as well. And same with as I I was in design school. You know, we do this classic Bauhausian intro foundational year thing where you're doing color chips and a gray scale. And the temptation is to try to eyeball it and pick just 10 chips in the right steps. And the professor says, you're going to be tempted because it will save you hours, but you will not have a great result. It's the difference between a B and a C or an A and a B. And, you know, like Hans and Franz on SNL used to say, hear me now and believe me later. And it's true.

And just like um in design school when they said, we're going to show you how to do these things on Illustrator or Photoshop, but you need to know how to do some basic drawing to communicate your ideas. And in a lot of instances, for internal purposes, you can think quicker with a pen or pencil than you can with dragging, you know, the cursor around and doing all these things. I mean, AI's a little different. But so, there is you're only better AI can be at its fullest when you know when you have some sense of how to do the underlying thing. So, what it may be most uh long-term great at is um helping to teach you educationally to be able to use it better and to have the underlying skill.

Mhm. So, underlying skills is another one. I think here in the US, they're they're not they're no longer teaching people to write cursively. I mean, I was taught, probably you were too, you know, running writing, very fast, um and still readable. I still make notes every day, all day, every day. And when I turn up to meetings, I've always got it there. And if something comes to mind that I need, I can just write it down really quickly. I look at younger generation, and and I guess it's I don't want to say that necessarily, but a lot of people any anyway, that's just don't turn up with a pen and paper and they come to a meeting. How they going to remember everything that was said, right?

How they going to remember that, you know, the certain number that was maybe mentioned? A pen and a paper to me is such a fast way of keeping me engaged and documenting what's needed, so I can go back to it that I can't dismiss it even though I'm using all the AI tools. And I think this is maybe it's just a a personal thing, but I think it's also very indicative of what you're saying is that learning those base level skills um still carries over very well into the use of AI. And perhaps young people when they come through they use AI so much they don't have that underlying foundation and therefore it's even more difficult for them to see when things are wrong.

Yeah, well, you're talking to an enthusiast and collector of contemporary and antique fountain pens and I use them every day. I've got got some right here in front of me and you know, that perfect paper pen nib and ink combination and there's you know, thousands of inks and so I mean, so I'm a little biased, but yeah, and I believe that cognitive research it's widely accepted that suggests that the problem is is if you're a good typer, you can and I know this is true and I've tried to do this even with a pen and it really doesn't work.

You can accidentally sort of create a transcript which is not doesn't imprint on your mind generally as well as note taking that's more selective and you have to think about what you're writing because your hand goes a bit slower most people than your typing fingers can and it's actually more helpful to do actual notes than it is to try to create a transcript.

Mhm.

And I remember really early in my career, you know, being eager and bushy-tailed I'd go in with the managing partner, my mentor and we'd take a client meeting and I would write so furiously and in court, you know, when I was second chair, um, you know, I would do that and and it was like I would have this verbatim almost and you can't quite even do like a stenographer, but I'm like I didn't remember as much as when I occasionally took a note and I listened. Mhm. Uh, it's like watching a concert through your phone screen cuz you're trying to tape the entire thing.

Um, there's a place for both and you know, it's easy to say, "Oh, you know, you guys you're just old men, you know, shaking your fist at the clouds. " Well, no, there's a place for both. I love and embrace technology. I have a part analog, part digital. When it comes to search and storage, nothing beats digital. Calendar, nothing. Although my better half, she's tech-savvy and loves a paper Japanese planner and calendar, you know, great. But when you combine the best of both, uh, I tend to take notes in longhand and ink on paper and then I translate that and usually that review and translation and whittling them down and putting them somewhere digital, um, is really effective for me.

I think for most people, certainly for me. Yeah. Yeah. So, I mean, I don't know a lot about the physiology of the mind, but what kind of makes sense to me when I do that is this person I used to do it when I was studying. You you'd go to all the course the whole course for the whole semester and you had to kind of compress everything into summarizable formats that you could then remember the key things that you would take into the exam cuz it was often closed book and there would be certain acronyms that you would uh, that you would um, soccer tour was one of them for for the the trigonometry one, um, which which I remembered.

And um, you know, so there was this kind of constant summarization that came up that helped you gain access, um, to, you know, the whole year's learning in a very summarized format with just, you know, and I would find myself summarizing and then re-summarizing and re-summarizing. So, when I got to the exam, it was all like like there I mean that Soc 2 thing, that was taught to me at school. I still remember it. If I had tried to to to record everything and have it transcribed through AI and then have it presented to me, it would be completely useless. I mean, I couldn't do anything with it. No, absolutely.

I I don't know if you had this I remember in in high school chemistry, we would sometimes be allowed to have like one cheat sheet, one little index card. And it was about how many things you could fit on that index card. And the funny thing is in hindsight

site. As silly as that was, you learn a little bit of something about what can you cram on there. You know, what formulas are most important because you can't put them all. Yes. Um so, it's remarkable how even that can teach you something. Mhm. Yeah. Yeah. So, we talked a lot about sort of modern technology coming on and and and of course a lot of this is is AI.

So, in terms of the legal world, where what's happening and how is AI shaping what's happening in the legal world? Ooh, that's a big one. Um let's start with some things that are more top of mind. Um I would say you you have some folks that are increasingly finding the boundaries of leaning on it for legal citation and legal briefs. Um and that has its own problems. I I'm a little more focused myself on the IP issues in particular trademark and copyright. Um those have an inherent interest to me as as well as my daily work. Um there's just so much going on. And you know, there's privacy issues, data issues.

Um I you know, an an issue that's that I've been thinking about a lot the past couple weeks in particular is you see increasingly these new products. There's a new one at least every day, maybe every hour that is um uh you know, basically let this app have access to all your inboxes and it will truly be an assistant for you. Mhm. Mhm. And that is very alluring. The problem is is we lawyers have to be extra careful. You know, some some industries more careful than others. Um I am not going to just let any old third-party app that started a week ago have access to my client information. Mhm.

And so there's this sort of cottage industry that's fenced off of a couple that are more expensive with fewer choices that are sort of either bar association stamp of approval on or you know, and that lags. But I think what people have to be careful is there's a couple things there, right? First is that startup going to be in business in a month or two after you're on boarded and and get up to speed on it? Mhm. Maybe maybe not. And two, even if they're well-intentioned, what reason do you have to think that these strangers are going to do the right thing by your data or aren't? And they don't even have to be nefarious. They can simply just not be particularly with it on that.

I mean, um you know, if if if um Yeah. So so those are some things that have been on my mind.

inbox thing, I guess what the a term is data leaks, right? Between between one company and another cuz you might have several email address addresses for different accounts that you hold and you have to hold them each private. If you allow AI to see all of them, you know, one can get across to the other or it could just get completely leaked out. Yeah, absolutely. And and to be honest, it's a roughly it's sort of a digital version of the concern you might have if you're new to having uh remote assistant or a virtual assistant or in office. Um you have to know and trust that person.

And look, that's classic since the dawn of time. What human can you trust with your trade secrets and things? Right.

But, I think it's tricky because the innovation, the features, and the things that are exciting are on the edge of these brand new startups. Mhm. And these brand new startups are also who I can't really necessarily trust with my sensitive client data. Mhm. Um So, that is a needle to thread. Yeah, right. Exactly. So, I think one of the other things which maybe alluded to but didn't mention specifically is that to me that in the legal world there's a lot of contractual documentation type things which to me were always very hard to read, hard to understand sometimes.

Um what makes sense to me is I take that legal document and I put it into an LLM and say, "Make this simple for me. Give me the the top 10 things that I need to be aware of or whatever in this contract. " And then have a bit of a dialogue back and so I can understand it. Um that to me was a really good shortcut. Do you How do you feel about that? Well, it can be. Um I I I think summarizing a contract can be helpful. Again, not not as an endpoint but as just a starting point. Like, what does What does this even say? Or what are just the business terms?

The The irony and the problem is that some legal terms of art, although they may be seem impenetrable and arcane, the term They may be in Latin occasionally. Believe it or not, um this may not be the best example but in copyright when you're hiring um an independent contractor to do something, uh you know, let's say in graphics or logo or something, and you use the term work for hire. That has a very specific legal meaning in the US and I I presumably elsewhere. Um And again, not the best example. There's more impenetrable phrases. But, that has a specific term of art meaning that if you don't use the term of art, you can use even plainer English, let say.

And this comes up even more in real estate or estate planning. They have more Latin. At least in the US, but you might take a paragraph to say what those three words mean to other lawyers and a judge, right? And the problem is is if you reinvent the wheel of drafting, you could put it in plainer English, take a lot more space to say it, and then a judge has to lift their glasses and say, "What Oh, you're saying work for hire? " But I had to go back to my chambers and read this paragraph and So, believe it or not, sometimes a an arcane phrase in law can actually be a shortcut for people that are in the profession. That's a way to think about it.

So, that's an irony that I'm all for demystifying everything. The way I look at it is if AI or a form provider puts me out of a job, then I'm not particularly good or I'm in the wrong field or or whatever, you know. Well, that's interesting. I think that happens across every expertise, every field. There are shortcut terms that, as you say, have a lot of meaning. I mean, that one I talked about with the geometry thing, the SOH CAH TOA thing. That was a a very short little acronym that I could use to remember how trigonometry works. Um but in in in my field, I guess in tech and in AI, there's lots and lots of terms um which have, you know, a lot of embedded meaning.

And when you talk to someone outside of that field, uh it's often quite difficult to convey the meaning, right? Um cuz you you used to I mean, we often shorten things to, you know, a couple of letters, so we don't have to keep saying the the term. Right. You know, and then uh Yeah, so I think that's a very common problem, and it kind of you can't really dumb it down. I think that's the thing. You can't really dumb it down without making it um exhaustive and then slowing everything down, and and then anyone who's an expert knows the language uh is waiting for everyone else to catch up. Um Yeah, that's very interesting. I hadn't sort of thought of it like that. Yeah.

Well, I guess the legal world is full of things which have that deep meaning, and of course it it's even you when you study law and you and I I did only a little bit as part of my MBA. The all the case law and the thing that's implied from the cases, you know, you have to kind of read the case to understand how that law came about. So, the whole history has has context. Yeah. Yeah. Yeah. Absolutely. I suppose the the flip side to that is it's all accessible, isn't it now? It's um Uh we've I mean people like to use the chatbots. I know I do. Because it takes it in a direction that I want it to go. Um when I was studying, it was always however the lecturer wanted to present it.

In in the way that they wanted to present it. Which often meant that it was the worst way for me to receive it. Whereas if I've got an interest, I know some certain topic, and they're talking about I can't don't tell me anymore about that. I know that, but now answer this piece. I'm going down a very quick learning path, cuz I'm just focusing on those things that that I need help with. Absolutely. And I'm all for anything that sort of democratizes or levels the playing field for people that have less fewer resources or less privilege. I mean, so in that way the web, you know, sort of cracked the door, and maybe AI is further pushing it open. So, I'm all for that.

Um but you know, occasionally I'll see I'll show you I'll tell you a perfect example. It wasn't that long ago. I met someone who had a certain personality, which was very chill and vibey and creative and awesome. But then I receive an email from him. Like the kind of person you never picture in a suit tie, and you know, God love him for it. I Unless I happen to be in court, I'm I'm generally not anymore, either. Um but then I receive an email that said, "Dear sir, we're here to for I was pleased to make your acquaintance. " Like AI wrote that? And maybe because I'm a lawyer, I mean, despite me being casual in demeanor, um it was nothing like him. It wasn't him.

He didn't happen to know it, and I just said, "Hey, you didn't This was, you know, almost 2 years ago, and AI using AI was brand newer. But I said, "You didn't happen to use AI? " He's like, "How did you know? " And I Well, because it wasn't you. I mean, it wasn't even close to you. Right. And so, I think we have to be particularly careful about you know, the the button that's like make it more formal, make it less formal. I'm not saying it doesn't have a place. And then you hear about uh specially teenagers turning to AI as a companion about serious issues they can't talk to parents or adults or about romantic advice or things, and you're like, "Ooh.

" That just sounds really potentially dangerous. So, I think we really want to think about and at the same time those kids need advice that maybe they're not getting from adults or all their counselors, whatever. So, it's a very sticky wicket, as they say. I I don't know the answer. I I'm I'm lucky on a good day to know the questions, but you know, they need something, and and but you know, it's like a tool. Gosh, the oldest analogy I go back to is a sharp kitchen knife. It can be the most wonderful tool. You can prepare a meal on a holiday for loved ones. You could feed underprivileged folks on the street. Or you could use it to go slash someone. Yeah. Yeah.

It is simply a tool that without knowing how to use it and without caution can be as dangerous as it is beneficial. Yeah, I think that's a good way to to put it. In In my world, I I double down on using LLMs a lot more this year. I said, "I'm going to commit to more time, more time. Um and it's become almost like a business partner to me. And and a business partner in this in the good and the bad sense, right? Um in the good sense, it's tireless. I can keep on asking it and I don't have to be polite, I don't have to arrange meetings. Um and it's very, very productive in terms of giving me, you know, what I need quickly. Um but also it kind of is a really good bullshitter as well.

And I think sometimes you meet someone who Well, we all have a tendency to I guess show our good side, right? To show our good side, to show that we know even if we don't completely know the answer. And I said there's a little bit of human in this thing as well, like um where you're you get your business partner to comment what they believe to be true and they kind of make it up on the spot and I think this is true and I heard this over there, but actually fundamentally they could be completely wrong. And it's I guess it's kind of like the same human condition, but just a little bit more amplified. Absolutely. In fact, I I think it's in part the art of the prompt.

In fact, something I was kind of doing, but didn't know it was called this more recently. I see a lot of articles, mostly in the tech world, but not just about the three-step or three-phase prompt method, which is something I had I hadn't called it that and wasn't that formal about it, but always asking for pro and con, but it more importantly giving the AI a persona. Um in other words, you know, you can say don't be a sycophant. Uh be devil's advocate, push me. Um things like that.

Mhm. Because you're right, it it will have a default to be sort of flattering and things. And when it makes a mistake, which it occasionally does, I'll point it out and then it'll say, "Oh, good catch.

" And you you know, but it's it's yeah, so flattering. And you'll love this. Maybe 2 years ago, year and a half ago when I started really using it, I would put in prompts that were in or respond to it in like full gentle sentences. Please, thank you, and you know, and and months ago I stopped doing that. Just like, you know, it's just not necessary, right? It's more more characters. It's like you forget, you know, this isn't truly a a a person. Um so, I think that's funny. That's interesting. I'm still on that kind of borderline where I sometimes am polite. Oh, I mean, I'm never Right. Right. Sometimes I'm depending on my mood, I think I can be short or I can be a little bit more polite.

I don't want to completely lose my sense of being polite to people, and I won't I worry that maybe it'll make me too blunt. And not not sounding like I'm caring about the person who I'm talking to. It's funny you say that. One of Yeah, I can be terse or short, especially if I'm tired. Never unhinged, but just not Would you please? That was so lovely, and you know, feed You know, I mean, it's like, come on. For people, absolutely. You know, one of the weirdest exchanges I had uh with AI was uh I was asking it to create some slides, and it was something it simply could not truly output, you know, those graphics.

In fact, what it did is it had some other layer it had to send the output to the file, but what would happen is it took so long to do it, the connection would time out. And it would say, you know, I'll have this for you in 5 or 10 minutes. And like, a half hour would go by, I'd come back, I'd make a cup of coffee, and I'm like, "Hmm. " And if I started to say, "Hey, look, I'm not trying to be on your case. " I wrote something like this. I was tired. I said, "You say 5 minutes, but a half hour later isn't back. Why are you telling me one thing and doing another? I just want to understand. Do you understand what I'm saying?

" And I really was was reasoning or rationing, and that was one of the most interesting exchanges I had. And it would say, "Oh, sorry, you're right. My bad, or I got you, or you know. " And then it finally explained to me the scenario. It said, well, in all candor the connection is timing out between that other module in the software in the background. And I'm like, okay, you know, going forward just tell me that. If you can't do it, don't don't tell me you're going to have it done in 30 seconds and a half hour later and I don't mind you telling me you can't do it. Right? Just tell me you can't do it. Yeah, yeah. It would be nice to if it was his onus.

Is that what I But it was one of the most interesting exchanges I I had, you know, about asking it this sort of either meta question or introspective question if if that's if it's introspecting, I mean, who knows, right? Even the creators don't know, but that can get very interesting. Yeah, yeah. So, um, we're talking about prompt engineering there a little bit.

When I was in Dubai last year, I just opened up the newspaper in the hotel and there's this article uh such and such Dubai you know, cuz it's run by um, this you know, the uh the families, the the the oil families and they're getting together about putting together an initiative to have 1 million prompt engineers done uh uh you know, through the uh through their system in the next 12 months. A million prompt engineer engineers. Um Sorry, I just got a missed phone call. Oh, no worries. Forgot to do that uh Such is life and 26.

Or 23 or 22 or Yeah, so you know, that whole you just brought up this I hadn't heard of that before, that three the three steps, but I think what you're saying is you you prompt and then you get it to critique it itself and um and and I mean, for people outside of a natural English language country, that's a big part of it is getting the the in the language English language part of it cuz predominantly these are based on the English language. Um so, you know, a lot of what they were teaching with is good English, but then also on top of that I guess the other half is actually how you do it, how you prompt it. Yeah. Oh, I'm in awe of people that know multiple languages fluently.

I've studied a few in school and I've probably forgotten more than some people, you know, know on a bad day. Um but when people are doing it through the lens of of a language that isn't their native language, I'm just in awe. That is a skill and you know, they may not think much of it if they've been doing it for years, but I just have so much respect and patience for anyone not doing it in their native language. That is remarkable. And you know, as English speakers we kind of win the language lottery. Right or wrong, it just, you know, and it can make us a little lazier about learning other languages. Um but um Yeah. Oh, so do you think the the legal system's being upturned by by AI?

Do you think it's affecting it negatively? Um Well, I don't know. What what I suspect is happening is because AI is so new and at least for practical purposes, not sci-fi novels of course, that most judges are of a certain age and not to be ages, but you know, they the profession itself and then on top of it that demographic on average. I stress on average cuz there's always exceptions. Has not so much embraced it and and may or may not be ignoring it and the law always lags too. So, I think we're very much at the beginning, although we're about, yeah, roughly 2 years in maybe. Um very much at the beginning and it will be interesting.

I I think what may be interesting even though it's one orbit out are the recent cases against social media that the biggies seem to be losing as of a week or two ago. Mhm. Um you know, meta in particular. Yeah. Um even though that is not identical, it's it's closely related and it'll be just so interesting to see you know, how that plays out. Especially with AI content moderation and um you know, things that are offensive about children and things, you know, to see how much AI can or can't do with that and I don't know, there's just so much Yeah.

that we're on the cusp of, you know. Yeah.

Well, I brought it up because I was talking to a guy that does web development, web designing stuff and I said, you know, it's making your everybody's going out and doing their own websites now and he said, no, no, not at all. He said, in fact, what's happened is my team have just become far more capable instead of having, you know, businesses you know, our clients at this level, we're now able to take on clients at this level. So, we're able to compete at a in a bigger game, you know. So, he's I know, he's got 10 people and he's able to compete as if he's a 50 team because they're so much more productive uh with these tools. Um is that what we're going to see with legal professions?

Do you think it's going to be more cost-effective, I suppose, to get legal um legal consult? Maybe. You know, it's funny you say that. There's something also counterintuitive I see where someone and AI certainly helps this as well as more forms being available online, where someone will bring me a form thinking it saves them money, but the truth is sometimes I work on a flat fee when it comes to something certain things. Mhm. And you know, it's my time, my expertise, my experience that's the primary thing more than the drafting itself, although that's in there.

But the irony is if they bring me a template that's not my trusted template, it may take me as much work or more to get acquainted with their template and usually there's more to fix than my trusted template. You know, that's reasonable. So, the irony is thinking they're helping. I mean, they may have educated themselves by reading it and studying up on it and that's that can be good. You can have better time with your advisor, especially if it's hourly, right? So, there's efficiency there. But, the irony is bringing me a template or a form that isn't particularly good and and isn't mine, um, doesn't necessarily save you money or save me work. Um, so I think there's an irony there.

But, um, I think it's going to push it's already it's pushing everybody up the value chain and I'm okay with that. Some people may not, you know, it depends on how experienced you are. If you're less experienced, you might have more to worry about. Right.

Um, if you're more experienced, you may have generally less to worry about. But, it's all about perception and I see it breaking down roughly by age demographic. Um, I see and and this is intuitive. Again, always exceptions, but on average I see younger people, you know, much more comfortable young startup founders using AI more and and questioning professional advice or it will fees, you know, classic fees.

So, there's pressure on that, of course, uh, to be more value-based. I mean, I already try to get away from hourly when I can. Sometimes you, you know, some things like something really open-ended, a dispute or something. Mhm.

Um, but then I I tend to see people more my age and older on average tend to be more like, well, you sound like you know what you're talking about. Your fees, you know, reasonable given the market. Okay. Um, so those are things I see. So, yeah, I think it'll push us all up on the value curve out of necessity. I think it'll put pressure on fees, of course. Um, Yeah. Yeah. That might be that might be a good thing, right?

I mean, if you can be more efficient as a business providing a service and and you can pass that on, then everybody wins, right? Yeah. We're going to hit one more topic before we start to wrap up, and that's um your interest in music. Mhm, okay. So, have you seen anything come out in in in the AI world which interests you that you're dabbling in in in the music field? Absolutely. I um I I sometimes work with a a writing partner and bandmate, and he uh even more so than I have is started using Suno among other tools. And what's so interesting about that is I think And you know, I see on forums for songwriters, etc. This whole Oh, you used AI.

You're you're crap, or you're fake, or And then other people are like, well And then there's some kids just doing amazing things. You know, for instance, it it it further helps people who say, well, I can't write songs cuz I don't happen to play an instrument. Well, your voice is an instrument. You don't need theory. So, that's a whole other thing. I mean, it can be nice, but you know. So, it's really helping people to express themselves. I definitely like that. Mhm. But of course, there's a lot of AI slop, as they call it, right?

Mhm. But what I see it doing And my dad's a visual artist, so is my partner, she is um So, they have their own experiences with it.

But I love the idea that it can help uh for me break writer's block. I mean, look, Mhm. Usually what I do is I come up with lots of music, and then I have to be really thoughtful, and it's harder work to come up with lyrics that I really like, right? And some people feel it's the opposite, but I have never had AI come up with something I'm like, oh, I'll just lift That's awesome. No. But it can create writer's block. Maybe it comes up with something that's awful, but I laugh at it, and it sparks another idea cuz the mind can bounce around. So, I see it as a someone It's like having a partner to bounce things off of and get you unstuck and just point you in other directions.

And I really don't see it as all or nothing, you know?

Right. Uh Um You know, like I've said, "Okay, I'm I'm at the end of this bridge. How do I get back to the chorus? " And you know, it'll know more music theory than I. And usually the straight music theory answer is usually it works, but it can be not particularly inspired. It can sound typical. Um But anyways, I I feel like that's one thing it can help with. But oh, I know what I was going to say. Rapid prototyping. So, for instance, I find uh even GarageBand to some extent, certainly you know, the the the fancier digital audio workstations to be kind of complicated.

And because it's not my day job, I barely have time to write songs and play music. I have an academic interest in, but you know, I'm not worried about if I learn the software, even though that would be nice to be independent. And a lot of of young people coming up are doing that. But my point is so something like Suno can help you with sort of rapid prototyping. If you keep prompting it to get it what you want.

So, what I'm in the process of doing is more or less writing the old-fashioned way, and then dropping in a Suno so that I don't have to either pull my hair out with the software to record it the old-fashioned way, at least for prototyping, you know, demos, or go to a studio that's, you know, 150 plus an hour across town. And so, you know, will it ever get exactly the way I want? Suno's maybe more likely than others cuz you can do a little editing along with the AI. But I think it can be a way that even if you're starting out like I do and writing the old-fashioned way with a pen and a notebook or pencil, it can help you ideate fast. You know, you can put in a prompt.

You can say, "Oh, that sounds overproduced. I want less compression. I want this. I want that. Or take out the strings. " You know, cuz I like really simple stripped-down production. And so, it can be it can just do that like that. You know, just you let it render and you come back, and there it is. Doesn't mean it's perfect. But then even if you take that to a studio and say, "Here's what I'm going for, but this or that different. " It's a time-saver. Wow. So, Suno is it's helps you with lyrics, it helps you with um Well, honestly, ChatGPT or Gemini can help you with lyrics, and it's amazing what it can do. Suno is just one, just the one I'm a little more familiar with.

Um it's an AI um It can do two things. It can create songs out of prompts. Right. But Suno's a little different than some, although I don't think it's totally unique, and it also has like the classic audio interface, you know, as if you were just editing on normal software, and it sort of allows you to go back and forth. Mhm. And that is um that's what's kind of different than most. I wouldn't say unique.

are you playing an instrument or singing into it, and then it it's kind of turning into a song? I haven't quite got it yet. What what is Suno?

What anybody can do is literally just write a prompt. "I want a punk song. Here are some lyrical ideas or theme. I want it to be this long.

" And and not be able to play a thing, and it will spit something out. Now, the less you give it, the wilder what comes out. But what I do is I sit there with a pencil and a notebook and I write a song with a guitar on my lap. Mhm. And then what I can do is and again, I'm just at the beginning of this, unlike my writing partner, I haven't done it fully yet. But what he's done is taken some of our ideas or my writing and drop it into Suno and come back with three versions and say, "What do you like? " It's a little bit like um Well, it isn't like just doing prompts for a website because that would be more like the first example.

But the point is you can literally start the old-fashioned way, but then use it for ideation or demo production in a way that is much faster than doing it the old fashioned way. Right. Right. You know, if you go into a typical audio workstation, I mean, don't get me wrong. I love the people that know it well, but like I get fatigued with 30 types of plate reverb, and I just I'm like it's just overwhelming. And and that's even GarageBand, the Apple version, which is actually pretty stripped down on the phone. I I do a little easier. But I want the software to get out of the way, and something alike Suno, there are others, Mhm. It it seems to be helping to do that.

That's a very good point, isn't it? I mean, ah, you said getting out of the way. I like the way you said that, because a lot of the things that we're seeing now with AI are taking the the learning curve, the complication, the difficulty, the barriers of entry, they get taken away, so you can get to making some progress. I tried writing music, cuz I was in a band as well, and I was at school. Just couldn't get over, you know, I wasn't a lyricist. I I could fiddle around with my guitars and do a few things. Had some ideas, but the idea of writing a song was just just a gap away, just too big a gap away.

Never could close that, and I think maybe if the tools like that had existed today, had existed back then, um maybe I would have made some progress. Yeah, and there are some real purists really poopoo people that are trying I love the idea that, you know, a kid in their dorm room or anybody of any age, any any demographic can literally say I have this melody in my head. You don't even have to really sing well, and it will let you create things. And and I think some people that are gatekeepers, and they have that law and tech and other professions, the gatekeepers don't like it, of course, and the the cork sniffers, as we say, but I am really neither.

I'm much more specially in music more than law. I'm especially just DIY garage basement band, and so I I'm no gatekeeper. So, you know, I'm I obviously love that accessibility. Wonderful. Well, one last question for you, Skyler. Is AI hype or help? It is hype. Mhm. And it is help. Okay. It is absolutely both. Yeah. Um it really is two sides of the same coin. Mhm. You know, it is um like I said, there's what I call intelligence washing or AI washing where everything has AI baked in and Mhm. You know, and like, okay. Uh just like I don't even know if prompt engineering will be a career title in a couple years. It's not that we won't all be prompt engineering.

It's just like who says I'm an e-commerce expert anymore? 20 years ago, that was a niche, that was a title. Yeah. Yeah. Anyways, it's getting baked in. So, I think there's a ton of hype. Um but, you know, it it helps me every day. It helps a lot of people. I'm trying to get my dad on board with it as a visual artist, and he hasn't bitten yet. Mhm. Um but, you know, the idea of remixing, like I love the idea of putting in your work, and it could be true for writers, you know, not just traditionally creative, but then it outputs something that you react to. Mhm. And you react back and forth, and when you come out of it, this mix of machine and human is terribly interesting. Mhm.

You know, and then it gets into issues. Should we have a disclaimer? You know, like France and Israel, I believe, have laws that if you Photoshop a model's image in an ad, you have to have a disclaimer or whatever. You know, there's issues like that that we haven't resolved yet. Do we like I'm tempted to do that, like disclose in the liner notes of an EP, you know, what was AI used and how? Mhm. I'm happy to be clear about that, but I'm intrigued by the interaction, the back and forth. Mhm. Um you know, to put something that I created and then say, AI, go wild with it. What do you come up with? And I may absolutely hate the output, or I may love it.

So, that's why I say it's help, it's hype. In the wrong hands, it can give you overconfidence. If you have a rough idea what you're doing, it can really be a help. So, it's absolutely hype and it is absolutely help. I love that answer. That's great. You really unpacked that for me. Thank you very much, Skyler. It's been a real pleasure. We could go on and on. It's been a really good and long podcast. Thank you for being on the show. Maybe we can have you back soon when we got some more things to talk about and we certainly will have. That sounds great, Peter. I'm very like you passionate about these these topics. So, thank you for having me. Well, my pleasure. Thank you. Bye-bye.

Well, I hope you enjoyed that chat with Skyler. He had an interesting few takes on how he uses AI and he recommends certainly using AI to understand contracts. Having it put into simple terms that you can understand and read and that I find that very helpful, too. But, he also warned about going too far with them, getting it to generate contracts because specific words mean specific things in the legal framework and they may mean something different to you. So, be careful with that. He's also someone who I met for the first time who uses Sumo, which is a songwriting tool. They talked a little bit about that and how helpful it was for him unblocking the idea of creating songs.

So, it made writing songs accessible whereas in the past perhaps it didn't. So, that was some great tips from uh Skyler. Look, if you want to know more about Skillion, please follow the links in the description. If you want to know more about Skyler, we'll leave those links in the description, too. If you could subscribe to our channel, that would be great. Press that like button. Hope that helps and thanks for watching.