New Jobs of the AI Age | AI IRL

do you think your job is safe from disruption by artificial intelligence what about ours will future media outlets need people to sit on passably comfortable stools and read self-referential scripts into a camera we should all be thinking about this AI is affecting almost every line of work you can think of whether in the form of recruiters automatically screening applicants for Vital Blue Collar rolls or white collar workers being replaced with increasingly flexible software but that's just the surface and contrary to popular knee-jerk opinion it's really not a doom and gloom scenario already entirely new types of jobs are being created on this episode of AI IRL you'll hear from this person who's already doing one of them we started interacting with these chatbots especially chat GPT when it came out was you know doing the very basic prompts you know and and getting the outputs I'm like oh this isn't what I want how can I prompt it in a better way and from people whose jobs involve hiring them my belief is that every role will require some level of prompt engineering skills we might stop calling it prompt engineering so if you've never believed you'd hear a genuinely exciting conversation about the future of recruitment prepare [Music] yourself Dr Freda Polly it's great to have you here thanks so much for coming in you have a long background in technology and ethics and Ai and Neuroscience and all so we could ask you a huge amount but we want to talk about jobs and employment and I'm curious to get your view of where you see the working landscape changing at the moment as a result of AI I think there are a lot of different ways that things could go um I think obviously with the Advent of generative AI you're going to see a lot more of that used right so I think there will be a time probably now but if not now in the not too distant future where you will have chat GPT writing the job descriptions and chat GPT answering them um so you're going to have computers talking to computers uh you know which I think I'm not sure what that will do for the quality of you know applications and recruitment but definitely you see a lot more usage of generative AI on both sides of the spectrum and then when we think about sort of non-generative just sort of plain old regular boring AI uh I think that will only continue to um be more and more prevalent in the recruiting space for sure as we think about you know automating resume scanning automating other types of candidate evalu ation interviews etc etc and I think some of those Trends are fantastic because I think it can really lead to um looking at populations of candidates that currently are not being evaluated I think there's a huge opportunity for really opening up the aperture when it comes to the employer side what have you seen happening this year when it comes to AI based hiring tools what are they doing now that they couldn't do in the pr pre- chat gbt era so my perspective in any case is that that is more where you know we're using AI to write job descriptions or we're you know using it in ways that you know while might be making you know a hiring manager a recruiter's life easier is not going to have as much of an impact I don't think on the candidate perspective right because at the end of the day like who wrote the job description I don't know that that matters as much it's just making things more efficient on the employer side I think what candidates are much more concerned about right and employers are also increasingly focused on is the AI that I'm using to make decisions not necessarily to write job descriptions or you know sort generate sort of stuff that I'm putting out there to the candidate but the decision-making aspect of AI which is not the generative type how is that how is that working because I think that's what you know people are most concerned about let's move into the interview process because and this is funny actually because our exec producer yesterday we were chatting around all this made a really good point that in future you basically want to have an interview to figure out do you want to sit next to this person for 40 hours a week not somethingi is going Sol how does the interview process change in this context like do you need do we should we be exhibiting better more human personalities or characteristics in order to be getting a job assuming we've beaten the AIS who are fighting each other to even get us in the door well so let's say we've gone through the first process which could have some sort of an automation to it and by the way generally speaking when people are using these tools to select candidates at the top they're not just using an AI they are using an AI in conjunction with whatever other screening tools the recruiter will be looking at so it's very rare that it's like zero1 decision based on an AI at least from our perspective at harvor once you've gotten into the interview stage then I think generally speaking we work with companies that are not automating that process because I think then it starts to feel pretty creepy to the applicant like there's been a lot of talk about you know facial analysis and you know sort of you know automated interviewing uh you know sort of analysis that I think generally tends to be a little creepy for a lot of employers just facial analysis in an interview process what I mean maybe I heard I've not seen what is this so there was a fair amount of controversy I would say maybe a few years back there was a company called Ti view that was doing a lot of facial analysis meaning they were sort of analyzing aspects of a person's face then using that information to automate a yay nay decision for you know whether someone should pass the interview so the entire interview didn't involve any people in terms of you know on the hiring side and what they were looking at was pretty questionable I think people felt uh in terms of its scientific Integrity they've since discontinued it there was a lot of uproar about the whole thing and they sort of said you know we're not going to do this anymore not to say that there aren't other people that maybe are still doing it so we've seen where that can go yeah so that would be kind of an example of like hopefully something that won't get repeated right because it was probably ill advised from the beginning when we think of disruptions to the labor force in the past I think the car manufacturing industry in Detroit for example had a really big impact on the folks that built their lives around that one city do you think that there's potential to see kind of an AI Rust Belt of sorts as AI continues to gr I mean my personal opinion is no um that might be somewhat contrary to what others are thinking but you know I have a couple data points for that first of all I mean I think there was the last decade people felt the same way with sort of like non-generative AI everyone's like oh my God it's going to take away all these jobs and the robots are coming and you know all these like scary headlines and I don't think we saw that you know what I mean and if anything the labor market now is way better for um an applicant than it was 10 years ago and that well that actually has nothing to do with AI it just has to do with Trends right it's just basically um you know demographics basically that we've known about for a long time with a lot more people having retired from the workforce and we just haven't sort of you know don't have that in influx of of folks coming in as well as you know a lot of early retirements due to covid etc etc right we really didn't see this massive disruption of like all these jobs going away because of AI right and I would say that that I believe very strongly is going to be the case this time around I mean think about I was thinking about this just a second ago think about an airplane okay an airplane is basically flown by a computer did you guys know that I mean the pilots don't do a whole hell of a lot right um however I would be hardpressed to find a person that would fly in an airplane without an actual pilot or Co and co-pilot terrifying right in a plane we have trains in London that completely drive lless but it's not the same as being exactly 30,000 ft up and so my point is simply that I think for so many of the jobs that people are worried are going to get automated away whether it's you know physician or lawyer this the decisions that are being made I think people would really worry about there not being any kind of human oversight we saw that in recruiting as well I mean you mentioned um demographics people early retirement from Co things like that do you think there's a risk that AI could actually get blamed for impacts on jobs that just because it's convenience but actually the reality is is that some people are just lazy or that there's covid or there's other broader issues well so they did this really interesting study where they looked at you know National attitudes towards AI right um so in China for example the attitudes were very favorable and sort of in Western you know countries they weren't and really what they found is that it was really to do with how people's lives had either improved or not improved during the time that AI came to be does that make sense so in the last 10 years generally speaking you know lives in you know people's lives in China had improved from you know sort of more you know poverty levels to you know better versus in the west as we've all talked about you know there's income stagnation growing inequality Etc and I do think a lot of that tends to happen right like what is new oi is new oh that must be the problem right so I do think that there's unfortunately some of that going on thank you so much for joining us thank you guys y Justin you're a senior director at gitlab how do you describe prompt engineering to the uninitiated yeah great question I use an analogy of a course I took in university called interpersonal Communications and what that course teaches you is how to communicate with a human and understand Nuance tone empathy nonverbal cues all these things that aren't obvious if you just read a transcript of a conversation like we're having right now uh those contexts those pieces are often what you need to pass to a large language model to get better results so if I simply pass Bland factual text and expect a great result I may not get it but if I can pass context tone parameters of what I've I'm expecting or foundational knowledge about the topic to the large language model then the results I'm going to get are better and it's going to be a better conversation right so there's a human value an element that you can't just capture in one basic fact I mean the joke I've heard this year is that the most popular programming language in 2023 is English I'm curious because you've had some conversations recently about whether kids should go to study prompt Engineering in school what would you tell their parents if uh if they said should I let my kids go to school to study prompt engineering it's a good question I I would steer them away from um pursuing an explicit career in prompt engineering inste what I would tell them is hey go go pursue your passion what are you passionate about are you passionate about teaching or philosophy or psychology maybe it's traditional software engineering but then figure out how to layer in those prompt engineering skills they can be selftaught most are we've all learned these things over the past several months um but learn those skills on the job or within the context of the work you're doing in University uh versus you know pursuing an explicit degree in that I don't know I've seen some of these Buzzy prompt engineering jobs pop up that have like $300,000 salaries I would kind of maybe steer my kid towards that direction but what what's your take on it the challenge is these large language models will get better and more sophisticated over time so if I was entering University today four years from now I don't know what that's going to be and I honestly don't know if there's a professor teaching those courses in University that can train me effectively so the job might not exist in a few years yeah and really the benefit the reason why those folks are commanding such great compensation is that they also possess domain expertise in their field so the foundation not just how to trick the that's right exactly so if I know everything about Building Bridges right um but I don't know anything about prpt engineering I can understand if a model is hallucinating if it's giving me the right answers if it's helping assist my firm in building great Bridges um and that is more powerful to use in that context than someone who's just really good at manipulating a model now that bridge expert should also gain those skills to use that tool just like they've had to learn learn how to use Microsoft Excel or some other tool right um in their career but it's that domain knowledge that's more valuable in my opinion than pure prompt engineering skills so they take the major in architecture and design but they do a minor in prompt engineering not the other way around that's a great way to say it yeah ex I would I would advise that I agree yeah a lot of the conversation around artificial intelligence revolves around the fear that it's going to displace people's jobs and so prompt engineering whether it's as a job description or a role overall is one of those jobs that's emerging that could potentially gain some traction as AI continues to grow however it doesn't necessarily require a college degree and you're seeing a lot of YouTubers and other personalities basically teach this stuff to people online without having to go to school what do you think about that I love it and so typically these very technical things over the years have been reserved for very good Engineers to implement but now you can watch YouTube video you can learn a little bit about how to wire things together and you can build something really cool without any formal training um that is magical to me and I think that future is in front of us empowering more and more people to contribute in different ways that they never have been able to before I've never discounted learning from YouTubers ever so why would you hire a prompt engineer versus somebody who can just do the stuff without needing to prompt yeah you know I think about it is an efficiency gain so someone who has prompt engineering skills that can leverage these large language models they are could be 10 times more efficient than someone who doesn't have those skills feel strongly about that person having the domain understanding in context to prompt that effectively so if all they were good at was prompt engineering I probably wouldn't hire them but if they had a really specialized skill or some domain understanding of a topic that we needed and then those prompt engineering skills they will likely be 10x more effective in their role than someone who doesn't have the prompt engineering skill and what do you think prompt engineering will look like in 12 months time yeah in 12 months time um I think more and more people will use it we're so early right everyone in Tech is talking about it but it hasn't sort of crossed the chasm where everyone else is using it and so even within a technical company like the one I work in you know our marketers aren't fully using it yet our sales folks aren't fully using it yet the people that aren't inherently technical in their role aren't quite using it yet to the full extent and so what we're going to see is more and more Fields more and more domains more and more roles start to leverage these tools to be effective in their role do you think these skills are going to be taught say in elementary school middle school or even high school um should they I think like in anything teaching students how to think critically how to ask good questions questions how to challenge an answer and a result and deeply understand why is a very very transferable skill to any field right apply that to prompt engineering and learn the technical nuances of it from YouTube and those people will be successful in the field so let's say someone's watching this and you've convinced them that prompt engineering is a fantastic way for them to earn their next mail uh what are like three top tips that they should have to becoming a prompt engine sure yeah uh the first one I'd say is get a free account of any one of these tools and experiment the best way you can learn is trial by error so start doing that today um the second I would say is actually do a little bit of research to understand how these models are actually built you don't need to become an expert by any means in a ml but know that these actually aren't sentient beings you're having a conversation with they're just machines that are really good at pattern matching and then lastly start to experiment with multimodal or multi- prompt queries so have a conversation with one of these systems and pass it more data and expect to get the result you want out of the 10th query not the first one because when you do that you're giving the model a lot of context and domain understanding to yield better results down the road but you may not get that on the first query well Justin Farris thank you for prompting such an interesting and time had to discussion I've been sitting on that since we came in this room very good Tiffany Jansen thank you so much for joining us thanks for having me here so you're a software developer and the founder of tiffan tech and if I understand correctly it's a community that talks about coding culture all things technology across YouTube Tik Tok and Instagram you didn't start off in the tech world so why don't you walk us through a little bit about how you ended up in it when I first graduated high school I thought you know I I really want to get into fashion because it includes you know uh building different things the creativity side of things but also using the Lo logical side of things as well and and from there ended up going to school for fashion uh got an opportunity to go overseas to Hong Kong for modeling I'm like you know what let's go travel the world let's see what's out there so I did that and it was during my time there uh I was like I really am craving more furthering my education again I just don't really know what that looks like it's like okay I'm going to go back to Canada where I'm from and go to school and I went to University for graphic Communications management which is essentially what it sounds like it's a bit of graphic design a bit of management uh courses did you know what role you wanted to end up in I knew I still wanted to stay in fashion I was like okay I'll apply this to the fashion industry to some capacity and it was in my last year of University it was required to take very basic coding courses I was like I would never have taken it otherwise I like this is incredible like it's I get to whatever I comes to my mind I can actually build like this is amazing so then I started teaching myself some JavaScript and I'm like oh wow this is amazing and from there went to a coding boot camp and uh from there ended up working at large tech companies startups and it was really through that Journey which led me to starting Tiff and Tech so we've talked a lot about prompt engineering which is something that I I know you're very interested in and do a lot with can you sort of describe a little bit about how you used that um skill set for sure it's really you know Nate it's evolved over time like many of us when we started interacting with these ch Bots especially chat GPT when it came out almost two years ago now uh which feels like on one hand just yesterday I was you know doing the very basic prompts you know and and getting the outputs I'm like oh this isn't what I want how do I how can I prompt it in a better way and that's really how it started and then I started using uh tools like uh github's co-pilot which essentially you're prompting it as well but it's more related to code specific tasks the same kind of goal with prompting and then having a specific output and it was through that that I kind of got obsessed with understanding how I could prompt better and started doing a lot of reading on it ended up taking some courses on it online and it's kind of evolved ever since just now and now you offer courses kind of on your own YouTube channel for others to learn I mean so it sounds like it took a lot of trial and error there wasn't like a specific course you go to to college to kind of go back to school for this kind of thing it's independent kind of work correct exactly it's something that you know if you are interested in it you can learn at home on your own you know I I do come from a technical background prior to getting into uh prompt engineering but you don't it's not required necessarily you can come if it's something anyone is interested in it's there's so many resources available online that you can you know start learning it today essentially isn't that an amazing part of this new era it is and I think you know you're a really good um example of of of a key Point here which is you know you you spent a lot of time pursuing fashion and modeling but actually you ended up doing this and now essentially your job is doing something that did not exist and could not be aspired to um and I wondered can you just tell us a little bit about how long you think what you do now has actually existed for because a lot of people talk about there being new types of jobs and New Opportunities being created and you're kind of you're kind of one of the people that's got one of those jobs it's super interesting and it's kind of evolved over time from you know from Pure software development to starting to bring prompt engineering into my day-to-day I think sometimes when we think about the future and future jobs it's this Stark that all of a sudden one day we'll be like oh this job exists but I feel like from my experience it's evolves over time and that's what I found with uh incorporating prompt engineering into uh into my day today um you know I know there are so many companies who are hiring specifically for a prompt engineer and then there's also people like myself who use prompt engineering within their career within their day-to-day for someone that is preparing for a role in the future in this new age of AI what is the greatest skill that they should be developing now you know it's funny I actually asked this recently on um my social media and I just wanted to hear what other people had to say and it was a poll and it was between is it it was General it was between would it be a soft skill or a technical skill and because my audience for the most part is pretty technical uh it was quite significant people uh voted that technical skills are going to be the most important skills in the future uh however even as a technical person I really disagree with that I think soft skills especially the willingness and ability to be open-minded and continuing to learn be a forever learner is more important than ever and now this might not be a specific skill but having that mindset in the sense ofch technology and artificial intelligence is changing and growing so rapidly that there are two ways you can you can look at it you can embrace it and grow with it use these tools or you can do the opposite but then in turn someone who's using this technology will continue to excel are you talking about the value of human like the value of being a person basically having having longevity still exactly so really there are always going to be certain roles where people want people is that your you believe that 100% it's that interact human interaction that uh We crave or that we there's this trust Factor as well with um you know interacting with other humans versus technology uh so yes I I think it will remain important and even maybe grow more important as technology continues to become more ingrained in our lives we'll crave that even more Tiffany thank you so much for joining us thank you so much for having me this has been great I [Music] oh [Music]

Channel: Bloomberg Originals

Views: 34,603

Keywords: News, bloomberg, quicktake, business, bloomberg quicktake, quicktake originals, bloomberg quicktake by bloomberg, documentary, mini documentary, mini doc, doc, us news, world news, finance, science

Id: pI7eO7X_wug

Length: 24min 0sec (1440 seconds)

Published: Wed Nov 29 2023