In this episode of The Geek in Review, we interview Josh Kubicki, Director and Professor of Legal Business Design Hub and Entrepreneurship Program at Richmond Law School, and the creator of the daily newsletter Brainyacts. Kubicki discusses the speed of change in legal technology and how it is now measured in days and weeks, rather than years and decades. He also talks about the need to embrace technological changes and how he created Brainyacts to force himself to learn about generative AI in the legal profession. Kubicki shares his experience of writing 100 consecutive days on generative AI and how it has helped him stay accountable and take a deep dive into the terrain. He also discusses the reaction from the industry side of things and where he digs in to find good content.

The advancements in generative AI tools like GPT, Bard, Bing Chat, and a slew of new products that seem to be launching daily has overwhelmed many of us. Resources like Brainyacts helps put a little bit of stability in this time of rapid change. So, take advantage of Josh’s hours of work on a daily basis and subscribe and read the very practical newsletter.

Links:

Listen on mobile platforms:  Apple Podcasts LogoApple Podcasts |  Spotify LogoSpotify

 

Contact Us:

Twitter: ⁠⁠@gebauerm⁠⁠, or ⁠⁠@glambert⁠⁠
Voicemail: 713-487-7821
Email: geekinreviewpodcast@gmail.com
Music: ⁠⁠Jerry David DeCicca⁠

Transcript

Greg Lambert  0:07  

Welcome to The Geek in Review. The podcast focused on innovative and creative ideas in the legal industry. Marlene Gebauer usually gives that part of the introduction, but she cannot make it today. So I’m Greg Lambert flying solo. You may have remembered last week or the week before, we talked about just the speed of change in legal technology and how we used to measure that in years and decades. But if you’re really lucky, we get it down to quarters within the year, but the way things are these days, it’s measured now in days and weeks. And today’s guest is helping us kind of keep up to date with all of the changes that are going along in the advancements of generative AI and its direct effect on the legal industry. So we’d like to welcome back to our show Josh Kibaki, Director and Professor of legal business design hub and Entrepreneurship program at Richmond Law School. And, more importantly, the creator of the extremely popular newsletter Brainyacts, which is a daily update of all the changes and happenings in the legal generative AI world. Josh, welcome back to The Geek in Review.

Josh Kubicki  1:21  

Hey, Greg, thanks for having me. It’s great to be here.

Greg Lambert  1:24  

You know, there must be something in the water in rich man, at least when you visit there from Cincinnati when you make that drive every week because we just had Roger Skalbeck and Ashley Dobbs from Richmond on the show a couple of weeks ago. So and I looked back, Josh, you were on in back in October of 2021. And I have to say, I reviewed the transcript. And you know how many times we mentioned AI in that conversation? Zero times. So does that surprise you? 

Josh Kubicki  1:58  

No, not at all. Not at all. 

Greg Lambert  2:00  

Yeah, it’s just man, we are at breakneck speed here. It’s so crazy. But the one thing we did talk a lot about was your work with design thinking with your students and the need to, you know, how you were teaching them to embrace technological changes. So you know, you had the platform there. And it looks like you’ve been able to build on that platform with your previous experience. And then just catapult right into this daily Brainyacts newsletter. So I got to ask you what, you know, what motivated you to take on this new daily task, which looks like a lot of work from my side of things?

Josh Kubicki  2:38  

Ah, yeah, it is a lot of work. And to be completely honest with you, I knew I saw the power of generative AI early on, and like most of us were working, and we don’t have time to learn. And I was thinking, well, how can I create something that’s going to force me to learn every day and show up? So I stay on top of this. And I was like, I’m going to publicly commit to write 100 consecutive days on generative AI in the legal profession, I have to do it, I have to show up every day, and I’ve got to write good content. And so that was it. And I’m on day 40. I’m reading issue 41 As I speak, and it’s a lot of work, but it’s keeping me accountable. And it’s taking me very, very deep into this terrain.

Greg Lambert  3:24  

Yeah. And this, this isn’t like 100 business days, five days a week. This is seven days a week for 100 days, right?

Josh Kubicki  3:32  

Absolutely. Yes, it’s every single day I am sitting down. And it’s probably two to four hours of work, because I’ve got to sort through and we’ll probably get in as all the noise. There’s so much hype, there’s so much noise. There’s a lot of BS out there. And I can’t serve I know, right? Shocking. And I can’t serve that up. So I a lot of that time is spent testing and going deep into the new stuff and just figuring out what’s what and what’s not. And so yeah, it’s been a journey.

Greg Lambert  4:02  

I want to dig into that here in a minute. But first, I wanted to just ask you, how’s the reaction been from the industry side of things? How have people reacted to the newsletter?

Josh Kubicki  4:15  

From the newsletter perspective, it’s been it’s just been amazing. I don’t know how many folks out there listening, have their own newsletters, or subscribe to newsletters, but it’s a, it’s a tough game. You’re competing for people’s inbox time. And we take that, especially in the legal domain, it’s very precious real estate. So to get someone’s attention in their inbox and compete, it takes a lot of effort. And I’ve been very thankful to the readers and subscribers and they’ve been spreading the word and listen, you know, you do a podcast, your goal is to put good content out. Right, that’s all I worried about every day is good content, because people will share it. They’ve been very, very helpful. 

Greg Lambert  4:57  

So where do you dig in to Find good content?

Josh Kubicki  5:02  

So, a number of places. So Twitter, obviously, I’ve got a number of searches and lists on Twitter, Twitter is, yes, it can be a dumpster fire. But if you train the algorithm by usage and all that you actually start to curate a good feed. So my feed and Twitter now serves up things for me, I run specific searches in Twitter, because a lot of the breaking news in this space, a lot of the tools, a lot of the startups are Twitter, first news, right? So I do set up Google Alerts. So I’ve got things in my inbox. Now, readers of Brainyacts are sending me DMS and messages and all that. So it’s a broad array, because you don’t no one’s got a lead on sourcing this right now. It’s kind of anyone’s game and news could come out of anywhere at this point. It’s kind of crazy.

Greg Lambert  5:54  

Yeah, it is really hard to keep up with. So I’ve, I, for one have been appreciative of you know, whenever I see that newsletter hit my inbox, and it could be early in the afternoon, or it could be late. Later in the day, I noticed that it may take you a little bit longer some days to get it out than others.

Josh Kubicki  6:17  

It is. Well, it’s the fight for quality, right? It’s I won’t press Send until I’m secure that it’s good content. So yeah, I try to get it by 6:30pm. Eastern Standard Time. That’s my goal. Sometimes I miss it.

Greg Lambert  6:33  

All right. Well, it’s good. It’s good to have goals out there. 

Yes, exactly. 

So while you’ve been pulling this all together, and again, I know you’re building on top of just the content, it looks like you’re also building a community as well. So what’s been you know, what have been some of the things whether it’s, you know, the products that you find, or the advancements that, that keep popping up every day, or even your personal interactions? You know, what, what surprised you in this first 40 ish days of Brainyacts?

Josh Kubicki  7:11  

So I think one of the big surprises is just how enthusiastic everyone is. Now, I think you’d say, well, of course they are. But they really are. You know, when we think about some past technology developments, there’s like certain, well, you had to be a geek to get certain things, you know, like, like, no code, or all lawyers must learn how to code, right? We sort of lived through that several years ago. Because ChatGPT, in particular, sort of immediately became available and democratized access. It was free, it still is free. It’s somewhat intuitive, people are still a little bit intimidated with that first prompt window, they don’t quite get what I’m supposed to do. And they kind of treat it like a Google search bar.

Greg Lambert  7:58  

Two-word searches?

Josh Kubicki  8:00  

Right, exactly. But I think once they, I walked my wife through my wife as a, as a president of a nonprofit, and she was creating this campaign to create awareness for certain congressional leaders. And she painstakingly going over this letter as you should. And I said, Hey, why don’t you just ChatGPT it? And she’s like, what? I said, Yeah, let’s put your letter in, let it know who you’re writing to, and just say, Hey, make the sound polished, pithy, you know, attention deficit audience, usually the staffer, so get the point, and it rewrote it and its uses like, Oh, my God, like, I cannot believe I’m seeing this. I said, Wait, there’s more. Now tell it to be unconventional and creative in it, and it spun up another version, and she was like, Oh, my gosh, so the surprise to me, it’s not so much a surprise. But it’s like seeing people’s joy, like their actual joy when they see the returns of this, that to me is really surprising to see technology create joy is awesome.

Greg Lambert  9:09  

It’s been a while since we’ve seen technology create joy, I think. So it’s probably since streaming services, maybe? Yeah, yeah, exactly. Exactly. So as you’ve been going through and doing this day in, day out and weeding through all of all of the things that are out there. Have you had any big aha moments? Or on the flip side of that had any Oh, oh, crap moments that that you can share with us as you’ve been pulling this information together.

Josh Kubicki  9:43  

Honestly, one of the big aha was and this is probably the biggest challenge for me right now with Brainyacts. But for most people, and that is we’re kind of, we get it, those who’ve been exposed to it, get it, they see the magic of it. There’s lots of news about it, but people still struggle Look for use cases. And so at first everyone was rushing to like prompts like, You got to design a prompt this way. And then we had all these people, some huckster, some legit, like selling prompts, and all those kinds of stuff. And I bought into that a little bit, because I didn’t know. And so I was like, okay, and then I just learned through prompting, the aha for me is, it’s not so much the prompting, yes, you have to know how to prompt these things. But it’s actually the use cases coming up with the different ways to think about using this tool. Because, you know, like, the marketers were on this the first part, so a lot of the stuff was how to create more content, which is not really useful to legal professionals. I mean, it should be for marketing, but like you run out of use cases. So it’s like, no, my job is to like, really help you think about different use cases and the business side of law on the practice side of law, and all. So that’s really been the thing that most readers are Brainyacts. And even, you know, now I’m getting calls from like big law firms to come on. And like talk to partners about what generative AI is, and a lot of it is they’ve read about what it is they just want to know how to use it. So that’s really been the AHA, how do I use this? To answer your second question? Like big O’s, we all know it hallucinates, it makes things that we know that like, if you don’t know that by now, you should like, do not take it at face value. The big O for me is more about where we’re going. And my observations on the tools. And pardon me this, this, this might take a minute for me to unravel here. But

Greg Lambert  11:37  

the floor is yours, Josh. 

Josh Kubicki  11:39  

If you think about ChatGPT, let’s just focus on that tool. Because again, there’s so many tools. That interface that we use right now is a direct portal into that underlying technology. There’s no interface between us and what they call the LLM, the language learning model that sits behind ChatGPT. That does what it does, like we’re at the coalface of this technology, when it first was released to the public, people being people did some pretty odd things and showed a lot of the flaws that I won’t share here because I don’t want people to repeat them. But just we’re asking ChatGPT to describe things that we probably shouldn’t be sharing on Twitter and all that kind of stuff. And jailbreaks there’s a term jailbreaking. ceppi ChatGPT. And we know that the EU, the US, you know, the UK, we know that Italy banned it, Germany’s on the on the fence about banning it. It all has to do with privacy, and safety. So all these nations are considering this. And it’s because of privacy and safety reasons. So Oh, initially, when I started Brainyacts, I put in a master services agreement, right? Because I was like, Okay, well, let’s see what it can do. And I literally cut and pasted a master service agreement between me and a client. So and the moment I did that, the lawyer in me like I was like, oh, no, like, Where does this go? Where is this information now? And I quickly learned, oh, no, I can’t do that. I’ve got to anonymize anything I put into ChatGPT. Right. And it’s, and it’s because these privacy and safety concerns. So that was one Oh, the second one, and then I’ll turn it back to you is, in response to this open AI, the parent of ChatGPT has been correcting the model has been blunting the model. So a prompt or use case that worked three weeks ago, may not work today, because they’re constantly changing the model, the back end and how it returns responses. And so we have to be careful, I have to be careful about that by not over promising that a use case or prompt will work because it might not. And I think that’s something that we have to talk more about. People don’t realize that they are literally changing it day by day, to make it more safe. And you know, to be more clear about disclaimers and all that kind of stuff.

Greg Lambert  14:03  

That makes sense. I’ve kind of had this theme since January, that using tools like this really takes imagination and creativity. And just really a different part of the brain being used than most lawyers and legal professionals have have used their entire career. Are you finding that the people are being creative? Or do you think that that’s kind of a roadblock for this industry in this type of technology?

Josh Kubicki  14:38  

I think you’re 100% Right. I think the analytical brain, right? So if we think about traditional search, or how we use online tools, it’s a very right brain. Activity. Think about how lawyers do searches in in Lexis and Westlaw and even Google like we’re thinking, what’s the core like, what’s the core or what I’m trying to find? And how can I be specific? And you’re 100%? Right? The generative AI is a left brain activity. You, you it’s not intuitive, like I said, it goes back to those use cases, how can I? How can I think differently? And how do I how do I create a dialogue? Which is actually what you’re doing? How do I create a dialogue with this technology? And communicate something that I might not even know that I’m looking for? Which is, it’s a really weird concept. I don’t know what I’m looking for. Can you help me identify what I’m looking for? Can you help me think about how to phrase what I’m looking for? And you’re right, people just aren’t used to, we’re not used to interacting, because that’s something you and I would talk about over a beer or over coffee, like Greg helped me, like, I’m thinking about this thing. I can’t crack this nut. We’re not used to thinking about technology that way. And that’s what ChatGPT is doing.

Greg Lambert  15:53  

Yeah. And are you for your newsletter? Are you just using the free interface? Or are you jumping into the paid version as well?

Josh Kubicki  16:04  

So so the designer in me for a long time, I have resisted paying because I was like, No, I have to use what my audience is more than likely going to use I have to stay completely centered, empathetic on what their challenges are going to be. The simple fact is, I was learning and running so many different versions of use cases and prompts, it was very slow. So then I bought the $20 GPT plus, and use the 3.5 legacy. So I still use the 3.5, which is the same as the free version, slight differences, but it’s faster. I will use GPT for for like my own creative process, you know if to try to organize my thoughts about a number of things. But I’ll always use 3.5 for the newsletter.

Greg Lambert  16:51  

Any real noticeable differences when you use four versus the 3.5?

Josh Kubicki  16:58  

Yes, I mean, four is just lightyears ahead of 3.5. It’s more eloquent, it’s more sophisticated. It’s more complete, understand your prompts. It’s more intuitive. I could go on and on. There’s a stark contrast. Again, I’ll say a few months ago, that contrast was massive. It shrunk a little because as they work safety into it and protocols and disclaimers, they’re starting to blunt the edge a little bit. And so now they’re kind of coming closer

Greg Lambert  17:29  

together. Okay, so I know that this is 100 Day event for you. What do you hope to achieve? As you get closer to the 100 days? And then is 100 days going to be it? Or is there plans? Beyond that? Do you even know now?

Josh Kubicki  17:51  

I’m on this journey Greg! I don’t know. I’m a the journey is that adventure, not the destination kind of guy. So right? I’m gonna take it where the community, my subscribers, want me to take it were sort of my nose for this goes, I will do something at the end of this because I think it’s so profound. And I don’t want to be I’m not being, you know, hyperbolic, or anything like that. It’s so profound of a game changer for knowledge workers, whether they’re in the legal profession or not. There’s so much that you can do with this tool. And that’s been my message to like, big law partners. It’s like, first of all, it’s not a threat. It’s not a threat. It could be, but no one’s gonna let that because there’s so many problems with relying on it for legal advice. Like we all know that, like the novelty has worn off. Right, okay. But it’s now being baked into tools and apps that are more specific. We know Dan Katz, right, who’s another one who’s been deep into technology for years, like, like, for years before it was all the buzz, you know, he’s got a new LLM, a new ChatGPT Like, model coming out, you know, soon. We’re starting to see the the rise of the specialized GPS, right, the distinct and those are going to be safer areas for us to be in. So that’s developing that ecosystem is being built out rapidly right now. So at the end of this 100 days, I don’t think there’s any shortage of things to cover, because we haven’t even talked about Bing chat. We haven’t talked about Google barred. You know, and what Google’s doing with Deep Mind and they just released Maggi, which any Google consumer out there that uses Gmail or ducks. They’re going to be embedding. They’re not going to ask you they’re going to be embedding this into their tools. So it’s here, so I don’t think there’s gonna be any shortage. Have of use cases for us on the practice or business side. So I don’t know what I’ll do, but I’ll be doing something Greg, definitely.

Greg Lambert  20:10  

Before we get to the crystal ball question, let me ask, what about your students have one? Do you find that they subscribe to the newsletter? Do they give you any feedback? Do you and have you used this at all in in your class and teaching to the students?

Josh Kubicki  20:26  

So I yeah, yes, I Absolutely. We don’t have a policy at the University of Richmond yet, for or against or restricting it. Like, I know, some universities have that we don’t I encourage my students absolutely it. How can I know I felt it was a disservice and disingenuous for me. So they have a paper coming up their final paper, and I said, you’d be a fool not to use GPT. Now, it was on me as the professor to design the assignment in a way they couldn’t use it for all of it, right? You just can’t. You’ve got to be thoughtful, as a teacher on this kind of stuff. But I want them to embrace it. I want them to understand it. To answer your question. Yes. So I’ve also created a free five day email course it’s introduce introduces knowledge workers to ChatGPT starts with the basics five consecutive days. My students, I didn’t even tell them about it, they found it because they follow me. I got students in that I’ve got students reading Brainyacts, one actually just posted on LinkedIn today, like, you know, for any law student out there, you’ve got sort of read this. So yeah, it’s really it’s, it’s really nice when they do that. But I’m, you know, there’s so much value in it for them to understand this stuff as well. So it’s kind of a no brainer.

Greg Lambert  21:45  

Yeah, I’ve seen the cartoons where it’s, you know, the student who goes, Oh, I just turned in this assignment using ChatGPT. And the teacher is like, well, that assignment was written by ChatGPT. And now I’m going to grade the the exam using ChatGPT. So it’s like this an endless loop of, you know, opportunities to use it. But I think you’re right, I think the novelty is just like with anything and technology, whether it’s games or this, you know, that first week of playing around with it is like super excited. And then all of a sudden, you kind of get into that rhythm of you know, either it doesn’t do what you want to do, and you never use it again, or it just becomes a part of your overall process. And I think this, this is one of the few things that has such a speed of becoming just a part of the process for the legal industry, which, you know, again, is notorious for being five years behind everybody else. And it seems like, that’s just not the case, the time. So I’m pretty excited for how it’s just kind of weaving its way into the daily interactions. So well, let me get to the crystal ball question. And I really wanted to split this into two. We kind of talked about the the first 100 days, but what do you see on the horizon is a change or challenge as we go forward? And let’s just move it out a little bit and say to the end of 2023, and then I’m going to when you finish that, then I’m going to ask you, where do you see going beyond the next two to five years? So let’s let’s start with this year, what was some challenges you think we’ll run into this year?

Josh Kubicki  23:35  

So one is just regulation. I think policy and regulation, we know that Senator Chuck Schumer is already sort of leading the effort in Congress to define us regulations, again, not know so much on what we can do with it, but what these models are trained on the data, the privacy issues, and and that will impact the quality of the responses that we get from these things. Right. So in the EU, they’re they’re already way ahead of this and other countries. So I think, by the end of 23, most developed countries will have a stated regulatory policy and beyond some level of enforcement or monitoring on this as we speak. So that will fundamentally change how we are building tools with this, the quality of the responses, what we can expect from safety and privacy. So sort of sort of the meta level, if you will, the micro level on the day to day. I think it’s just going to continue we know you just sort of said it as well. It’s seeping into our daily lives. It’s it’s going to be on our mobile phones whether we like it or not. Yeah, Surrey it we have yet Apple has been really quiet. They have not said a peep about any of this now they’re always worried about this stuff? We know Amazon just released, you know, the CEO released his annual letter last week, he was all over this. And he hinted, like we’ve got more, I’ve got more to say. So we’re gonna see the more consumerization as well as more enterprise specific plays, I mentioned Dan Katz and his effort in a space as well. So I think the rest of 2023, we and your listeners, continue to learn it. But sit back and sort of watch things unfold. I don’t think it’s going to change our lives drastically. But there’s a lot of opportunity to learn and a lot is going to be happening between now and the end of the year.

Greg Lambert  25:41  

Yep. Oh, how about further out? Were, you know, you’re looking at 2025 and beyond where? Where do you think we are by that time? Is there even a way to guess?

Josh Kubicki  25:53  

Well, I will, no, but I’ll try. I’m going to constrain it. I’m going to constrain it to the legal profession. And I think a lot of there’s been lots of varying opinions on what is the what is actually going to be the impact. And we at first we saw robot lawyers and replacing lawyers. And now the notorious Goldman Sachs report that said, lawyers are the most at risk. No offense, I’m sure very bright minds at Goldman Sachs. But I don’t believe it for one second. This is doing one thing, it’s creating more complexity in our daily lives. It’s democratizing access to information, it’s, it’s helping people become aware of different things they might not have in the legal context. For instance, one of the things I do now, when I’m testing out all these tools and apps, is I actually go and cut and paste the privacy in terms and conditions for each app, I try, put it in ChatGPT, I’ve got a standard prompt, that gives me red flag yellow flag green flag for any sort of provisions. And I’ll tell you, there has lots of apps I won’t use once I run it through there, because you can tell these T’s and C’s have just been copied and pasted from something else. There’s not even a named entity in them or anything like that. So and that’s me. But I’m becoming more aware, as these tools continue to make everybody more able to understand legal rights. For instance, I’ll give an example. So think about a neighbor who has just put up a fence. We know if anyone’s ever done that going and understanding your zoning laws are it’s a horrific user experience, finding them, reading them, and applying them are almost impossible, you usually have to get a zoning commissioner or a contractor who knows how to do that. Well, the generative AI, you can solve all of that right now with tools that are out there right now, not just ChatGPT. But you can take a picture, it’s context aware, it can read context and tell you what it’s looking at. So how is that good for lawyers? Because that doesn’t displace lawyers, you’re now equipping everyday people with the awareness that they have more legal rights and more legal tools than they probably see we’re, you know, ignorance is bliss. We go through our lives, think about how many contracts are written every day by businesses with no lawyers eyes on them at all, right? This happens all the time. If people become aware, and they’re able to use tools that generative as giving them to understand I have legal rights, I can do this, I can do that. That’s not taking work away from lawyers, because lawyers aren’t doing that work. Anyway, today. What it’s doing is memorializing relationships. It’s planning in people’s brains, I do have legal rights. So when things go sideways, instead of just shrugging it off and going, well, that’s life, they actually would probably pursue a lawyer for that, right? So I think this nets very positive for the legal profession. I think we’ve got a bridge to cross here in the meantime. But overall, when you introduce complexity, when you equip citizens with more information and all that, it’s not like they’re not going to get into arguments. It’s not like they’re not going to need legal services. Right. So I do see this as an absolute blessing for the legal profession, quite honestly.

Greg Lambert  29:27  

Well, good. I like that prediction. I like it a lot. So. All right, well, Josh Kubicki, author of the outstanding Brainyactsnewsletter, and Professor of Law at the Richmond Law School. Hey, thanks for coming back to the show. This newsletter is great. how can listeners find you and how can they subscribe to the newsletter?

Josh Kubicki  29:51  

So the best place is to just go to JoshKubicki.com. Everything is right there. You’ll see the Brainyacts down below and just hit that you’ll be able to subscribe to the Brainyacts or find me on Twitter, at @JKubicki.

Greg Lambert  30:07  

So I’ll make sure that we put the links there on the show notes. And of course, I want to thank everyone for tuning in and listening to The Geek in Review podcast. If you enjoy the show, please share it with a colleague. Josh, I know you’ve gotten over 1000 subscribers now to the newsletter. Believe it or not, we’re reaching right at 1000 subscribers to the show. It only took you 35 days to do it? It took us almost five years.

Josh Kubicki  30:38  

Got to do it every day, Greg, it’s got to be released every day. What can I tell you?

Greg Lambert  30:41  

That’s true. Well, maybe I can talk Marlene into doing 100 shows and 100 days and tell you right now that that is not going to happen. So but if you haven’t subscribed already to the show, you can click that button that follow button on Apple podcast or Spotify, we’d appreciate it. And we’d love to hear from you directly. So feel free to reach out to us on social media. Marlene can be found at @gebauerm And I can be reached @glambert on Twitter. And if you’re a technophobe and you didn’t like anything you heard today, you can always just call us on the phone at our geek and review Hotline at 713-487-7821 and as always, the music you hear is from Jerry David DeCicca Thank you Jerry and thanks again Josh!