Cybernomics Radio!

Is SaaS Dead? The Hidden Cost of AI Agentic Companies

• Bruyning Media

Share episode

Josh Bruyning and Richard Stiennon explore how AI agents are fundamentally changing the SaaS industry and traditional software models. They discuss the shift from feature-based to outcome-based solutions and why companies must adapt to survive.

• Richard Stiennon shares insights about Security Yearbook 2025 and IT Harvest's database tracking over 4,340 cybersecurity vendors
• Introduction to Compliance Aid, a disruptive platform supporting 366 global compliance frameworks through conversational AI
• Analysis of why traditional SaaS companies must incorporate AI agents to remain competitive
• Discussion of the "agentification" trend: adding AI capabilities to specialized software platforms
• Exploration of how specialized knowledge maintains value when combined with AI capabilities
• Deep dive into AI's impact on employment, particularly for future graduates
• Warning for educational institutions that resist AI adoption instead of embracing it
• Comparison of the AI revolution to other technological transformations like desktop publishing

Check out IT Harvest at it-harvest.com and connect with Josh Bruning on LinkedIn to stay on the cutting edge of technology and AI.


Josh's LinkedIn

Speaker 1:

Welcome to this episode of Cybernomics. Thanks for tuning in spending your time with us. I'm your host, josh Bruning, and today I am talking with Richard Steenan. The one, the only, the man, the myth, the legend. Richard Steenan hey, richard, thanks. Hey, josh, thanks for having me. Congratulations on the Security Yearbook 2025. Every copy sold out at RSA correct.

Speaker 2:

Correct. Yeah, all the ones in the library or the bookstore. And then various vendors bought 250 copies to have me sign and I signed them all and people took them here. It is Awesome.

Speaker 1:

How many vendors are we up to right now?

Speaker 2:

Yeah, 4,340. But quite a few of those 500 of them we've archived. In other words, we go through all of our vendors every week, or LinkedIn's page is not there or points to an acquirer, maybe whatever reason that I get an alert and then we decide whether or not to archive the vendor. So it's usually because they've been acquired, but sometimes they just disappeared.

Speaker 1:

Yeah, yeah, and I know sometimes they fall through the cracks like the really new players and the disruptors. Yeah, is Compliance Aid on your radar? Is that in IT Harvest?

Speaker 2:

I will have to check. I don't even know if I'm allowed to talk about them, because they're like there was a time my brain was so pliant that I would remember, if I ever, because I have to enter the original data. We have compliance cow, compliance scorecard and compliance risk. All right, what'd you say? Compliance aid?

Speaker 1:

Compliance aid, compliance and then A-I-D-E Nope, we do not. All right, you're going to want to put them on your radar because they are disrupting things in the GRC space and even beyond GRC, because Randy Blasick, the co-founder there at Compliance Aid, created and have had him on the podcast. You and I had talked to him before and things have changed so fast over the last couple of weeks, even since you and I and he all got together and talked you and I and he all got together and talked and he can support now the compliance aid platform can support 366 frameworks, so like global frameworks. So a customer can just pop in and go hey, what do I need to be compliant with across the globe? And can you crosswalk that to all my evidence and all my controls? It's like, yeah, sure, we can do that in the afternoon. Can you generate policies? Yeah, sure, we can do that.

Speaker 1:

Go through the feature list of any competitor like Drada, vanta, aptiga Audit Board, whatever it is All gross. Yeah, aptega Audit Board, whatever it is. Allgress yeah, go through their feature list and you can just go to the compliance aid. It's like a chat bot, like ChatGPT or something like that, and you can ask it hey, I need this output, I need a risk register, I need a framework assessment, I need X, y, z and it'll just spit it out and I'm like holy crap.

Speaker 1:

Which brings me to the topic of today SAS is dead. I mean, llms and agents have not just disrupted the SAS model or the SAS industry, but I think it's on the verge of just obliterating it, because every traditional tool or solution out there cannot compete. They're not as flexible, not as customizable, ai is inherently fast and there are governance issues which we can talk about, and there are the hallucination issues and reliability issues. There are privacy issues, there are security issues, there are implementation issues and usability issues, which are all things that SaaS have conquered and are really good at. So, richard, what does it mean in 2025 to say that SaaS is dead?

Speaker 2:

Yeah, I would couch that in terms of look at it if you have SaaS, you better be looking behind you for what LLM is going to displace you quickly. So incorporate it right. You've got to change to survive. You might have to drop your price by a factor of you know two one-tenth what you were charging before to keep up. Just think of all the poor startups when ChatGPT first came about and they discovered, hey, you could embed a whole bunch of data in a RAG model and turn it into a knowledgeable LLM based on your stuff. And then, all of a sudden, chatgpt says oh, here's your own GPT, make your own, just upload your data. I created something I call Socrates and I uploaded all my books and papers and articles. Uploaded all my books and papers and articles, and so now you can talk to a bot that has all of my writing and maybe will respond as if it was me.

Speaker 1:

Okay, Big win for anybody who wants to be an industry analyst Sure. I wish I had that like three, four years ago. Sorry, ed, don't we all?

Speaker 2:

So yeah, so there's that challenge Like right now. It's like when you say SaaS, you have to think of Salesforcecom, right? The kind of the anchor for all of SaaS. And do you really need it? Right, it's horrendously expensive. You could build your own in an afternoon using a little help from ChatGPT or Cursor or whatever you like to use, and it'd be customized just to what you need today, right? Then you add a feature later if you need a feature. So, and maybe use no code for the user interface. So, no matter what, it's going to be cheaper than getting a five-seat license to Salesforcecom. So I wouldn't say SaaS is dead, because it would be a long time before that huge building in downtown San Francisco is leased out for apartments or whatever. But I think SaaS is completely challenged. It's definitely not the hot thing anymore. The startup world has shifted immediately to doing all this chat, gpt stuff.

Speaker 2:

There's another corollary to that question and it is. This came up actually this morning. So a friend sent me a link to a product called NextRose. So NextRose, you know, it's immediately apparent what it does as soon as you see it. You just go to NextRosecom something like that, and you can upload any file so you can input a URL, you can input PDF, you can input an image of data in any form and it will turn it into a spreadsheet. Just a minute, and what a brilliant idea, right Well?

Speaker 1:

that's great. I'm glad that somebody's specializing in that, because I try to do the same thing in ChatGPT all the time. It does not work Right. I mean it may work sometimes, but it's not reliable yeah resolution issues something.

Speaker 2:

But they've, you know, I don't know what they're using on the back end. It could be a combination of things. They could be doing pre-processing, you know, to just extract the text and then later on do the columns in rows. But I tested it out. So Jeffrey Dubs is the VP of sales at Gartner for the Gartner Security Conference which is coming up next month, and he's been posting an infographic of all the sponsors. We track all the sponsors for most of the big conferences. So I grabbed an image, uploaded it to JustRows and it immediately created a spreadsheet by category. So you know, each column was a different category application, threat, intelligence, et cetera, and then all the vendors in each category. The way it was, it is infographic. And then I said, well, I just want a row of company name and category. And it did that, just redid it. For me it was simple interface, so simple.

Speaker 2:

Now my friend who pointed it out had made an investment in them and my question is you know, if you put on your curmudgeon hat and you're an investor, you immediately go. Oh, you know, chatgpt will replace this. You know, within months People will use ChatGPT because they'll be so good at interpreting what they want and doing stuff. So the question remains is that true? You know, here's guys with a fantastic product. They don't even have a pricing page, so I can't even buy it, they just use it. I would pay for it. Yeah, are they doomed? And that is a very interesting question to me, because I have the same question Am I doomed with IT Harvest? Because with IT Harvest we've taken you know what originally was just the directory in Security Eurobook of 4,000 plus vendors, and originally with manual processes and now completely automated with the help of, you know, manus and Perplexity and Anthropic and OpenAI, we can populate a database with information on all products and align all those products with NIST and whatever and just make this deep and rich tool for exploring the industry. And so, and the fear is, and I totally believe, and I use it in our strategic thinking that the foreseeable future and it all takes two years to get to be 100 times better.

Speaker 2:

So right now, if I ask any of those Manus, perplexity to tell me you know, list all the cybersecurity companies in Canada, they can kind of get 30, right, we've got 138. You can't tell the difference between Fortinet, which is a big player in Canada. They've got a lot of employees there, but it's definitely not a Canadian company, right? They're in Sunnyvale, so it messes up on a lot of stuff. In other words, a lot of work.

Speaker 2:

If you wanted to create a directory of cybersecurity, it's such a concern to me that we are seriously considering making this the last year that you can get the directory in the security yearbook that directory if you just OCR this and I shouldn't be saying this, but anybody could buy a copy right now OCR this. They'd have the list, they'd turn it over to ChatGPT and they could make tools worth millions of dollars. You know it helps to be an expert in the cybersecurity industry, so you know what you're doing. But you know ChatGPT is probably more of an expert than I am now. So that's the opportunity and the threat. Chatgpt, if it's 100 times better, can categorize vendors. They can't do it now, right? They can't tell an errand between. They wouldn't even know. They can't do it now, right? It can't tell an error in between. It wouldn't even know what a digital risk protection vendor did, right.

Speaker 1:

Right, so you're saying that the more granular you get with the categories, the harder it is for ChatGPT to catch that?

Speaker 2:

Yeah, and it doesn't have the same categories, of course, as an industry analyst comes up with, because it's swayed by what the vendors say.

Speaker 1:

Yeah, yeah, and I think that that's one of the greatest selling points of IT Harvest. Well, this is really the great selling point of Richard Steenan is that you came up with a lot of those categories and that was the first thing you did in the foundational days of IT Harvest, so that remains your IP.

Speaker 2:

Yeah, yeah, completely.

Speaker 1:

Yeah, but I do have a silver lining for the problem and I think that this is where the sas companies will have to go. I don't think that it's a choice. You have two there. There will be two main models going forward. I think. I'm interested to hear what you have to say. Number one would be the AI agentic factories, so the folks who are going to build agents and build models that others can then productize and the enterprise can basically customize this to their heart's content. There might not be a box around it, but it could be a chat bot in a CRM. It could be a chat bot for I don't know for handling customer calls, customer support, whatever. They're already using some of that.

Speaker 1:

The second model is going to have to be and if the SaaS companies want to survive will have to be the agentification of SaaS. Harvest is going with the need to catalog and to understand the vendor space, especially for private equity that might be interested in cybersecurity companies or anybody who's looking to invest in cybersecurity Virtually the people that are interested in IT Harvest. Now they will need to incorporate some sort of agent into their software, right, and this only makes the software better, and so you can't go to chat GPT. Like you said, there are those limitations within chat GPT. They don't understand the categories. That's still all your brainpower and all of the work that you and your team have put into IT Harvest.

Speaker 1:

But what you do is you, I call it, you know harvest. But what you do is you, I call it, you know agentification. It's kind of like gentrification but agentify. Agentify it harvest, such that it there's a chat bot in there. People can ask it questions. It'll spit out everything that you need to know about any vendor in the space. It can create Monte Carlo simulations, it can create forecasts, anything that an investor needs in order to make wise decisions. But it's all specialized and optimized using the IT harvest methodology. And I think that all SaaS they're going to have to go that way, because where they specialize and where they excel all SaaS platforms is in the maintenance, in the customer service, the specialization, the security, the privacy, all that stuff. And now all they have to do is kind of bolt on the agents into that so it becomes a value add instead of a competitor. So to your point, in that way SaaS wouldn't be dead. It's just that the current SaaS model might be dead.

Speaker 2:

Yeah, yeah, I think it's easier to make a big claim like that and you're totally right. I think the one thing I learned at RSA because obviously I'm not building agents, I'm not building AI technology right, I'm using it, I'm applying it all the time but I talked to a bunch of the SOC automation AI guys so Bricklayer, quiller, a lot of those guys and they all independently came up with kind of the same methodology is have the AIs that, and then it assigns tasks to the specific agents that are best suited for doing that task.

Speaker 2:

And then it collects the results from multiple agents and turns it into a report that you can read. Everybody's taking that approach and it must be the logical thing to do because everybody's doing. It Makes sense. But at least one of them Quiller was building agents that can be deployed to your desktop. There's an agent that will kill security awareness training and it's unfortunately. It reminds me of Clippy.

Speaker 2:

But if you, for instance, are cutting and pasting some information and it happens to have social security numbers in it, it will intercede and say well, are you sure you want to put that in an email? Because you know it violates company policy and all sorts of things, and they made it templatized. So the company that's deploying these agents would set their own policies and things in the way it interacts with their employees, and that's you know. I immediately think, oh wow, can I do that? You know I need an agent to give to people to deploy, so I can see that they're reading about. You know a cybersecurity vendor and say, hey, we've got a lot of information, would you like to click here and see it? Or that kind of thing. So I truly believe that agent templates are going to be available. Maybe not a shopping mart for different types of templates. But who knows, that could be next.

Speaker 1:

Yeah, I was talking to somebody about this very thing this morning where they're looking to. They're like, okay, do I do? I have a blank page where my customers come to the screen that looks like chat, gpt, and then the customer might be lost. But then I was like, why don't you have a prompt library that basically says you know, or there could be a thing at the bottom of the screen that says what would you like to do today? And then there are a list of outcomes that you can choose from, and then you can look at the prompt library and say, okay, this is what I want to do. So, to your point, everything would have to be outcome oriented. So you'd have to approach it with. You'd have to approach the platform either already knowing what you want the outcome to be, and if you don't know what you want it to be, there might have to be like a Clippy type assistant that says I don't know, that's optimized to whatever domain that AI agent is specializing in. Right, right.

Speaker 2:

Yeah, anything that's got repetitive tasks for sure should be right there so you never have to type anything more than once in your life. So you know. So we have created a IT Harvest agent it's called Harvest IQ and then a third LLM adjudicates between the answers and figures out the best combination. We had to do that because we launched it with just one and we told everybody hey, sign up. And now you've got a agent that you talk to, or chat bot that you talk to, that will be completely informed on all cybersecurity products. So go for it.

Speaker 2:

And people are asking questions that were not about cybersecurity products, just general stuff. So that's when we added a call to perplexity, and so perplexity and answer anything. So we'll get you a perplexity answer and if there's anything in there that our data can contribute, we include that as well. So we get really long answers. But you know, now you ask it for SWOT analysis of a company and it'll just do it for you or create a battle card or compare two companies, or you know, here's the top secret one, we'll see if anybody's listening. Yeah, or you know, here's the top secret one, we'll see if anybody's listening If you are making an argument that one company should buy another company. You just ask HarvestIQai to do that for you, and it does it with a breakdown of how the products work together.

Speaker 1:

Yeah, yeah. And the two companies Optimized and specialized, right, right yeah, because the agentic companies can't do that, the ones that are building the agents, they don't care about that and the current sass platforms don't have that. So specialization, um, and I've seen that trend since chat gpt 3.0 came out and I called up one of my friends and I was like I can make this thing optimized to teach your kid how to do math, to tutor your kid, whatever your application is. You just have to train it on that thing and you're already ahead because you have so much data already to train that. So IT Harvest is far ahead of the crowd in terms of intellectual property methodology.

Speaker 1:

Here are some things that I would like to see in IT Harvest in this agentic age. Here's something that I would like If I were a PE firm. I already have a portfolio of my best performing companies over the last 10 years. I want to drop those companies with their attributes into IT Harvest and say, based on my best performing portfolio companies, what company should I invest in next? Or what are the top five companies or the top 10 companies that are most likely to generate a return like the ones that these companies have? What do you think about that?

Speaker 2:

I think it'd be worth a try. You know why not, Do you think?

Speaker 1:

that's just conceptually. Are there any problems with that?

Speaker 2:

None at all. Yeah, because we've got the data there right. We've got our health scores, we've got the growth of every company. We could, if it really helps. If their portfolio companies are in our database, right, if they're long gone, then they wouldn't be in our database, we wouldn't know anything about them. But we could line up on geographies, categories, growth rates, size, funding levels, all that, and then it would look deeper into the products. And yeah, I think we could do that. We should try it. I've got most of the portfolios of most of the companies, so I'll just take one of their portfolios and analyze it. I'll be the one that decides which one is a good investment, but I do that all day every day you can tell which ones are not good investments.

Speaker 1:

Yeah. So you've got the human in the loop and that's the model. It's outcome-based versus deterministic. So I don't think it's sas versus ai. It's deterministic, where the company has a set of features that customers kind of already have in their mind what they want those features to be, and then they come in and then they measure what they want against the feature set, and that's's how most SaaS companies are sold today. That's why they say don't do it by features, because if you did it by features like everybody's basically got the same features Right, yeah. But with AI you don't think about features, you think about outcome, right, right yeah. So if you thought about outcome with SaaS, they're going to charge you $100,000 to customize so you get your outcome.

Speaker 2:

That's right. Two years ago, ever since I saw what AI could do, it was burning in my brain. I had to just write about it on Substack. So I did, and it's very similar to what you just described, but it's for intermediating, disrupting the entire publishing industry.

Speaker 2:

So if you publish a book, you quickly discover, especially in the fiction world, that there's gatekeepers between you and the publisher. So, simon Schuster, if you're famous or something, you can go directly to them. But typically you get an agent. So you have to find an agent. So you spend a lot of time querying agents. The agents are just people who hopefully like to read books and they are looking at whether they can sell it. Right, they're not looking at whether they like your writing or anything like that. They just want to know if they can sell it.

Speaker 2:

So what if you had a LLM that had all books and most LLMs do have all books in them and somebody could submit their manuscript and it could say hey, you know, here are the five publishers that are most likely to be interested in publishing your book. And oh and, by the way, you know, here's an analysis of what you submitted. You can make it a lot better if the third chapter did this and this because, of course, you could do all that too, things that ChatGPT would be happy to do for you today. And then, on the publisher's side, they'd be the ones to subscribe Because you would have all these manuscripts. They wouldn't have to filter it through humans that are notoriously bad at picking bestsellers. Now, in the publishing industry, all you want to do is get a bestseller. The only reason you publish 100 books for every bestseller is you don't know what you're doing. You don't know what the market wants and you can't provide it.

Speaker 2:

So you're going to lose money on all the non-bestsellers and you're going to make a killing on the one that sells 2 million copies. So the publishers would sign up and when there's a hit that they should publish, then they would reach out to the author and make a contract and cutting out the middle person, the agent, completely, and they'd start. You know they'd actually become profitable if they made a few more correct guesses on good books.

Speaker 1:

Yeah, yeah, again, outcome based, right, right, what is it? Think about what the outcome you want. I want a bestseller, all right, let's optimize for that. And if you're an agent and if you're listening to this and richard just scare, scare, the, the bejesus out of you, I think that there might be hope, just like it harvest. Maybe there would still be, maybe not like a junior agent, if there's such a thing, but maybe there is someone who has the final say and kind of puts their human touch on it and says yeah, the baby step is simple the agency, the agent.

Speaker 2:

now we're mixing up our terms again. But the people that the literary agents should use agentic AI to weed through their slush pile, which is what we call the ones they haven't read yet.

Speaker 2:

Oh, gotcha, gotcha and find the gems in there, find the ones that look like they could be developed. You may find a manuscript that was submitted a year ago from an author that wrote another book that got published and was a bestseller, and you can go back to that author and say, hey, we're interested in that. Now that you've figured out how to write bestsellers, maybe you'd like to take another glance at your old manuscript. There's billions hidden in there.

Speaker 1:

Yeah, if nobody's working on it, get on it, because I think that's going to be great. So there's another factor to this. Now you've got my gears turning on this publishing thing and we can apply this across all kinds of models, even cybersecurity, it, whatever. There's another factor here that when we're picking winners and losers yes, the quality of the work, the author, the relevance of it okay, that's all good, but there's one factor which is timing, and I suppose timing and relevance are the same right, because you can write the shining in 1995 and write the shining in 2005 and make it. You may have two different outcomes.

Speaker 1:

Yeah, for sure so per or with, in the case of it harvest. You may invest in a company in 2005, invest in the same company in 2025, and get two different outcomes, as I did. Is there an AI application for that? Is there something that you can think of that solves the timing problem?

Speaker 2:

If anything, it'll come out of the people who are applying agentic AI to the stock markets. Right, Because that's all timing.

Speaker 1:

Yeah, maybe some like correlations or something.

Speaker 2:

Yeah for sure. And you know, if you had all social media going on fed into which at one point you did, grok does, then you can kind of tell. And you look at historical stuff too. You know during the World Wars what kind of books sold the best, when was the heyday of mysteries, whatever, and what caused that? Yeah, you could draw those. You definitely do that timing, yeah. Yeah, you definitely do that timing, yeah.

Speaker 1:

Yeah yeah, I was tossing around the idea of creating a tool that just tracks trigger events. So in sales, one of the biggest factors is really timing, because you can reach out to somebody 20 times a year, but if they're not in a buying mood, they don't have the budget, they don't have the need, they're rocking and rolling already with another tool, whatever.

Speaker 2:

That's totally why salespeople do reach out so consistently is because it works. You know the 20th time somebody goes. Oh, I was trying to remember your name. Thanks for reaching out.

Speaker 1:

So a free tip to all my sales folks persistence is key, because underlying that persistence is the fact that you don't know when is the right time and people don't tell you, and so you have to just keep hitting it. And you know, trigger events are a part of that. So let's say, for I don't know, vulnerability scans or whatever, they probably peak in their sales when there's a big breach. Right, sims, probably the same thing. If AI and cybersecurity all of a sudden are hot which it's always been hot, but let's say all of a sudden it's hot, then IT Harvest is a more valuable tool, because now they can weigh through all of the vendors so they can make their investments. So, yeah, outcome-based, all right. Is that all? On SaaS, we've determined that SaaS isn't dead, it's just comatose. So we can give it some smelling salts, which would be AI agents, and wake it the hell up.

Speaker 2:

Totally. And you know, don't forget, if you've got a SaaS product and you say you're Canva, right, and it's really old, but everybody just says I do use that once a year, so I'm going to keep paying the monthly subscription or whatever. You'll survive for a long time in the monthly subscription or whatever. You'll survive for a long time, right. But obviously Canva is incorporating AI now. Right, they're giving all those tools, as is Grammarly and the rest. But you could just kind of sit on your laurels and slowly fade off into, you know, the end of life for your product and that's the most profitable time of your business is when you're not spending money on the product or marketing or sales. So, yeah, you take advantage of that. Go on. And retire.

Speaker 2:

Or you can just start sweating again and come up with a better solution.

Speaker 1:

Yeah, yeah, yeah. If you want to check out some AI shops, hit me up. I can make some good recommendations in addition to the ones that Richard had mentioned. Is AI going to take your job? And I hear a lot of different narratives around this, but the prevailing narrative is that AI will not take your job. It will make the humans better. And I hear this like clockwork and I've been repeating this for weeks and weeks and weeks, and it's not because I'm trying to drive a particular narrative. It's because I hear competing narratives.

Speaker 1:

If you see, going back, bill Clinton, if somebody's pointing somewhere and they're looking somewhere else and they're telling you something, it's, there's probably something screwy, right? So Bill Clinton, looking this way, say, I did not have sexual relations with that woman. Pointing one way, looking the other way, all right, turns out he was lying about that, and so it's not that I think that people are lying. I just think that vendors are scared and they don't want to push AI too quickly and lose sales. That's one thing. Enterprises want to integrate AI into their work, but they don't want to push it too hard and move too fast and encounter the security and the privacy issues. Okay, we also don't want to scare employees. You do not want to scare the job market. If you scare the job market, it screws up the economy. You don't want to do that. That's another reason why I think people are pumping the brakes on AI and saying that it won't take your jobs. But you and I can, yes, and I think that my platform is small enough to where I could say that and not do too much damage. But let's say, on CNN, they're coming out and saying, hey, ai is taking your jobs. You can imagine what that's going to do. Here's my thesis. Let me know if you agree or not.

Speaker 1:

Richard, my thesis is that AI will take your jobs, but it will not take present jobs. It will take some present jobs, but it would not so much take the present jobs as it will kind of halt the job market for future employment, which means that colleges and the future job pool would be in trouble. Right, we're hearing the narratives around. You know AI will take your jobs, but then, or AI will not take your jobs. But at the same time, people are saying that AI will replace labor, it will make labor basically go away manual labor. So it's very evident and it's very clear that AI is doing something with jobs. Will that impact the average person, or the junior employee, or the senior employees or, I don't know, maybe even CEOs? What do you think, richard? Will AI take your job? If so, how, and if not, why?

Speaker 2:

Well, I have to fall back on historical references. Every technological revolution has definitely displaced workers, like no question, right From the cotton gin to automated sawmills et cetera. And in my experience in the workforce, when I got out of school we still had typists and these were people, invariably women, that would type everybody's memos in commercial communications. So 1982 is like just the end of that, right, because that was roughly when the IBM PC AT was introduced and that was the microcomputer revolution that was happening. And probably five years later there were no more typists All gone.

Speaker 2:

There used to be tens of thousands of people whose chief skill was typing and they did that for corporate offices, completely obliterated. And the other one is the kind of you remember Apple creating desktop publishing? That's actually what they went to market with. Is we're going to make it easy to create nice graphics and print them and talk to printers and talk to printers. So you know a lot of graphic artists. You know that used to literally cut and paste things on a board and then somehow turn that into something that could be printed. They were out of that job, but most of them just transitioned right. They loved the computers and digital art and made a good transition. So I think we're in the same stage now.

Speaker 2:

Somebody the other day said well, you know, lawyers aren't going to be displaced. Are you kidding? They're going to be the first right. It's like you do not need a lawyer to create a contract between two people who just want a legal-sounding agreement between them. Right, right, the chat. Gpt is so good at that and it's complete, and it throws in all the bullshit language about. You know what.

Speaker 1:

You know just all the stuff you know Definitions and.

Speaker 2:

Definitions and supersessions, and statutes and.

Speaker 2:

All that stuff. Now, mind you, I wouldn't want to use chat GPT to argue a case for me in front of a court, but it sure helped inform that That'd be great. It could take all the depositions and weave them together and come up with a good story for the jury, but for sure, contract law is going to be significantly challenged. You still need lawyers, you know. No question Lawyers will remember how lawyers were so slow to get on the internet, like they wouldn't have email addresses for you probably don't remember this, but for the early first five years of the internet, lawyers refused to do it.

Speaker 2:

Because you know, in today's world you send an email to your lawyer and say, hey, can you look at this clause for me? And they give you an answer right, and you know that it took them two minutes to do it. So it's really hard for them to bill the 45 minutes they used to charge for that kind of work. So they just hated the idea of having a conversation with their clients. It was so fast and friction-free. So they got over that they had to, because the young crowd of lawyers that were coming behind them were perfectly happy to do that. No problem.

Speaker 1:

Yeah, they're bobbing and weaving, yeah.

Speaker 2:

Imagine a patent attorney. Right? It's like patent attorneys have to write in a specific language and they just take what you tell them about your idea and turn it into that language. It's come on chat. Gpt can do that in its sleep, so you don't need them. You know what other things? Financial analysts? I don't think I need those anymore, so yeah.

Speaker 1:

What about accounts, the number crunchers?

Speaker 2:

Yeah, you know, the number crunchers are the analysts, right? Actually, you know, the accountants are the ones that set up the books and put in the controls. And yeah, I think we'll need those for a while for sure.

Speaker 1:

So can I ask you this, if I am correct, in that the college pool or the workforce that's coming from colleges, they're the ones that are going to be the most affected If we were to apply it to this scenario? So your scenario of the lawyers right, let's say you're not going to need lawyers in the future. The lawyers right, let's say you're not going to need lawyers in the future Um, what would that do to law schools? Like, if I were to be a lawyer, if I wanted to be a lawyer, and if I'm a law firm, let's say, okay, on one side, I'm a future lawyer, on the other side, I have a law firm.

Speaker 1:

The law firm is thinking we already have lawyers. We're not going to fire them because they've been around for 5, 10, 15, 20 years, but we're going to give them these tools, we're going to give them some AI technology and they're going to be like flying, rocking and rolling. We're going to do more business, more clients, more money. But if I am a law firm, will I hire a new lawyer and have to train them for a year or two years and have to worry about them becoming partner and displacing another partner, which you don't actually become partner unless one partner leaves, because the firm's only got so many partners right. So now, as a poor college student wanting to become a lawyer, do I enter this job market that's dominated by cyborgs essentially Humans, using AI?

Speaker 2:

Well, right now you don't have a problem because as a student, you can know more about AI than the fuddy-dud old attorneys. So now you've got something against them.

Speaker 1:

That's a great point, Richard.

Speaker 2:

You can 10x their productivity right, so you can outproduce them by a factor of 10, maybe without working 80-hour weeks your first two years. So you've got that going for you. In the future, there will just be a market that lawyers are 10 times as productive as they are now, and that tells me that probably we don't need to graduate as many attorneys. So right now we're at this cusp and this is where everybody should listen up. If you're going to a school that is frightened by AI, if they start saying no AI in these assignments, no AI for this test, and we're going to check it for AI production, we just invested $100,000 in a tool to check whether you're cheating by using AI. Change your school right now. You are going to just be so ill-served by your degree from that school.

Speaker 2:

There are plenty of schools I find schools on the East Coast in particular, from DC up through New York and Connecticut, that are embracing AI wholeheartedly. Here in the Midwest, pure fear. The professors are afraid of it and, you know, mostly in letters and science, less so maybe in physics or something. You've got to grasp this right now. It's like graduating school in 1990 without, without you know, having a modem at home. It's like how do you do that?

Speaker 1:

yeah, it's like back in the day when books were invented. Imagine showing up to school where everything was just. You had an orator who gave you a speech or taught, like play-doh or you had a slate slate, a wax tablet.

Speaker 2:

You had to write stuff down.

Speaker 1:

Right, you had to write some stuff. And then somebody shows up with like a nicely bound book with all of Plato's speeches and then some, and you've got a test to take and Plato or whoever the guy, goes hey, you're not allowed to read that book, you can't use that book, that's cheating. You've got access to knowledge that nobody else has, that's right. How crazy would that be.

Speaker 2:

Yeah, good point.

Speaker 1:

Nobody would do that, but chat, gpt, it's like the, or LLMs, or AI in general. It's like it's a technology on par with the invention of the book. Yeah, all right. The invention of the book. Yeah, yeah, all right. Well, I think we solved world problems and we've brought peace to the world, richard.

Speaker 2:

Anything else that's burning on your heart. Yeah, we got to worry about warfare and peace, you know, because it is. It's a little unfair that we're at the verge of a world of abundance. You know that AI can create, you know. At least the visionaries are thinking this way, and yet the world's falling apart around us right, and war is breaking out and it sounds horrible. And why couldn't it all converge? You know we're post-Soviet Union was the chance to create a liberal world where everybody plays nice with each other and trade is the reason countries interact and not land grabs, and we blew that opportunity. Here we are.

Speaker 1:

I think the information, or I think the problem is information. It's a world of abundance, but there's too much abundance. Have you read Chris Hayes' book Siren Call? No, so Chris Hayes, the MSNBC guy, wrote a book called Siren's Call, which basically he's arguing that we live in the attention economy, the attention age, where attention is bought and sold, and in the world of information and abundance value is found in not the amount of information that you present to people, but your ability to condense that information, to reduce it to its simplest components and be able to present that to people.

Speaker 1:

So I think that the world is devolving into chaos in some way. Number one our understanding of that chaos is a function or a result of too much information. So it seems chaotic because your brain can't process everything that's happening. So I think the future belongs to those who are playing in the information and the attention economy and is incumbent. It is incumbent on them to simplify, and so with that I will wrap up this episode of cybernomics. So if you want to stay on the cutting edge of technology and ai in this chaotic world where everybody's trying to get your attention, we try to break it down and simplify it so that you can understand and make critical business decisions. Richard, thanks for joining me again on this episode of Cybernomics.

Speaker 2:

Thanks, josh, mind opening as always.

Speaker 1:

All right and thanks for listening. If you want to learn more about me, check out. How do you learn more about me? Check me out on LinkedIn and it's LinkedIncom Josh Bruning. Just look for Josh Bruning and Richard. If people want to find you, how do they find you?

Speaker 2:

Just ask ChatGPT how to get a hold of me.

Speaker 1:

Oh great.

Speaker 2:

Good, all right.

Speaker 1:

Thanks, bye, bye, all right.