2822: JLL - Reimagining the Workforce: AI's Role in Job Creation and the Evolution of Skills

[00:00:00] Is the fear of AI taking over your job a reality? Or are we on the brink of a new era of job creation and innovation?

[00:00:30] How AI is reshaping the workforce? Also, why there is a surge in demand for new skill sets? What this means for the future for knowledge workers in 2024 and beyond?

[00:00:43] So our conversation today will be around AI, artificial general intelligence, AGI as well. We'll discuss a few implications around that.

[00:00:53] So I don't want to question, is the AGI that we should be worried about or the immediate impact of AI on our jobs? And those ethical considerations such as bias that surround this topic.

[00:01:05] So we'll shed light on all this and also draw some of the, I'll also hopefully draw from his extensive experience at JLL where he's at the forefront of integrating AI into real estate advisory services.

[00:01:19] Now, before I get today's guests on, it's time for me to mention the sponsors of Tech Talks Daily and in an era where digital security is non-negotiable, legacy managed file transfer tools. They simply don't cut it now. So that's where kite works comes in.

[00:01:35] Revolutionizing the MFT landscape with unparalleled security credentials, including the much coveted FedRAMP Modra authorization. This isn't just about compliance though, it's about offering a secure, efficient platform for today's remote workforce.

[00:01:50] So with kite works, you can benefit from advanced file sharing, email security and customizable integrations all within a platform designed to safeguard your most sensitive data.

[00:02:02] So don't let outdated technology compromise your security. Step into the future of secure managed file transfer. Get started today by going to kiteworks.com. That's kiteworks.com where security meets sophistication. But now it's time to get today's guests on.

[00:02:19] So book a look and hold on tight as I beam your ears all the way to Tel Aviv where we're going to try and demystify the complexities of AI with today's guests.

[00:02:30] So a massive warm welcome to the show. Can you tell everyone listening a little about who you are and what you do?

[00:02:37] My name is Orr or Hilch. I am the chief data in AI architect JLL. JLL is the language style, CRE service of his company.

[00:02:47] I have the joy to company following the acquisition of my startup, style and AI where I was one of the co-founders in CTO.

[00:02:55] So I've been doing technology and prop tech for the first six years now and part of the bunch of other other cool stuff. Yeah, so excited to be here today.

[00:03:06] I'm excited to have you on here because there's so much hype around AI at the moment. But I think we're starting to look beyond that hype and start to talk about how can we use this?

[00:03:16] How's it going to make our work better? How's it going to make our life better? How can we work alongside it rather than replace people with it?

[00:03:23] With AI predicted to reshape job market, similarly to past technological revolutions and we have here. We have been here many times before.

[00:03:31] So how do you envision a future of human AI elaboration in the workplace? And how can organizations better maybe balance technology adoption with that intrinsic human need for social interaction and trust in professional settings?

[00:03:45] It's such a huge balancing act, isn't it?

[00:03:49] Yeah, it certainly is. And I think AI has been actually helping us out to achieve our goals more efficiently for quite a few years now.

[00:03:59] I think it is getting smarter and smarter and it is hard to keep pace. But if you look at historically the way we have been using AI and how it changed our work,

[00:04:11] it's certainly safe to say that the way that we do our work is always changing. But usually it is the case that because of AI we are able to do more work and then it always ends up with humans just doing more and more business.

[00:04:29] And I can give you an example like a personal example, right? So I started my career as a software developer and that was about 20 years ago.

[00:04:38] And when you consider programming 20 years ago, it really looked nothing like it is today. So we used to have the generation that preceded me was actually using punch cards, actually doing zeros and ones that machine code.

[00:04:54] My generation started writing a standard language which is like a real level language that requires you to understand feedback memory management and the structure of the processor and all these things.

[00:05:06] And certainly with the years of past, we have been introduced to higher level language, like usually we will see in the steepest plot, then see sharp.

[00:05:16] And now you have like super high level frameworks like the React from Meta. So but that essentially if you want to draw a button to the screen, you don't have to know about the quantum mechanics of all the traditional works.

[00:05:29] And I think all of that has been made possible with actually with AI, with things like compiders that you know they take the computer code that you write in some higher level language and compiled it all the way down to zero zeros and ones that can simply understand.

[00:05:47] And if you and then I think there are two observations here, like the first issue considered like the number of the number in the world that when I started that 20 years ago versus today.

[00:05:57] So you have like you know thousands and thousands of water developers today compared to you know 20 years ago.

[00:06:04] And the reason is that it has become a lot easier to write code and a lot easier to build things.

[00:06:10] And I think AI is calculated the right step in that and we can see it with things like GitHub co pilot from Microsoft automatically writes code for the daughters.

[00:06:19] And I think this is kind of the next level of abstraction.

[00:06:22] And just like writing code let's say in 2020 was very different than writing code in the needed 2000.

[00:06:30] It certainly will be a different task to write code in the next 20 years compared to today due to AI.

[00:06:36] But my take is that it's probably going to proliferate in the you know the number of developers just as they did like you know 20 years ago.

[00:06:45] And I see I'm guessing we're going to see the same many other than the science so you know we certainly see it in GLL.

[00:06:53] We are already applying AI to make stuff more efficient and usually the cases where we just end up doing more business.

[00:07:01] Yeah, it's so true what you saying that as I said a few months ago we have been here before we go back those 15 20 years ago they were no.

[00:07:10] I have computing architects mobile app developers that podcast produces rideshare drivers social media managers UX designers virtual assistants such a long line of new roles that have come as a part of technology so considering the widespread discussion

[00:07:26] around AI's potential to both automate and create jobs any insights that you can share on or maybe what sectors you foresee experiencing the most significant transformation and how workers and companies alike should be preparing for these changes right now.

[00:07:42] Yeah, so I think first of all I think I know where should be preparing in the sense that you know you should be using the AI tools and no matter what you do so if you're marketing your controls you can use tools like Jasper

[00:07:54] and whatever to you know make your content writing more efficient if you're a coder you should be using it have copilot or the equivalent.

[00:08:02] And if you are in a CRE you should definitely plug your theory data sources e to some AI engine that will just make you more efficient in your work.

[00:08:12] And I think if you look at the segments that are probably most points for disruption it's it's a you know because I think that most of the AI revolution that we're seeing in recent years is about large language walls which naturally makes the language heavy areas more prone to change.

[00:08:32] So if you look at the you know seems like legal right to legal is super language heavy and even the gel with our proprietary along gel G.

[00:08:42] We are seeing a lot of our little stuff in getting it to analyze contra apps to make the little jobs process faster so every area which is language heavy is probably more points for disruption in CRE that also includes things like policing so you know there are a bunch of these like super long least agreements that could be causing.

[00:09:01] That could be complex to analyze and this is the great area for AI to help us with same goes for things that got property valuations where you need to write pretty lengthy reports so all these areas that are language heavy are probably going to be.

[00:09:18] Be the first to you know to enjoy the revolution.

[00:09:23] And something I was thinking about the other days the two different sides of the coin because on one side we've got the rapid advancement in areas like generative AI but on the flip side of that there's also relatively slower progress in things like robotics this got me thinking how do you assess the current state of robotics in a I development and are there any challenges and opportunities that lie ahead for integrating more advanced robotics into various industries because again is so much going on in this space isn't it.

[00:09:51] Yeah so that's a super good point and you know I actually this is the discussion I have a lot with my colleagues about.

[00:09:57] You know it like it's certainly true that the world while robotics is did the balance in the first the last couple of years the pace of progress in every and every second that's hard we needed.

[00:10:09] He's always much slower compared to software right because it's so much harder to eat with hardware you know we have to build things right.

[00:10:17] So I think that in that sense you know the relatively if you look at like the type of work that required is actual speaking of presence from people to do right so operating machines.

[00:10:32] You know everything that's you know everything that's you know that actually demands like you know these emotion type of skills I think these would be the last professions to be disrupted by the AI because robotics is simply not there yet.

[00:10:47] I think generally we are actually is helpful in the progress of robotics because many of these robotics labs these simulators to improve the hardware and Jedi in that sense is really helpful because it can generate a lot of synthetic data so these.

[00:11:04] These training simulators could be a lot more efficient now but stay I think it is the different to state like it's lagging behind as you as you mentioned.

[00:11:15] So yeah and the thing is the ability that we've always been between hardware hardware software and AI is the same as AI changes the nature of jobs such as anything from programming and document review in the legal industry.

[00:11:30] What are your thoughts on how professionals can maybe adapt to these evolving roles and how significant the role of creativity and innovation is to staying relevant in an AI driven job market because there are so many different skills more human related skills that AI cannot compete with it's almost like we need to leave the machines to do that repetitive mundane boring stuff and get back to being human and move away from the robotic tasks right.

[00:11:57] Yeah that's exactly true and I think like the skill that are going to be more and more important.

[00:12:03] I think around the things like language that's like the first right because finally enough the way to get these AI is to do stuff using natural language and then the better your grass that languages you know the better you can the results you get from the model and because most of the written content on the web which is what it is.

[00:12:26] So I think that these models are trained on in the neighbors and it's actually probably the case that you know the better you would be an English demorty could leverage these technologies.

[00:12:37] So I think having a really good grasp of the language and the special English would be more or more important.

[00:12:44] It all has always been important but I think it's just going to get more important.

[00:12:50] The second thing is the kind of what I can't like seeing at a little better praising AI is that many of us are kind of like becoming analysts in a sense.

[00:13:01] So you know we're talking to the AI we were asking the AI to do something maybe I want the AI to write a function for me or to do something else and then I spend most of the time reviewing and the kind of like making slop fixes and actually asking the item fix that.

[00:13:19] So I think it's definitely the case that you need the end to end understanding of things to be able to get the AI to work well at least in things like software engineering.

[00:13:33] But I think it's kind of like it's probably better to be the case that you know how excited you are in universities they always kind of focus on theory and not so much necessarily practical stuff right.

[00:13:45] The one of the one of the of the complaints students always had for university that people put aside that you know we focus on being like algorithms and theory and not so much about programming right so when they when you are like a fresh graduate out of the university.

[00:14:02] The problem skill are okay but then you actually learn to program in the job right and then this will be like a complaint students would always say but now it's I think that's kind of like the the ideology so moment of university because that it's kind of like seems like the actual technical programming still is becoming less and less important and the actual understanding is becoming more important because in this age where most of the coding could be done by AI.

[00:14:31] You'd use it actually the understanding that's more important so that's a that's a fine bed interesting.

[00:14:37] And also of course with the emergence of prompt engineering is a critical skill set in the tech industry anything you can elaborate on the unique blend of skills required for success in a field like they said is anything that you foresee in the demand for prompt engineers influencing the broader tech hiring landscape because that demand for prop tenderness is still playing more and more in our new speeds now.

[00:15:00] Yeah so I think I think it's it's still in the case that you know it doesn't matter to what business line you would hire someone yeah probably still the case that you would want them to have a domain expertise right though you could be you know I would like to let's say you should consider one of the many different lines that we have a gel which is around CRE.

[00:15:25] I'll take anything for example so while good prompt engineering skills are definitely important we definitely also want the candidate to have a good grasp of the business.

[00:15:36] So I think it's still it's still quite fun the future until we see like a pure prompted during jobs to always be coupled with domain expertise at least in the short term yeah but that's very good thing to think our earliest conversation about.

[00:15:54] About the no the grasp of language right because eventually prompt engineering is it's all about kind of asking the right questions the right way and you need to have a good grasp of the language and in order to achieve that.

[00:16:08] I'm curious if we reflect for a moment on the evolution from writing machine code to using high level programming languages how do you see AI further transforming the role of program is there any implications that it might have for the education

[00:16:22] and training of future software developers to because I gave so much going on I know he's moving it such a fast pace right now and it's difficult to to know where he's heading but how do you see this evolving.

[00:16:34] Yeah so so the definitely I think I think it's like I would say that I would kind of divide my answer to you first like what we're seeing right now and what we'll probably see you know in the next 35 years I think that what we're seeing right now is that AI is making the veripers at least twice as efficient right.

[00:16:51] So the same developer could probably write twice much code or produce twice as much delivery in the same time and that's super helpful but that's still requires the door print the loop because the AI is really is really is really great at like writing the you know the boarded stuff but when it comes to end to end to actually being able to deliver a working let's say working application or something like that.

[00:17:19] It's still quite lacking. I think that the way that we see this evolving the signature is that and that's I'm talking about at least let's take five years from now.

[00:17:31] I think that most of the veripers will probably be and I would say there would be two types of jobs for developers like one would be the architects ever will right so you would tell the developer actually you know they would need to know

[00:17:48] what they want to do and not so much how they want to do it right let's say they would they would need a specific or hey I want my application and it needs to have a database it needs to have maybe a need needs to run on the cloud right it needs to have

[00:18:04] a security measures and authentication methods and all these things I think many of the question of the how would be left for the AI so in that sense the veripers which need to know about the different components of the application.

[00:18:19] And they could teach specify them and that only takes sense because you know there are different requirements between writing software let's say in startup versus the corporate in an enterprise right.

[00:18:30] So you always need this architecture function the developer does today they will probably need to do the future as well but many of the how would be left for the AI so that's like I would say one type of engineering job by an agent with future and the second is really something which is I can it's probably what I what I referred to earlier as a code analyst.

[00:18:53] It's I think it has something to it's also quite similar to quality assurance in the sense that you would need to look at the results you look in the code and collect verified and see okay that that action does what I what I expected to do right.

[00:19:07] And there are of course ways to automate that as well but I think that we will probably want the developers to actually do the notifications a thing like in the near future yeah so I think that's that's how we're going to see the profession change.

[00:19:22] And if we go back just a few weeks there was some controversial discussions that the world world economic forum about the potential for.

[00:19:29] AGI to mimic human thinking at what are your views are on the ethical considerations and of course safeguards necessary to ensure a GI benefits society without this empowering humans because whatever we talk about any form of AI it's not long until we start talking about the ethics and the dangers of moving fast and breaking things

[00:19:50] because we've kind of been here before again to but what do you say the developments here.

[00:19:56] So I think my take is that it's we have like I would say more urgent to ethical concerns about AI yeah for instance you know if you look at things like bias and potential ethical issues with you know the outcome to right of the Davis.

[00:20:13] And this is an issue that I think we already have to get right and we are working hard to mitigate.

[00:20:19] And I think that the these are more like practical issues that you know we're seeing in the wild right.

[00:20:24] And they are definitely the type of picture that I think we we as a kind of consume is humanity need to solve probably before worrying about a GI because a GI it's like I would say still very much unclear like if and when we are going to be able to build that I think yeah when when a GI happens it will change everything.

[00:20:42] But but like my point of view is that you know it's still so far in the future and it's really hard to and hard to even imagine what the challenges would be.

[00:20:51] So no you could go back as far as you know the development of linear algebra and matrix multiplication like you know like hundreds and thousands of years ago which is today you know the technology the technique that it's underlying all of the GPU conversations right so whatever eventually when you

[00:21:10] when you look at these the high level AI boggles that we talk to eventually they compiled to neural networks which are further broken down into matrix multiplication operations that are being run on GPUs right.

[00:21:24] And I think kind of started to worry like worry about a GI right now is it's almost like pouring about a GI back when you know someone will develop the metrics for complication hundreds of years ago.

[00:21:36] Because I think we are so far away from creating something which would be like a true you know general purpose type of intelligence thing that it's quite premature and part of that why is by the way because I think but times when we think about a GI we actually think about human evidence or something which is you know it might be super human but it is human study intelligence.

[00:22:04] But it actually the case that human intelligence is also it's also quite specialized right we don't really have an example nature or someplace else to you know and and actually generic type of intelligence.

[00:22:19] So I would say yeah we definitely need to worry about the ethical concerns of a way I but that's not something new I think that that was also concern in the days of deep learning when we know when we use things like what it's coragulism and stuff like that.

[00:22:36] They might have viruses based on the training data I think that that's a much that concerns me a lot more than you know kind of a futuristic thing with the GI.

[00:22:47] And something I try and do every day on this tech podcast is learn more about the impacts of technology the benefits of technology in every corner of the world not just Silicon Valley and the Bay Area and a somewhat based in Tel Aviv Israel or a K.A start up nation are in now tech or over there how do you perceive the impact of regional difference in AI development and adoption and are there any unique challenges or advantages that you face in Israel in the global AI landscape.

[00:23:17] So if we zoom out and look at that bigger picture.

[00:23:19] Yeah, so that's a very good question.

[00:23:21] I think generally speaking I think there is something quite uniquely local to AI and that's the fact that if you actually want to develop a you know a foundation model you need to have a lot of resources.

[00:23:34] And if you look at these type of like foundational model providers you don't see them coming from you know from most countries right you obviously you have the US eating this right.

[00:23:46] You have the UAE with TDI and the Falcon model which is one of the largest open source of large liquid models.

[00:23:56] And you actually have one in Israel too from the company called the 21 so I think in that sense I think Israel is an interesting thing in this game in this race for LMS even though it's a real people country.

[00:24:11] I think like garlic, like making many other areas of tech you would see the Israeli startup nation focusing on things like fiber security and indeed you have a bunch of AI startups around that field here.

[00:24:26] So that of a little insecurity more broad AI security but that's the type of things that we're seeing are here.

[00:24:33] Fantastic exciting times and I cannot thank you enough for coming on here and sharing your insights around everything AI and a G.I. today but before I let you go I always like to have a little bit of fun with my guests so I'm going to ask you now if you look back at your career what's the funny install most interesting story that has happened because I would imagine that you you picked up a few stories on the years but anything stand out that you could share anyone that you can share with anyone that is clean and you're able to share.

[00:25:02] Yeah so we're fortunately actually had a lot of funny stories about the states I made in the programmer and I think one that I might share today is that when I was like sort of like in the beginning of my career I was a software engineering one of the large banks in Israel.

[00:25:23] And I was working on this campaign to basically distribute a call from the call center to the bank's clients and offer them some self kind of banking service I think it was something like a pension fund or whatever.

[00:25:38] And then when I was writing the code to generate I was supposed to write a code that generates the actual phone collectivity on the C of the Rm of the call center.

[00:25:48] The call center had like maybe 40 50 people working working for it and just making the call all day and then you know how it is.

[00:25:59] It is actually should look at the database systems you have tables for example with the client names things like that and their phone numbers.

[00:26:07] And then you have the data fields that you know they're also always supposed to have a valid right.

[00:26:12] If you consider something like you know the idea of a person you would expect that to always have a value right because no person has no idea.

[00:26:20] So when I was writing the code to generate the phone collectivities for the CRM I was assuming that you know that that field will actually always have a value.

[00:26:32] But it did turn out that there was this one client who actually for some reason probably a bug or whatever.

[00:26:38] It did actually have a value for the Rmd and then forward might problem by program ended up creating like hundreds of hundreds of phone collectivities for the call center all words the same person.

[00:26:50] That person got like 100 gold phone call is like like doing one hour and he was not happy about it but eventually everything was settled so yeah that's kind of one of the story of their own remember.

[00:27:02] What a great story absolutely love it but anybody listening just want to find out more information about your work what you do any of the topics we explored today or just contact your team.

[00:27:14] Where's the best place to start and learn more about anything we talked about today.

[00:27:19] So first of all we have a lot of great content on the jail website concerning prop that AI and CRE and AI in general you could always feel free to reach out to me or my LinkedIn as well.

[00:27:31] Always have been a chat.

[00:27:33] Well I'll get the link added to your LinkedIn under website and the social channels for JLL tech too love chat we covered so much there from how AI will actually increase jobs as previous tech revolutions have proven in the past many times as we've said a few times here we've been here before and also our occupations such as programming document review and writing most won't disappear they'll just take a different shape a different form and that isn't unique to the AI area.

[00:28:00] And also of course looking to the future such a demand right now for AI skills prompt engineering really diversifying the tech hiring landscape and for the better as well I would argue but more than just thank you for bringing this topic to life with me tonight.

[00:28:15] Yeah that's my pleasure thinking about.

[00:28:17] I think it's clear that the narrative around AI and jobs is not about replacement but transformation because the rise of gen AI l l m's and diffusion models almost heralds a new age of job creation demanding a blend of technical and soft skills like we've never seen before

[00:28:36] or whether it be the emerging role of prompt engineers or the evolving landscape of programming the future is bright for those that are ready to adapt.

[00:28:45] And today's guest reminds us that the innovation knows no bounds and even the smallest nations can compete on the global stage this is not just about Silicon Valley that's being impacted here so for anyone eager to dive deeper into AI's impact on jobs ethics and beyond I'd love to hear your thoughts on this.

[00:29:04] And you see AI reshaping your industry good bad indifferent I want to hear it more and also how you're preparing for the jobs of the future so share your thoughts and join the conversation by emailing me tech blog right to outlook calm Twitter linked in Instagram just.

[00:29:20] At me or see you's love to hear your thoughts on this so I'll return again tomorrow with another guest but until next time keep looking forward to the future maybe learn to embrace the change that AI could bring to our world maybe it's not all bad let me know but I'll be back again tomorrow so until next time don't be a stranger.