Peter Morgan Deep Learning Partnership Building Human Like AI Agents in 2026

So what we're seeing right is entry- levelvel jobs, computer science undergraduates graduating with a computer science degree and finding it harder to get on that ladder, right? The entry- level jobs have been done by Chat GBT and Gemini and Cord, they're actually that type of work is the most easily automated by large language models. And it's not good or bad. It just is. It's a thing, right? And if we can embrace that as a reality, and it's only going to get more as we move forward, more and more jobs, entry- level jobs, and it will move up the stack to the higher like masters and PhD level. Hi, my name is Dmitri Bonichi and I'm a content creator, agency owner, and AI enthusiast. You're listening to the AI agents podcast brought to you by Jot Form and featuring our very own CEO and founder, Idkin

Tank. This is the show where artificial intelligence meets innovation, productivity, and the tools shaping the future of work. Enjoy the show. Hello there and welcome back to another episode of the AI Agents podcast. In this episode, I'm here with Peter Morgan from Deep Learning Partnership. How you doing, Peter? >> Hi, Demetri. I'm great, thanks. Thanks for having me on. >> Yeah, we really appreciate having you on the show. So just to kind of get things started um tell us a little bit about you know your background and and how you got into uh AI in general and then uh we'll we'll kind of dive into deep learning partnership. >> Sure. Yeah. So I started this journey into AI quite a while ago um over 10 years now. So I've been doing it longer than a lot of people in fact. So the way how I

got in was um I have a physics background and then I was a solutions architect in computer networking and I was getting a little bit bored to be honest. I did the sort of um computer networking for about 10 years and you know coming from a physic background after a while I felt I just sort of uh that had run its course. So I went back to uh my alma ma for three years and worked on a physics experiment measuring the mass >> and when I when I came out of that I'm like oh god what the hell do I do now >> and um that was uh 2012 and data science was just becoming a thing right uh the devport had just written an article for the Harvard Business Review data science will be the sexiest job of the 21st century so how can

you resist that right so I became a data scientist Um, now that's a bit flippant thing to say because there's an awful lot of work to become a good data scientist, but that's where my journey began. >> Yeah, probably 2013 really. And so I basically trained up on all the Python libraries, all the all the good technical stuff. Um, which is probably quite boring for your listeners, but at the end of the day, right, it's a very technical area. So I I trained up on all the things scikitlearn and numpy scypi all the python frameworks that are were used for analyzing data sets quite often in the scientific community but again data science was just becoming a thing and so we were applying it to businesses and so from there I've just watched it over the last 12 years uh get all these um frameworks

get progressively and these AI models get progressively more powerful and capable until probably the release of chatbt was a huge moment where it suddenly became into the public awareness. Before that, it was more the community, right? A machine learning community. Um, and but then suddenly, you know, it had 100 million users in downloads in like two months. So, you know, the fastest download of any app ever. So, people love intelligence. So, my the the trajectory t took a little bit of a change then. Uh, and it became really mainstream. So, I guess that's when I felt like well maybe all my work had paid off at that point. Interesting. Okay. Well, I I I'd love hearing everyone's journey and how it gets started. So, like >> um how did that lead you to and to tell us a little bit more about deep learning partnership?

>> Yeah, so I did start uh deep learning partnership in 2016. So, I was a data scientist freelance for three years uh and um 313 2016 and I was just taking freelance gigs uh contracting, consulting, that type of thing. Worked for some super interesting startups in London. And I was based in London at the time and still am. And um yeah, the startup world combined with data science was just quite magical for me actually cuz a lot of really super interesting and smart people work in the uh startup, you know, start their own companies, right? Because a bit like me, they don't really, you know, like working for big corporates or whatever. So yeah, I found a real affinity with those guys. Yeah. >> Okay. And um uh yeah just could you dive into more of uh what necessarily you are trying to do like

what's the goal of deep deep learning partnership on the vision on your website it says the vision is to empower humanity with safe machine intelligence um and the mission is to automate business and science with AI what does that mean to you obviously words on a screen or one thing but hearing it from you yourself >> yeah so that's right so in 2016 I decided to actually start my own company a consulting company called deep learning partnership and so the vision and the mission. Um, for me, they have to be big statements, right? To get, you know, people, you know, engaged in your story and your journey. So, you know, I said, okay, where where where's the what where what's what's the biggest thing that could possibly happen with machine learning? And I thought it's going to change the world, but how specifically? And it's

just I could see immediately it was going to change business, how businesses operate, how the government operates, and how science is done, right? So, it's the big picture. And I could see that early on and I and I will say I was kind of early. Um so that if I had to give myself any credit it would be for seeing that and and it's just come more and more true isn't it with CHBT and Gemini and Cord. Um I mean these things can really do what uh a lot of white collar work a lot of scientists and a lot of governments do. So yeah my my mission has actually is coming true as we speak and so yeah I predicted it perfectly. >> H very cool. Well, you know, uh you know, everyone everyone I we try to predict things here too on the show

uh with everything going on in AI, it's uh not the easiest uh thing to go, but um uh you know, it's there's so many different things right now with AI that um I think are you know uh top of mind, but for you with what you're doing at deep learning partnership, what are some of the you know different uh kind of things that are top of mind right now with what you're you're doing? >> Yeah, that's a great question. So really it's a disruption in businesses kind of wondering they don't want to get caught out right and so you know I wouldn't say panic but it's people really wanting me to come in and our company to come in and our machine learning engineers and kind of hold their hand a little bit and say this is what you your company could be doing with

your data so you don't basically go out of business uh from some from the companies who are embracing machine learning and training them on their own uh proprietary data sets and just using them and integrating these tools into their workflows and into their systems. So that involves talking to the uh a lot of companies don't have you know machine learning talent right PhD level machine learning they don't come you know they don't you know that's quite rare and a lot of them work at Google Microsoft and anthropics and open so it's very rare so yeah basically just kind of handholding usually IT departments managed business ex you know sea level and saying hey this is what we can do and this is what other companies are doing and uh yeah we should definitely get started building a prototype Okay. And let's see if we can

uh put that into production for you as well on a very specific part of a workflow and then just scale from there. >> Okay. Um you know, what are some of the I know you have a lot of partners. I'm looking at your partners page right now. >> Um you know, how how do you how did those kind of relationships come to fruition and what does like having a partnership with you look like? >> Yeah, that's a great question. So again because I started quite early you know these partnerships were formed quite early right with uh quite technical people inside of startups inside of the bigger companies like Google Amazon and Microsoft uh we're talking really uh Google cloud uh AWS and Azure and so uh cloud partnerships are very important um because most uh businesses run on the cloud these days some some are

on prem um a lot of hybrid but a lot are just using the cloud services. So that uh entailed just forming partnerships early with AWS uh Google and Azure uh and also Oracle. Now they're they're doing really well. They've caught up. They're way behind, but they've caught up really well and so almost a trillion dollar company now. Can you believe it? Um yeah, >> that is crazy. >> I know. They almost doubled in the last six. >> Yeah, I know. And it's all AIdriven this whole thing. So the cloud uh companies are really well positioned. I mean they're all two three trillion dollar companies now. um almost four trillion for Microsoft, Google just hit three trillion, Amazon's two trillion, Oracle 1 trillion. So these are the this is the the futures now and we're seeing it right in the numbers. So yeah, partnership and then

the other partnerships were with startups and not you know beside the cloud company were with startups who were had really interesting tools and frameworks that we could leverage take them into our clients. Yeah. >> Interesting. Okay. Um, and you know, just just in general with the the state of AI right now, where do you think um this uh world that we're kind of living in? >> Okay. >> Has impacted the job market, will continue to impact it. I think a lot of people are curious right now like what the impact is going to be for uh their jobs and whatnot. >> Sure. Sure. That's a great question too, Demetri. Yeah. Um it's gonna impact it is impacting already. So what we're seeing right is entry- level jobs, computer science uh undergraduates graduating with a computer science degree and um finding it harder to get on

that ladder, right? The entry- level jobs are being done by Chat GBT and Gemini and Cord. Uh they're actually that type of work uh is the most easily automated by large language models. Um and it's not good or bad. It it just is. It's a thing, right? And if we can embrace that as a reality and it's only going to get more as we move forward more and more jobs entry- level jobs and it'll move up the stack to the higher like masters and PhD level um yeah the better off we'll be right so we have to accept the world as it is so we're really going to have this is so disruptive right which is why my mission statements are quite big um that we're going to have to have a new economic system uh the government's going to have to step in and

and have some sort of universal basic income I know people don't like the sound of that because it sounds an awful lot like socialism and or you know it's not really the way humans are built I don't think so it's going to be so disruptive and we're at the very beginning maybe we're one or two% into that journey but uh eventually you know we'll see 50% of the jobs automated Dario Amade from Anthropic already said that in two to three years already 50% will be gone um I'm not just copying him that's what I've been thinking for a long long time and um yeah so when people like him come out and say that I think that's really quite brave and courageous because that is kind of the reality that we're moving into. Demetri. >> Yeah. No, I um it's interesting. I feel like I've

gotten a lot of different answers from people um on the uh uh you know like the >> trying to think about it like the >> Yeah. >> Um the li landscape for jobs right now >> is so tenuous it feels like in people's in people's mind, right? And you know, there's also some people though who think that it's going to get better before it gets worse, right? Like uh or like it's No, it's going to get worse before it gets better. What do you What are your thoughts on that necessarily? >> Yeah, I don't think so. No, I just think it's neither good nor bad. It's just we're going through an evolution here and um a revolution, too. So, this is the fourth industrial revolution. So, we had uh steam engines which automate a lot of labor and the ludite smashed up the mills in

the UK when that happened because there were no job. it was displacing their jobs and the government hadn't kind of uh set it up you know in advance so they had jobs to move into. So they basically lost their livelihoods um almost overnight and and then we had electricity then we had the computers and every single one of these has been so disruptive we but new jobs have been created to meet you every single time but this time is different because when you automate intelligence there's nowhere left to go these machines can do what we do the white collar jobs will go the accountants the lawyers the doctors the computer de you know the software engineers and um and then the blue collar will go too so the robots will you know a lot of companies are developing robots really fast. China, the US in

particular and um the I mean they they will automate all the blue collar work too. I mean, robotic factories are a thing already. These are rules based, but now we're going to have AI driven robots and yeah, it'll be lights out and the factories just let the robots do their thing. Human in the loop, but how many humans do you need in the loop, right? Not not like half the population now, which is four billion, you know? So, yeah, the writing's on the wall. The the only question is how are we going to manage this transition? >> That's where government has to step up. It's their job. And what do you think they can do to step up? >> They're going to have we're going to have to redistribute the wealth. Uh, and again, that sounds like socialism. And you know, that's the horrible concept

to think about because capitalism has got us here, right? Probably 400 years. Before that, we had feudal systems which weren't that much fun for 99.9% of the population. We're going to have to go through an economic transition as well as a technological one as well this time. Uh, and so yeah, that will be a redistribution of wealth, some sort of robot tax. already, you know, people at Google and Microsoft are kind of, you know, saying, "Yeah, no, we're not not going to do anything like that. We're got, you know, we're going to keep all our money and become hundred trillion dollar company." They're like saying, they're not saying it exactly uh directly, but yeah, the writing, >> well, what things have they been what things have they been saying? >> Yeah, they're basically saying, "Yeah, we're going to need some sort of new economic system."

And that's then then they stop because they don't want to go further because then the share price will will kind of tank, right? So yeah, they're playing it very wisely and the way I would play it too, right? You don't want to scare people too much, but it's it's inevitable and it needs to be managed and we've got a few years. It will take a decade, but a decade is not long in the big scheme of things, right? It's just not a long time. >> Interesting. Yeah. What about uh I have some thoughts on that. What about like >> Mhm. >> the impact of companies like Google being in this type of tech, right? Like so for example are you familiar with the recent argument that we have uh what is it like >> I don't know how many companies but the main one was

Perplexity vying to buy Google Chrome. >> Yes. >> What are your thoughts on that whole situation? >> Yeah. So we're seeing the the tail wagging the dog now we these companies like cursor and uh Perplexity just becoming like 10 billion 20 billion $50 billion companies within a year right and the ARR the annual run rate really you know being sufficient that you could almost justify these valuations so we've had these three huge uh technology transformations in the past but they've taken 10 20 30 even 50 years for the steam engine you know so we've had time to adjust these these are happening in a year. So a year ago, Flexi was worth about a billion. Now they're, you know, or maybe 100 million. Now they're worth 10 billion. Same with cursor, right? And you look at the dollars and it's actually justifiable, right? It's not

not even a bubble. They're like P PE ratios are kind of like in line with that. And so things are happening way quicker than we're we're used to. And they're only going to accelerate because this is an exponential technology. So we just have to get used to the fact that unless company big companies are going to be super nimble uh they'll get kind of disrupted by these uh smaller companies. It's a very interesting times but yeah it's the Chinese curse right you live in interesting times. Yeah we're living in interesting times. Yeah, I mean you know what for me >> China is maybe one of the concerning um >> factors for for some of the different companies but then I or well for us as uh well I'm from the states right? >> Yeah. So I I do wonder like what do you think China's

impact is on this whole this whole realm because there are >> a lot of uh companies out there that have managed to even with like the uh the the chip restrictions to do or to do really well like with what happened with um DeepSeek. >> Yeah, correct. That's a great example. I'm glad you br it up. So that was only you know nine months ago I guess December January time when deepseek kind of leaprogged everything that Google and Microsoft openai and entropic were doing and it's like they came out with that reasoning model which just such a thing of beauty right and also impressive technological prowess a hedge fund with 200 very smart people uh you know leap frogging Google and it's just like okay China's in the in the game now that was that was a signal we're there we're here hi you know

It's just not the US anymore. It's the US and China. And ever since they >> That's a good point. Yeah. >> Yeah. Ever since they've kept up the play. So it's very It reminds me of the USRussia kind of arms rate, the neutral arms race in the 50s and 60. Now it's going to be US China uh with the AI arms race. And so I feel personally because I'm kind of like an innocent bystander out here in the UK. I don't I'm not US or China. I actually think competition is quite good for the technology itself, but it's >> Oh, absolutely. Yeah. >> Yeah. It's not going to be the US having all its own way. It's going to be 50/50, a lot of competition. But for me, competition is good. But yeah, you you may see that slightly different. Yeah. >> You know, I

Yeah, I agree with you. I think uh I think it's an amazing um thing to have competition like this. I I um I was so shocked when it happened, right, >> with the the improvement that happened with the everything on the Deep Seek side. Like when when that came to light, what was so impactful to me >> was the cost situation, right? I don't know if you recall, but cost was hilariously low um actually from an API standpoint. And then a follow through of that was that um >> what was it the >> I'm not you know forgive me if I misspeak but >> there was like a 5 to1 discrepancy API costwise between like chatbt and deepseeek and then chatbt ended up like lowering all the model costs like 03 I think had an 80% cost reduction within the next like month which I

found fascinating. What are your thought Yeah. What are your thoughts on how the accessibility of this uh these different tools uh from a cost standpoint with the the pricing changes? >> Yeah, definitely. Again, that's why competition is so good. It drives the cost down, right? Otherwise, the US could keep the it be probably 10 times more expensive. Um so, yeah, that it's just competition is always best. Monopolies are never the best for the market or the consumer. >> And is that is that the main reason you're arguing for that redistribution? Is it just because it's like excessive at some point like with the size that they could reach? um without uh you know without that redistribute because I feel like you know even if there's like mass accessibility to some to some extent that amount of money can lead to you know um I what

are your thoughts on that situation? Yeah. No, no, I think uh yeah, I'm I'm completely neutral there. And if China hadn't released Deep Seek and, you know, had stayed behind for another couple of years in the US out in front, that would have been fine, too, right? Because the charge the cost isn't that bad. It's like 20 bucks a month for OpenAI. Now they they introduce these other two tiers, 100 bucks and a thousand bucks. All of them have AW, you know, sorry, Google Antropic as well. So if you want the really super high performant model, you're still going to get charged a thousand months, but companies can afford that. So it wasn't really the cost. For me, what's most exciting is it actually accelerates the rate of technological development. It woke Google and anthropic and open AAI. It woke them the hell up, right?

And said, you know what, we've got we've got the technology. We've been sitting on it. We haven't been telling now we got, you know, China's b forced us to show our hand. And there we go. That's the story. And what companies do you think uh I mean there's a lot of different companies out there >> that people view as like the top dogs so to speak. What ones do you think will remain in the pole position as it were uh moving forward? Cuz you know you have Claude, you have Anthropic, you have companies uh like that vying for things. Which ones do you think will essentially have and keep the pole position? Uh cuz you know a perplexity is on the list cuz some people are concerned you know they could become >> monolithic right with the and I'll get into why but I'm curious

what your thoughts are. >> Yeah sure. Um yeah I think you know the the race is already over. It was over probably five years ago. So AWS or Amazon, Microsoft and Google. Uh and that's reflected in their market cap right again Google just hit three trillion. Microsoft's just four trillion. Nvidia's 4.4 trillion. Uh they make the hardware, the other companies make the software and the models, right? So the race was over probably 10 years ago even. And then the Chinese companies um mostly 10 cent, Huawei by um yeah the uh 10 cent. Yeah. Huawei, not BU. Yeah, BU. And so the the Chinese have their equivalent of Amazon, Google, and Microsoft cloud provided. So they're they're going to win. They're going to scoop all the money. They'll all become hundred trillion dollar company. That's when the government has to step in and you know redistribute

the wealth because otherwise there'll be nothing left. See this this this force industrial revolution again is a little different than ones before um where you could just sort of you know tax them a little bit but now they're really going to have to redistribute because it you know it's just exponential curve. You follow that curve and and the three big cloud companies end up with with all the money like and we have not been ever here before. Um so you know Google will buy complexity and I mean you know or Microsoft they'll just buy just you know 40 billion's just pocket change they're worth four trillion for god's sake you know it's 1% of their market they can they could have bought them you know whatever they that's >> there's a there's actually funny there's a rule that I just read in a in a

report it was like if your net worth right if if a purchase say you're worth $500,000 if you >> have uh to you shouldn't bother wasting the time to think about a purchase that's worth like half a percent, right, of your of your net worth, for example. And it's what you're basically saying is that's pretty close in regards to >> these large large corporations. Is that fair to say? >> Yeah. Yeah. Also, Demetri like Amazon and Google have both invested heav Microsoft. So, Microsoft bought 50% of OpenAI or 49%. Um, Google and Amazon both own a sizable chunk of anthropic. Um, so yeah, they've already own half of these anthropic and open eye are already half owned by the three major cloud providers anyway. I mean, I don't know if people know that, but they already own them. >> Yeah. Speak speak to the the

foundational quote ownership that AWS sort of has on most things because I don't I don't uh some aren't aware of this. I mean, many people like just are out here learning about tools and stuff, but how does that necessarily work with uh AWS because it is the foundation behind >> Yes. most things. >> Yeah. Yeah. So they basically have the platform, right? The cloud platforms as to Google and Azure. And so Azure's perhaps more focused to the enterprise. AWS is the first mover in cloud and Google have a great cloud because they have great engineers. So there's uh they're in a three-way race there. So within AWS, they have the platform already. They have, you know, 50% of the cloud market in the West. And um but they also own you know a big I think they've invested 10 trillion or something of that order

into anthropic and so have Google maybe $10 trillion each and so yeah they own a platform and then they run anthropic models on their platform AWS and Google do you can log in now and just spin up an instance if you know what that means and um yeah you can you can it's all really fully integrated the day they released Sonnet 4.1 called Sonnet uh it was on Google Cloud and AWS the day it was released Right. So they they're working hand in hand now anyway behind the scenes. So yeah, it's all it's it's sewn up. The market's sewn up there. I can't see any other company coming in that will be able to shoot past OpenAI and Topic because they're already half owned by these other players anyway. And Perplexity was probably the last one we'll see. They were a little bit of anomaly

and they've done very well. Uh and Cursor as well with the coding, but yeah. Well, do you feel like there's going to be cuz I mean just to state it again about um perplexity, right? Like their solution specifically >> uh is >> probably the the research and the rag component a little bit more than the other tools. Do you feel like while there might not be obviously a really big improvement um from some like small obscure company that ends up creating a competitor to open AAI's chipt or anthropic or perplexity? Do you feel like there will be smaller micro tools that'll be able to compete for market share in in specific use cases though? >> Yeah, I think where it will be it will be like fine-tuning on a specific data set. It could be a business data set or or say say you know

curing cancer. I mean uh Google deepmind have spun out a company called um isomorphic labs. They're the basically commercializing alpha fold right which is the protein folding which is basically the uh fundamental way that you know description of disease. Every every molecular process in our body is just is just you know proteins interacting with one another or or molecules. And so they've already spun out this company. So that's kind of biology, you know, they already have their fingers and everything. Is there an opportunity for other companies? Of course there is. Um but these are verticals now, right? Um but yes, there will see a ton of companies in different verticals like Isomeorphic Labs. But yes, so yeah, maybe, you know, I'm not saying there'll be nothing. No other companies ever formed, but yeah, I I don't think they'll go to a trillion dollars. They'll probably

maybe hit 100 billion and get bought by someone. Uh I mean the big farmers are in a good position right for for because biology is one of the killer apps of AI right lets who doesn't want to live longer healthier life everyone right and people are willing to pay a lot of money to get another 10 20 30 years in their life as well so the pharma companies if they play it right um they might be in a good position too the big farmer yeah >> okay so um just to speak to that a little bit big pharma >> healthcare medicine, >> that type of realm. >> Um, I've had a couple people on the show we've interviewed, right, for those different uh types of applications. What are your thoughts on in general the benefits of uh AI on uh healthcare in general in the

in the upcoming years? >> Yeah, that's such a great question. Now, we're in very early days, but we've I mentioned Alpha Fold. There's um other uh frameworks out there uh from kind of startups and spinouts as well. There's a group from Meta started a company called might have been evolution AI or something like that or I forget the exact name but the EVO3 is their framework and so there are a bunch of companies building these frameworks training on this biological data cancer data disease that so it all comes down the only moat left is the data. So it's really down to the data now. It's not the hardware. Anyone can spin up an Nvidia instance on AWS. It's not the models because the models have already that race again has been won by open Google etc. So it's really about the data. It's training these

models on data sets and that's really what it's going to come down to Mitri and and that's I mean that's it. That's the next 5 to 10 years is just building these verticals, right? Uh AI verticals and we I can't predict who's going to win in those, but I I I'll guarantee that the three cloud providers will be, you know, part part of the winning uh picture, but there may be some startups too that manage to, you know, go through and become hundred billion dollar companies. and and with these startups and you're saying a hundred billion dollar companies. I had a conversation recently with a with another founder um you know just to be frank a lot of these companies are >> essentially difficult to value right now. >> Um right cuz SAS >> probably much more easy to value versus now when you have

AI tools >> they have this unique >> there's this unique bubble going on. where essentially there's the value of the tech that's on top of the AI and then there's the LLM that they're essentially using to >> make their AI tool work. >> Care to speak on that a little bit more? >> Yeah. Well, I don't think we're in a bubble. Um, so this time, yeah, I don't think we're in a bubble. There's way too much upside. Um, which is >> Oh, I meant Sorry. I meant like um >> Yeah, >> I think valuation wise is maybe the only place I'd be. I'm not saying that there's no value in what the AI is doing, but >> from my perspective, it seems like they're double counting the value of like what these tools are providing where in in essence they're essentially >> counting the LLM

value from OpenAI as OpenAI and then they're counting it again for the company that they're giving they're funding and stuff like that or they're acquiring. >> I don't think so. That's yeah, they're they're the foundation models or the frontier model companies and there's only four there's only five in the west um and it's equivalent number in China uh Google, Microsoft, Amazon, um Anthropic and OpenAI. Um in Europe we have Mistral and that's about it. Maybe you could count deep mind in there but they've been bought by Google a long long time ago. And then in China they have the binary tens. So there's only like 10 foundation companies in the world and yes they they are going to do very very very well. Thank you very much. I wouldn't be surprised if they all got to 10 trillion valuation. Um yeah. Yeah. I I mean

I'm almost put money on that. Um now so is that so that's not you know that's what they're doing. And now let's say there's a startup a bio startup or comes along and says I'm going to train I have access to this data set no one else has access to for this specific type of cancer. I'm going to start a company but they have to be using these foundation model from Google, Amazon and Microsoft. So the the big cloud providers can't lose. It's just that like I say that race was over a long time ago. And so is there going to be any scraps left for these startups who will train their models on different types of disease and make make our lives happier, healthier, longer? Yes, there are. Will they get bought by the big guys? Probably. So yeah, it's tricky. I mean that's

where yeah we're going to have to do redistribute wealth at some point but I mean it's not just in biology it's in chemistry it's in material science it's better batteries it's better solar cells to fight climate change it's you know better cures for all the different diseases it's it's all the different workflows and um every you know operations and every uh enterprise in the world it's every single workflow we can think of science business and government as well governments will be fully AI in 10 years time as well you which is great because AI is actually better at making decisions than humans. >> Speak to that. A lot of people did I I don't actually disagree with you mattering on the level of decision uh it is, but uh yeah, speak to that. Some people might >> not agree with you there. >> Well, may

they might want not want to agree with that. >> They might not want to agree. Yeah, that's the the clarification probably. Yeah. >> Yeah. Change is hard, right? Yeah. There's going to be resistance to change. is always there is like the gray smashed up the mill the ludites they called right uh we're going to have like AI ludites for sure and I totally empathize with those guys too I won't be one of them but I I can see because the machine learning engineers with these PhDs and working side of Google demo they they just so far ahead in knowledge and how could we expect the guy in the street to even figure out what that means right maybe when they saw a mechanical mill they could say yeah I could see the piston moving and I get it that's an engine to replace you our

arms and our muscles, but these things running in data centers distributed around the world, right? They had no clue. They've never been inside a data center. They don't even know what one is. And so there will be some push back for sure. And uh yeah, it's going to be interesting to say the least. I don't >> Um >> it's got, >> you know, I I uh I am curious though >> with decision making, right? Usually we connect that to higher level employees right whether that be seuite um uh people who are coming in as large level consultants. What do you think about >> the peak of decision- making, right, versus the valleys? Because like I I can say pretty confidently I feel like a lot of entry level work is going to be >> eradicated here soon due to the uh the decision-m capabilities of

>> AI. Like my my thought is and my joke anytime is someone says like, "Oh, I'm not sure if they can replace entry-level jobs or like I'm not sure if I trust them. They're only like 90% accurate. My joke in response is find me an associate who's better than 60% and I'll and I'll I'll pay I'll I'll pay double what you're paying them. Right. >> So that is the you know cuz not to not obviously there's a lot of high but the average one like the average associate in any job does a pretty poor job practically. That's why they're an associate not uh not a man a senior manager or a manager in whatever category they're in. So >> yeah, what do you think? And so I think everyone's if if you >> talk that talk track through with them, they'll agree that low-level type

reasoning is good. But what about >> yes, >> you know, strategy and high level seuite decision making? When do we think that's going to be something that's capable of uh improving? >> Well, it's a little >> Yeah, it's a little paradoxical, Demetri, and counterintuitive. It turns out that those jobs are easier to automate using lang large language models than the uh other types of work especially >> really. >> Yeah. So the sea level are the first guys to go. Yeah. Honestly. Yeah. >> Why is that? >> Yeah. Because it's basically training. You put all the training data, all the strategic training data, all the Mckenzie, everything McKenzie, Bane, you know, Deote PWC has ever produced into the LLM. And I've got a suddenly I've got a McKenzie GPT, right? uh and it'll be better than anyone at McKenzie and I say I want a president

or prime minister GBT. I just train it on all the 250 years of American history or thousand years of British history and suddenly I have a prime minister GBT or president GBT. There's no there's yeah this time's a little different than in the past because we we're automating intelligence and at the end of the day just to kind of frame it in the bigger picture. We are um biological information processors and we we are just putting making what we do. We're putting it into silicon and the silicon can scale almost infinitely um into these Nvidia uh these data centers chalk full of the latest Nvidia processors. Um and so yeah, we just can't compete with that. No, no matter what we do, whether we're an entry- level coder or president, you cannot they they can outprocess us. They're faster, they're quicker, they're cheaper. So the

only question is what are we going to do when everything can be done by uh you know silicon better than carbon up here and and that's I'm quite optimistic about that. I wrote a blog a couple weeks ago about uh humanity unchained. I think we're just going to go. We can do whatever the hell we want, which is going to >> Yeah, I think that's a very natural next question I was going to ask is what, you know, from a >> work life balance standpoint, from a health standpoint, like where do you think this this puts us in as a as a world, you know, because you know, there's been many of these points where whether it be physical automation with the industrial revolution that have caused for some improvements here, there, and everywhere. Like where do you think this is going to make an

impact on us practically? Yeah, I just think we'll do human stuff like what what what am I in? I love, you know, learning stuff. I'm a curious person. I'll just go to read all the ancient philosophers, Socrates, Plato, which I don't have time because I'm too busy paying my mortgage. And same with, you know, you just find out maybe some guys love playing basketball. They're going to do that. Others might nerdy ones want to play chess. They can do that all day long. And we can do what the hell we want basically because uh, you know, everything will be lower marginal cost and the cost of everything will be marginal. Um yeah, there's just a whole different economic system and so we'll just be able to do what the heck we want and um yeah and it's yeah so yeah money is not so the

capitalist system and money is super important that that will sort of dissipate. Uh so the most important things will be the moments we have. >> Yeah. No, I think that's that's interesting. And and do you feel like just from a work week standpoint, uh people have been claiming even like a two week workday could be coming or sorry two-day work week could be coming at some point here soon. Do we feel like any like drastic um >> changes like that for the average knowledge worker would would come to >> Yeah. Knowledge gone in five years probably 90%'s gone, right? All the McKenzie they're out of jobs. Thank god because they charged like you know so much overcharge. So yeah, thank god all that crap will go away, right? And these people who overcharge will just be like, "Sorry, you know, I can I just type

in a prompt and it can do like a thousand times better than you can, but thousand times the cost." So all of that knowledge work crap is gone, right? Just completely gone. I never like that word anyway. What is knowledge? It's all So that's gone. That that's the first thing to go. The hardest thing to automate a plumber is electrician like the fiddly little, you know, with the hand. But robots can do are getting really good and they're on on an exponential curve too. So what I'm trying to say is that um yeah it I it's you know I've been in this field 12 years so I may sound flippant but I've had a long time to think and hear every argument from every direction and I think I see the trajectory that we're on and unless something really bad happens like you know we

have some sort of nuclear war I think we're just going to end up with all this free time and and the only thing we'll have to do is figure out how to use that. I mean, personally, I love free time, so I'm not going to have too much of a problem finding things to do, but yeah. >> Yeah, I'm happy for you. I mean, I, you know, people uh tend to be fans of free time usually, um, but >> yeah, >> I think it'll be really good for, you know, the free time, not only of um you from uh, you know, I you what I think is going to be very interesting >> is I feel like arts and stuff like that as much as >> Yeah. uh maybe AI kind of took away from it a little bit somehow originally when we didn't think

that was going to be the case. Art's going to come back in a big way, I think, because of the the the free time, right? And and not only that, the work that we do in general, I think, will be more quality andor more what we will enjoy to do. >> Yeah, definitely handmade go back to the artisan arts and crafts, right? Things we do because we love doing them gives us such a great sense of satisfaction. Most people they go in they do they sat in cubicles you know I'm not trying to diminish anyone but um it doesn't really give them that much of a sense of purpose like most people if you ask them I'll get you know you can live for free and uh you know do what you want or or go into your cubicle every day get your sense of

purpose there most people would choose the former and um that's going to be flippant but yeah it the arts will come back the music you know I mean art uh the writing you know just go to town just do you know do do create music for two years and then then move on to painting and then go and play chess for six months, you know. So, yeah, it just, you know, I guess, you know, what I'm trying to say is think of the Star Trek future. The robots will help us build starships. We'll travel the universe, right? Yeah. >> Yeah. No, I think I think that's going to be uh be interesting. And it's so funny that um you know, and and maybe this is my skepticism. I I hope that this is the case. I think this is the ideal of of how it'll

turn out. And I'm >> I would be very happy with that right to be to be honest. But um there is a bit of skepticism in me in regards to whether we'll eventually get to that point cuz I feel like we have had massive strides forward in so many ways and other advancements um in time and then like it ultimately still led to the you know improvements in work life balance but not to this level. Would you just say that the technological improvement is just so outstanding that it would be nearly impossible for us to not have this occurrence of free time? >> Yeah, I think this time's different again because we're automating intelligence, not just manual labor. So, we've already automated most power manufacturing anyway and a lot of factories are automated. All the assembly lines are pretty Henry Ford invented it pretty much

all been automated just with mechanical robotics and you know rules-based computing systems uh programming these things. But now they learn as they go. These new uh AI is different in machine learning is from normal rules-based computers and that they learn just like we do as they go along. So you train them. We have to train these things from you know McKenzie consultants or factory workers. We have to train them but they can be trained very quickly within a year. They can be kind of PhD level, senior Mckenzie associate level, you know, presidential level. um because they we can tra it's the silicon is uh much quicker than biology is that's that's just a fact. Um so yeah so yeah we we we will get just get used to this future um yeah unfolding as we go. I mean we can dumb things can happen. We

we still have primate brains unfortunately. Um so yeah we'll probably you know the wars that we see in Gaza and Ukraine are so unnecessary given the trajectory I've just described. absolutely no need. That old uh reason to go to war about territory is just sort of not a reason anymore um when everything can be manufactured and done you know super low cost. So, but that it's going to take a while for us to adjust, right? Our monkey brains to say, "No, you know what? If we try to go to the stars, do the Star Trek thing instead of just killing each other." And why? Um, yeah. So, so that's, you know, that's the thing that I'm most passionate about is to try to steer the future in that direction rather than we go back to being cave men and women because that that's a possibility,

too. And and Yeah. But we Yeah. I don't want to let that happen. >> All right. Well, well, you know, I think that's a that's a good place to to kind of to close things out. My last question would just be for you. Is there any other things that you'd like to >> to to state, you know, about what you're doing and anywhere that you'd like to plug for people uh to go to, especially with Deep Learning Partnership and um I I deeplp.com, but um anything else you'd like to to plug? >> Yeah, deeplp.com. Yeah. So yeah, we've been we've been a little bit esoteric here in our conversation and yeah, I hope that didn't put people off too much. I mean, we still live in 2025. Um, you know, we still have rent and mortgage pay. So, >> we actually help people on this

journey. So, every few months a new models released. Gemini 3's neck cloud 4.5 and GBD6, right? They just get better and better. So we help people kind of look a little bit ahead of where they are now and what's it going to be like in three months time when these models can do maybe you know a hundred times more tarts than they can today and that's what we've seen right since release at Jet GBT in 2022 but even before that like I say from 2013 to Chat DBT there was this improvement exponentially as well and people weren't aware of it because it wasn't in the mainstream. So what we do is we you know we've been on this journey for 12 years. We come in we we have a very good idea where the techn is going to be in 3 months 6 months. I

don't think anyone can tell you where where it's going to be in a year or two years. Um and we just hold your hand. We do um discovery. We say what data have you got? Uh if we train it on these models we you you can deliver X performance and save Y in cost. Um and then um we and then we do a pro prototype proof of concept and sandbox environment uh make sure the results are exactly you know as claimed and then we can put it into production for you. We can help you scale uh in in many different parts of your workflows and your company workflows and that's that's what we do today. Uh as well as looking forward to 10 years where we can you know go all sci-fi but yeah we we we're here to help. >> Well thank you so

much. I appreciate that. And you know, we're really excited to uh we're really excited to have you on the podcast and we thought you had I thought you had some really interesting and and cool opinions. So, thank you so much for kind of you know, guiding us through where your head's at with all this kind of stuff and uh we're really uh really happy to have had you on. >> Yeah, thanks Demetri. We really enjoyed it. Thank you. >> For those of you who are also not subscribed on our YouTube channel, please make sure to do that. We primarily get a lot of viewers over there, so we'd appreciate that as well. Thank you so much for watching this episode and we'll see you in the next one. Bye guys. [Music]