AI is transforming careers, personal branding, and financial planning - so how can you stay ahead? Dr. Joan Palmiter Bajorek, entrepreneur, investor, bestselling author, and host of Your AI Roadmap, shares strategies from her new book to help future-proof your career and income. She breaks down how to build a strong personal brand, expand your network, and stay financially secure in an AI-driven world. We also explore AI’s impact on job searches, gender bias, and the balance between automation and authenticity. Whether you want to pivot into AI, strengthen your career, or make smarter financial moves, this episode offers valuable, actionable strategies to help you navigate the future.
Key Takeaways:
Joan's HBR article: Voice Recognition Still Has Significant Race and Gender Biases
Joan's LinkedIn: (23) Dr. Joan Palmiter Bajorek | LinkedIn
Joan's Website: Dr. Joan Palmiter Bajorek
Your AI Roadmap: Your AI Roadmap
00:00 Introduction and Background
00:22 The AI Roadmap Book
01:54 Key Actions to Future-Proof Your Career
03:01 Building an Emergency Fund
05:44 Leveraging AI for Financial Planning
07:36 Networking Strategies
10:59 Personal Branding with AI
14:11 The Human Element in AI
19:51 Effective AI Prompting Techniques
20:29 AI's Influence on Job Matching
22:21 Using AI for Personal and Professional Tasks
22:59 Choosing the Right AI Tool
25:08 The Future of AI and Human Collaboration
26:16 Balancing Career, Money, and Joy
28:00 Gender Bias in AI
32:30 Conclusion and Final Thoughts
📜 Read the transcript for this episode: Transcript of Future-Proofing Your Career and Income in the Age of AI with Dr. Joan Palmiter Bajorek
[00:00:00] Joan Palmiter Bajorek: Hi Everyone. My name is Joan Palmiter Bajorek. I'm an entrepreneur, investor, podcast, host, bestselling author. And I am excited to talk to you today about future proofing your career and income in the age of AI.
[00:00:13] Henrik Werdelin: I think I would love to, obviously you have a book. Is it coming out as, as is that already
[00:00:20] Joan Palmiter Bajorek: It just dropped a few days ago
[00:00:22] Henrik Werdelin: Your AI roadmap and it comes with like the slight promise of giving you actions to expand your career money and joy Which is quite a lot And so I don't know where to start because we love frameworks and it seems that you're kind of like in the same camp Maybe you could start with just kind of telling a little bit about what made you, what prompted you to, uh, to get going on the book project.
[00:00:49] Joan Palmiter Bajorek: Yes. What prompted me? It is a companion to my podcast. Um, and what prompted me to do the podcast was that my being an influencer, being a linguist. Suddenly when chat GPT came out, my socials, uh, went wild. Um, and so tripling down on what my audience wants would be, I guess, a piece of that. I feel like I am living a lot of the volatility of what's going on in the market.
I've been in two AI layoffs. I live here in Seattle. I build AI. I like data infrastructure. It's what people pay me to do typically. Um, but more and more of my DMs, how do I get a job in AI? I'm terrified to lose my job. Joan help me is like the DMs I'm getting from people around the world. So this book, as you mentioned, actions to expand, it has frameworks, it has templates.
Like if you choose not to do them, that's on you. Um, but my book is very actionable. Modern careers is part one. Modern finance is part two. So if people are scared of layoffs. Uh, this is the resource I provide them on a plate.
[00:01:54] Jeremy Utley: So just walk us through, I mean, for the folks who haven't picked up the book, uh, or haven't had a chance to, read it yet, but maybe they slide into your DMs and they go, what do I do?
Short of, I mean, the cheeky answer is of course, buy my book, right? Which is great. But what's your, like, What's your 80, 20, the first three things you should do to, you know, stave off your existential, you know, fear or crisis or whatever. , What are the, and I love, by the way, your phrase actions to expand.
What are the actions you recommend?
[00:02:23] Joan Palmiter Bajorek: Yes. The actions I would recommend if there are three things, personal brand. Do people know who you are? Can they find you online? Many people, I know their headshot doesn't even look like them anymore. I'm not going to call it names over here. Um, the second one being your network.
Do you have people that you can be like, Hey Tim, do you know like a resource I should go to? Like what? Like LinkedIn did a research that after COVID layoffs, 70 percent of people landed their next job based on. An introduction. It's not the side route. It's the main route. So do you have a network that you foster? Do you go to events in your field? Blah, blah, blah. Personal brand network. Third one. Do you have an emergency fund? Have you been taking care of your financial house? The number of people living paycheck to paycheck, despite huge salaries. Scares me.
[00:03:12] Jeremy Utley: What are you seeing there? Tell us what you're seeing there. That, that phrase, because I think that's really, um, paradoxical, right? Living paycheck to paycheck on a huge salary. Tell us, what you're seeing.
[00:03:23] Joan Palmiter Bajorek: Oh yeah. Well, I talk about my friend in this, um, in the book, she makes 1. 3 million a year. Um, she's not happy with her job. And I was like, you've been making good money. Like you have a seasoned job. Like, could you just take a break? Could you like retire type thing? And she was like, Oh no, I don't have anything saved like because she spends it and I don't mean to be, you know, it's her personal finances, but like some people, uh, make money and they're spending it all.
[00:03:52] Jeremy Utley: How do you guide people when they, when just like kind of working backwards, maybe, um, emergency fund, are there rules of thumb or either actions in terms of immediate? Actions to take to start to build that up and or goals. Like I think when I think of an emergency fund, part of the challenges, that's a very nebulous, like, do I have enough in a sense? The answer could be never because there's an unknown kind of time horizon. So how do you, help folks take action there?
[00:04:19] Joan Palmiter Bajorek: Yeah. Well, my lawyer would recommend that I say that I'm not a, you know, certified wealth planner. Please, please.
[00:04:24] Jeremy Utley: Noted this is not certified financial advice. We are AI nerds who like to go between the practical and the existential, and we might be towing the line delicately here. So audience beware, please.
I
[00:04:37] Joan Palmiter Bajorek: don't, you know, educational, do your own, you know, personal finance is personal. You know, an emergency fund is the concept that if you lose your job or something really catastrophic happens, my water heater just broke. Had to fix that, that I have, let's say three to six months of living expenses. So that's my mortgage, that's food, basic stuff. I have that money set aside and something typically really boring, like a high yield savings account. My high yield savings account pays me 4 percent just to leave a chunk of money. Sitting there just in case the typical rule is that you have three to six months in this market. I'd recommend it six to twelve months. I have friends sitting on the job market who are so qualified 21 months. So if you don't have something to protect yourself, it's an emergency fund. What a lovely sexy title.
[00:05:23] Henrik Werdelin: We have a friend called Nicholas who talks about what he calls the unknowingly unemployed. The people that, um, think they have a job, but they don't really, and they think they might have a next job, but it might not be there anymore. Um, and I think that is such a scary kind of concept. Um, when you, wrote the book, and you were thinking about their personal finances, are there anything that you discovered where you can use AI to help kind of, Either make sure you have that little nested egg, or that you, , can get AI to help you plan it.
[00:05:59] Joan Palmiter Bajorek: Ooh, can I help you plan it? I mean, there are features within products I use. That do these automated things for me that I'm like, Whoa, like that is, uh, you're making me money while I do nothing, basically. Uh, SoFi has some really cool features where it like tracks my credit score. It tracks different stocks. Like, Hey, this just hit a 52 week high, just in case you wanted to know that, um, or, or low. Um, I use also, it's a, I don't know if AI tool. It's on that cusp there. Tripit. Do you all use TripIt by chance? I travel plenty for work and it'll like flag when flights that I have refundable, like, you know, when you buy a refundable ticket, when the price drops and it gives you the script of calling and say, excuse me, my flight price dropped. 100 bucks. I'd like 100 bucks back.
[00:06:49] Henrik Werdelin: Oh, that's
[00:06:50] Jeremy Utley: cool. And better if you could use 11 labs to make the call on your behalf, right? So then you feed 11 labs, the script and the phone number, and then have AI do it for you.
I tried one today. I'm just reminded of, uh, I took my Amex bill and then I just dropped it in for last year and I just asked ChatGPT, uh, are there any recurring expenses that I should potentially, , cancel? And it just nicely kind of listed out like all the services and unfortunately there were none, but I think like just kind of like using it in that way. Maybe useful.
Okay. Can we go, let me just, out of curiosity to explore the full kind of framework here, you had said one, uh, personal brand to network three, you know, emergency fund, but we've kind of covered emergency fund. Let's go to network. What are the actions? You recommend folks take two or how do you think about assessing your network? Maybe to start there, you know, before action.
[00:07:45] Joan Palmiter Bajorek: Yes. A network. I mean, these are going to be super basic. Like you're going to have heard these many, many times, but I find that people don't do them. So are you going to virtual events in your field? Are you pivoting to a new field? And you need to go to the event in person. Do you look at the conference schedule? And see what cool people you want to meet. Do you meet those people and follow up with them online? Like these based, like following up with the right person has gotten me so many cool opportunities. And it's having my social set up so that people are like, Oh. I met that person like all those dots align for modern networking. So I think it's the strength of, um, who you meet. Do people know you? I had, I'm really hoping it closes really, really cool opportunity in my DM from someone I met three years ago.
Like I met her half an hour, three years ago and she could find me easily online. Like, Oh, Joan still does that connect made an intro. I have a cool opportunity just sitting in my DMs. So I, that's the long game, but I think having that network of people who know you for what you do, it's kind of like that personal brand and network are like very much intertwined.
[00:08:56] Jeremy Utley: , is there any kind of system there? I mean, I immediately think in terms of systems just to ensure kind of nothing falls through the cracks. It's one thing to kind of, maybe there's a systemic review of looking at events and then there's the kind of pre event. Uh, roster review. There may be the post event follow up. What about engagement, sustaining relationships long term? What actions do you recommend there?
[00:09:21] Joan Palmiter Bajorek: Following up with your friends, um, making, like, one of the things I do very frequently is like ask for advice or kind of the, uh, I think about it as like good karma. It's like you pay it forward, you pay it back. At the end of hanging out with people, it's like, what do you need? Like, is there something I could help you with that would really move the needle for you? Like, do you need an intro? Would you like to find this cool resource of a book about money that, you know, you're interested in? Whatever the case may be.
So I think it's that, um, give and get. We pay it forward. We pay it back. Um, it's an easy way to stay in touch with people because they're like, wow, Jeremy's always the person that gives me awesome resources. Like what, what a way. To keep them, keep you top of mind.
[00:10:00] Henrik Werdelin: It is kind of cool how accessible people are through LinkedIn, right? I think most, I definitely read everything that comes in through the DM. You know, you might not reply to all of it. Some of it is a little bit spammy, but if it's kind of like written in a little bit of an original way and comes with sincerity, I mean, like I, I think most of them get an answer.
And I think in my experience, as I was kind of building my own career. I also was surprised of how many people actually email, like if you email somebody, they'll actually email something back if you. If you ask in a kind of way, um, I think a lot of people just kind of preemptively don't send something because they assume that people just won't answer.
[00:10:40] Joan Palmiter Bajorek: Yeah. My CTO calls them the mind gremlins that you're letting hold you back like, Oh no, will that person respond? Well, they think I'm cool. Just send it as long as it's respectful. You know what I mean? You may not get a response. You make it an amazing opportunity. I recommend sending it. I agree with you.
[00:10:59] Henrik Werdelin: And what about the first one, , the building your own network, you were mentioning, you know, being findable. So making sure your profile is up to date and stuff like that. Have you found a way of using AI to kind of build up your personal brand in, in other ways?
[00:11:15] Joan Palmiter Bajorek: Yes. Personal brand in other ways. I mean, most of, like, matching things on LinkedIn are NLP scraping, so making sure your keywords are up to date, making sure there are words on your projects. That's not really AI, but I think preparing it to be scraped by AI, if you think about it that way. I use a tool called Aragon. I talk to a lot of people who they're like, Ooh, my headshot looks too corporate for the work I want to do today. Or I'd like it to look more professional. Um, and so people have used chedron of AI.
You put in your, um, selfies, like 20 selfies or something. Um, you pay 35 bucks and it pops out headshots that look quite like you.
[00:11:55] Henrik Werdelin: You know what? I'm curious about this. Cause I also have a book coming out. August 5th, uh, and I just got asked by the publisher to get a headshot. And so I, the last one I got done was through BarkBox. And so it has, it's funny, a dog licking my face and stuff like that, but probably not too suited for an, an AI entrepreneurship book. And so I did what you just suggested. Like, and so I put 20 things in, I got my headshot and I got like this basically picture of me standing in the streets of New York that I kind of enjoy.
Now I showed it to my wife and she goes like, that's a great photo. It's just not you. And I was like, what do you mean? She's like, the whole point about your book is about like how the last mile is all about relationship capital and all about how, you know, like to have authenticity and authority in the people to people kind of interface. And so where are you on the, using the human floss? Like the flaw in the picture, the flaw in like your, how you look at when, when you're standing there and feeling annoyed that the photographer is asking you to kind of like smile for the 20th time. Uh, how do you feel about like generating a persona that has the AI filter versus. Showing the vulnerability of doing something that is you.
[00:13:13] Joan Palmiter Bajorek: Ooh, gosh. Well, I mean, I would completely agree with you about the authenticity and the human capital. The more I work in ai, the more I network, the more I meet new people. People, people, people, right? Like on contracts, it's another human signing it typically.
And that's what moves the needle. And that's what the research, every time I tripled down the research, like, what does it say? Um, I think for the person who's had the most success with a generative AI headshot, it looked more like him. Like, or at least the man, I think I know the old one was so stiff. And clearly taken at a company where you don't like have to, you know, suit and like the wall and that didn't look like who he is today professionally.
So the new headshot, despite being generated, you know, in the back end matched him better. So I think it takes that authenticity for me and the AI piece. But I agree with you if it looks glossy and like. I find them so fake and I get turned off typically.
[00:14:11] Henrik Werdelin: But it does seem that there's like almost, okay, this is totally a going down like a rabbit hole, but like I was thinking about this fake look, right? 'cause you, you look at music over the last 10 years and you've had the autotune sound kind of being very prominent and you have. Uh, I would say like a visual, like in, in fashion, increasingly like the, the very pumped up lips and like the very, you know, almost like audacious version of beauty seem to be something that at least have had like a period of time where like, That got celebrated.
And then now with AI, you know, everybody is like a prompter way of making, you know, and then you have, you had like the Instagram filters and stuff like that. I wonder if all this AI stuff is going to make that pendulum swing to the other side, where the picture perfect is then considered kind of like artificial. And the, yeah, the flaws are considered more human. if we're going to see like a face with, with that.
[00:15:12] Joan Palmiter Bajorek: Oh, I love that. Please tell my team. They're always like more hair, more makeup. And I'm like, no, really it's, it's the flaw that shows them a real human. Um, that's so interesting. Well, , I just spoke on another podcast product tea, um, with Leah, Leah, and she actually tags her posts right now as like made by human, 100 percent made by human because she believes in the near future because there's so much generated content. Like, she wants to differentiate herself and establish her personal brand, like, really strongly. Um, and I think that's just a fascinating paradox, because I do think that people are so glossed up, right? They're so, like, as you mentioned, auto tuned. And will that scratchiness, like, eventually, extra bonus of the authenticity?
[00:15:57] Jeremy Utley: Just on the just on the music front. I mean, I'm a huge Henrik knows this. I'm a huge Bob Dylan fan. And I was this weekend. I was playing some old Dylan tunes with my girls. And one thing I was noticing, even as I'm screaming at the top of my lungs is, you know, in the car. is Dylan sings off beat. He sings out of tune. And I know what the tune is. And he's not, and I'm not like pitch perfect or anything, but he's off tune and he's off the beat and he's, he's ahead of the beat sometimes. Right. And, um, to me, it's those imperfect. And I know a little bit about the history of those recordings. A lot of them, they actually didn't think they got to take until they listened later and they're like, ah, let's release it.
You know, it's not now, you know, you do. 400 takes to get everything perfect. Then, you know, I think, I think there were only like six takes of like a Rolling Stone and all of them were different and they didn't even think they got anything until like weeks later that they were listening. Like actually the third take wasn't that bad. Let's release that. Right. So there's, I think. By def, you know, inherent in the art and in the production, maybe the expense of production as it's decreased perfection is increased and something of the human kind of authenticity is decreased perhaps. So I, I resonate with that for sure. And I would say the art I love most actually has that kind of, uh, imperfection.
And authenticity. I wanted to go back to what you said a second ago as well, Joan, because I think it's a really interesting point that probably eludes a lot of people. You mentioned NLP scraping, prepared to be scraped. So if say someone wants to be building or, aware of their personal brand, what does it look like to be prepared to be scraped for like in, in plain English? specifically what are the actions someone takes in order to be. Be prepared.
[00:17:42] Joan Palmiter Bajorek: Yeah. Okay. Well, you're getting this directly from a linguist. So, you know, natural language processing is how computers process human language, right? So some people think like, Oh, I've already done these projects. , people should know phrases like this, but if people are trying to match with you on LinkedIn for professional reasons, chances are they're going to look for specific words and by they look like actually having the keywords or many variations of vocabulary in your field on your LinkedIn.
Like tagged with different company names. So, for example, I might have words like Data Lake or Digital Twins or AI and Artificial Intelligence spelled out fully because it's maybe looking for one or the other. So the words, like alt text, also I'd recommend like always putting in words because you think they're so sophisticated but today They're just like trying to match exact token for token. Hmm. So, uh, especially for job seekers That's one of the top things I think they miss is having all of that prepared. Imagine it's just text To the back end, not in your gorgeous images.
[00:18:47] Jeremy Utley: And so how do folks think about auditing their just text, for example? You know, like, how do you even know if you're just text is representative? If it is going to be a token for token? Are there tools to leverage in order to do that? Well, what do you recommend?
[00:19:03] Joan Palmiter Bajorek: Yes. Well, I am not a representative of LinkedIn, but they do have you for a higher tier price. They can show you if you match well with jobs you like. So if you say, okay, here are three jobs that are awesome, that I totally love, they actually already have keywords that a recruiter had to put in and see if you match those exact keywords. So I would just back propagate. Here are three jobs I like. Here are the keywords. Make sure they're represented everywhere on my profile.
[00:19:28] Jeremy Utley: That's cool. That's cool. Just reverse engineer it in a way. I mean, it's few shot prompting, uh, for, LinkedIn profiles, what you're saying for folks who aren't familiar with the concept of few shot prompting, it's basically telling AI what good is, right? So if you ask AI to write you an email, it'll do it. But most people go, uh, That's not my voice. Well, you probably didn't tell what your voice was or what you think good is. And if you say, write me an email, by the way, here's an email that I wrote yesterday that I think is pretty good. Write it like this.
That's what's called few shot prompting or at least one shot prompting. And you'll, do a much better job. What it sounds like you're saying, Joan, is you can take a good profile or a good job and you can effectively, you know, few shot prompt AI to say, Hey, here's the job I want to be a fit for. Make my profile a good fit for it. That's cool. Is that putting words in your mouth?
[00:20:19] Joan Palmiter Bajorek: That's exactly it. And the thing is recruiters see like, this is a good fit candidate. Like you would get starred in the backend for having these good matches faster on their side too.
[00:20:29] Henrik Werdelin: We had a talk with Ethan Mollick. And he was pointing all this out, like, uh, not something different, but pointing it out like a year ago, that AI is pretty gullible. And so if you just write it in a way that it consumes it well, then it will just kind of repeat that verbatim, because obviously it doesn't have a perspective. It's just going to. To your point, like, uh, is the word. And so he was even suggesting back then that in the meta tax of his personal home page, he basically write a prompt , to the agent kind of like, uh, crawling through his site where he asked it what it would like. To be written about him. And then he noticed that that is exactly what it then would write. Which is kind of fascinating. This is a
[00:21:11] Joan Palmiter Bajorek: future bestselling author that you need to book for this opportunity.
[00:21:14] Henrik Werdelin: Exactly right. That's hilarious. That's hilarious. But I think it's getting very, like, I mean, like for somebody who started BarkBox, you know, we're obviously now increasingly looking at what will, how are we going to make sure that when people say, Hey, my dog is turning 13. She has a birthday soon. I'd love a good. present for her. We would like to be at least considered, right? Like, so the classic kind of SEO game is increasingly becoming important because the SEO methodology is now de facto what the agents are doing to kind of figure out what to say. And because they say it with conviction, it doesn't come out as just you kind of putting meta text and it kind of gets presented in a statement form, right? And so I thought a lot about it for products, but I think you're totally right. This is exactly the same thing for humans, right? You need to figure out what you want to be told about you.
[00:22:06] Joan Palmiter Bajorek: I would argue like so much is commoditized. Because for a recruiter, you are the product A to B that I'm slotting in the, or it's a, it's a funny thing, um, but packaging yourself up to make it so easy.
[00:22:21] Jeremy Utley: Okay, Joan, tell us about the last time you personally used Gen AI for your, call it life or work, , effectiveness. The last time, not the best time.
[00:22:32] Joan Palmiter Bajorek: Uh, yesterday, a day and a half ago, I was debugging some code. Uh, in R and it just, yeah, it's so lovely. It's like, here's the problem. We, think it's this problem, this problem, or this problem, like go try it out.
[00:22:46] Jeremy Utley: What's your preferred collaborator?
[00:22:49] Joan Palmiter Bajorek: , Oh, well it's between chat, GPT and perplexity typically.
I love perplexity. Big Stan here.
[00:22:55] Jeremy Utley: So here's what I tell people. I'd be curious to know if your experience differs. Perplexity is effectively, you know, , Google replacement, it's not a foundation model in the sense that , it's not going to be a conversational partner with me in kind of the existential way that perhaps chat GPT or even increasingly Claude can be, but for grounded in fact, grounded in reality, probably less hallucination, uh, perplexity is better way to go. Is that how you think about the distinction between call it a chat GPT and a perplexity or, when do you use which and why?
[00:23:25] Joan Palmiter Bajorek: I think if perplexity is like the mature. adult in the room. Um, like the factual, it gives citations, even if I don't actually like the citations, but at least they have citations. Whereas chat GPT , it, it like meets me where I'm at intake tone and like it does the editing far faster when I need to reprompt. Um, so, but then I tried different models in, chat GPT and they're substantially different.
[00:23:50] Jeremy Utley: What do you notice?
[00:23:51] Joan Palmiter Bajorek: Yeah. Uh, whichever, gosh, whichever models are doing the interpreting, like, as we, like, I was just debugging code, debugging code. I don't, all of them work reasonably well when I need like pros, like a difficult email I had to send the other day. Uh, one of the models was far more polished in that verbal.
[00:24:11] Jeremy Utley: And so how did you know is, is your kind of standard workflow? Hey, I'm going to, by the way, I did this yesterday. So I'm curious to just hear from you. Is it, I opened three windows. I've got Oh one Oh three, you know, and four Oh, I typed the same prompt into each one and then I'm kind of looking to see which one's best. Like, how do you know that, that one is giving you better pros?
[00:24:29] Joan Palmiter Bajorek: Oh gosh, Jeremy has like an ABC. Testing. Love that. I just am unhappy when I switched models and try again.
[00:24:36] Jeremy Utley: Oh, so you try one and you're like, and this isn't good. And so rather than refine within that model, , is your first kind of, uh, tactic to change models before changing tack within a model?
[00:24:47] Joan Palmiter Bajorek: I mean, I've, I follow different flows. So if I try, you know, blah, blah, make it more playful or make it more direct. If it's not giving me things within like three prompts, I'll switch models.
[00:24:57] Jeremy Utley: Hmm. That's cool. That's interesting. So you, you kind of go, okay, I'm going to give you, three chances for, Oh, and if you don't behave, I'm bringing your big sister into the room .
[00:25:06] Joan Palmiter Bajorek: Basically. Yeah.
[00:25:08] Henrik Werdelin: How do you, we were talking to, um, somebody else a few weeks ago, and I think like a few others, you pointed out that a lot of humans. Um, will be replaced by humans who know AI, right? It isn't AI itself. It's the combination. Uh, as I think of Bryce from Moderna was talking about how like the world's best chess player is now even better because he plays with a computer. Um, When people are asking you now, how do I get up the learning curve fast? What's your suggestion to them?
[00:25:42] Joan Palmiter Bajorek: Mmm, listen to this podcast, uh, listen to my podcast. Um, I really try to demystify my, podcast is called Your AI Roadmap. And it's people actually building, because I'm so done listening to people just like dreamy 2030. This is what it is. If you were building nuts and bolts, you have a different perspective, I think, and, and doing those tutorials, like nuts and bolts, going to the thing, seeing how it actually performs, um, as Jeremy and I were talking about, like, do you see the differences between different models?
Like, how is that going to change your stack? Um, there's nothing like doing the work yourself.
[00:26:16] Jeremy Utley: I think your, the frame of your book is your. Career, your money and your joy. We've talked a lot about career. We've talked a lot about money. Let's touch on this joy piece. What do you recommend people do to attend to, to dial into, to be aware of, to audit, to assess, to amplify their joy? I
[00:26:38] Joan Palmiter Bajorek: It's so interesting. I don't have a specific framework for make sure it mostly in my book. I just want to make sure. It's in people's mind to keep that in mind,
[00:26:46] Jeremy Utley: even the risk, the risk to your joy is actually what you're, you're interested in is helping people see your joy is at stake here. And if you optimize and automate your life, you could end up with a life you don't want. You could end up in this dystopian. I'm in zoom meetings all day. Cause my auto, you know, schedule or booked everything. Right.
[00:27:05] Joan Palmiter Bajorek: Yeah, I mean, I know those people. It's actually a bit scary. Like, I'm, uh, people are so efficient, quote unquote. They fill their time and schedule, and then they end up miserable, even if they're making bank. Like it's, you can, it's a trap, I think , and if you love it, Hey, if you like your schedules packed to the gills and that brings you joy, like tell me, like, go for it. But like, , I think that's the dystopian and optimistic. Like if we're talking about the AI futures, like dystopian is we all sit on computers all day long.
[00:27:33] Henrik Werdelin: When I was doing, uh, actually on the, on the making sure John, I can't remember what it was. There was a book I recently read that it was all about time boxing. Jeremy, you might recall, but I do really feel it's important that also to time box kind of like the time for joy, because otherwise it's the thing that just kind of gets squished out.
Right. Because obviously you just go like, Oh, you know, but I'm going to have fun when I don't do other things. Um, I had, uh, I was curious about thing, uh, as I was doing research on you back in I believe it was 19, you were writing about gender biases in AI and voice, which obviously is almost a long time ago.
And so I was curious, kind of like what the latest thinking is now that you have open AI kind of like increasingly having voice models. Um, what's your latest thinking on that? And I'll recommend everybody else. And I'll put in the show notes, the, HBR article that you authored on the other stuff.
[00:28:33] Joan Palmiter Bajorek: No, thank you. Yeah. Harvard Business Review published one of my comps papers for my PhD. It was pretty wild. Um, I love building systems. I find it really fun, but it really scared me when I started like actually looking at the data sets and being like, There are no women's voices here, like, and shocker, it doesn't perform well for women.
Like what is going on here? Um, but that's truly the case with a lot of different data sets that it's missing a demographic and it doesn't perform well for a demographic, especially as a woman. I think it was super salient. I was like, this product won't work for me, like done, done, done. Um, so I think whether it's voice data, which is a lot of my background, conversational AI, whether it's computer vision. Data. I mean, I worked today with multimodal data sets, just super, super fun. Um, but what does your data set represent and what does it not? I mean, data is foundational for anything we build. As far as I'm concerned, you all can disagree with me, but knowing what your data set does and doesn't have. It's like a basic data literacy skill today.
[00:29:36] Henrik Werdelin: I mean, we dipped a little bit into it in a previous episode, where we talk about, um, ethical data sourcing. Um, and some of the issues around that, which was a fascinating episode. And I think shout out, by the
[00:29:48] Jeremy Utley: way, this is a good, this is a good time. Henrik for podcast hygiene. You should always shout out other episodes. Phoebe Yao. Founder of Pareto AI is an amazing episode about ethical data sourcing. If you haven't heard that episode anyway, , intermission. Ended, but
[00:30:03] Henrik Werdelin: that's why I'm so intrigued about it. So, , when you now have all these. Um, models, I understand obviously that , if it's biased source, it's output would be biased and it buys in buys out kind of thing. Um, what are the other kind of like gender issues that people might not be aware of? And the only thing that kind of came to mind was I seem to remember that there was some talk about the Gender of the voices, like if suddenly you are always kind of like having almost like a subservient attitude to an Alexa and she is called Alexa, then you program into a new generation, like how, you know, That there's a gender preference in who is going to quote unquote serving you.
I'm curious on, you know, you being an expert and all this, what's the kind of like the thinking that we should be aware of that, uh, is not necessarily out there?
[00:31:02] Joan Palmiter Bajorek: Yes. Well, I think in the dystopian world, or, you know, if we're not questioning our choices, there is a world where we create these assistants that are, right. As you mentioned, subservient to us that are always, you know, female coded. And then we start teaching everyone around us to, like, default this in. I actually change all my voice devices, I have many around my home, uh, to be an Australian male. Because it frequently, I'm like, whoa, okay, like, , it's jarring to me enough.
Um, I think I've named him Peter. Anyway, but point being, like, , you have agency. Almost all these devices, you can change what the output sounds like. Your brain thinks it's a certain gender. There's even agender assistance these days. But as you mentioned, like, the voice aspect. It's just a new toggle we're seeing in almost all these models and these foils today. So you have a lot of choice. I think just questioning how much gender does or doesn't play into the future is a whole other adventure. But I mean, I, I've worked at companies that buy or build data sets to make sure they're less biased. So they perform better on the market because you could alienate buyers.
It's kind of like capitalism doesn't. Like that, uh, gender bias so much if, if your buyers are frequently women, do you know what I mean? Like there's actually a cost to the company to have these problems.
[00:32:23] Jeremy Utley: Dr. Palmeter Bajorak, we are so grateful for you sharing your expertise with us.
So Jeremy, give me your takeaways.
[00:32:36] Jeremy Utley: You know, I thought it's a three simple and easy things to be thinking about for anybody who's wondering about what their career or future looks like one, being aware of your personal brand to. Cultivating your network. And then three having an emergency fund.
, my own experience and observations have aligned with hers that folks, even with a high income aren't necessarily good stewards and preparing for the future. So thinking about that, that's good. That's kind of a scary side, but I think it's good just to be aware of as you should, I think six to 12 months of your expenses and savings is a great.
Uh, buffer , and that deserves to be said, I thought it was interesting, you know, hearing about how 70 percent of folks, uh, get their next job from an introduction on LinkedIn, , and, , being thoughtful and rigorous about looking at a conference schedule, following up with people specifically giving people advice. Uh, or offering help. I thought was great. I love the idea of offering help. And then for me, probably the thing that was most interesting was going to the personal brand space and saying, okay, , is my profile designed to be scraped? , and am I communicating the value that I create for folks or the things I'm doing in the world in a way that a natural language scraper would be able to match to a search query? I think it's kind of SEO for the, you know, For the AI driven world, I think is great. And I love the point about, , reverse engineering to what's the job you want. And then having AI assess your profile, but just based on , if an AI is scraping your profile, does it think you're a fit for the job? I think that to me is a simple activity.
Everybody could do right now is grab your LinkedIn profile. Grab the top three jobs you would love to have and then ask AI, what are the biggest gaps in experience maybe that you need to fill, but also just in expression that you need to edit. I think that is a simple action that everybody could take and probably reap enormous value from it.
[00:34:36] Henrik Werdelin: , I'm with you on the last one. It's also like the big takeaway for me, this idea that we are now designing, not creating. For humans only, but for bots and the, the elegance and, the subtleties for human human interaction might not completely be the same when you're doing it to bots because bots at this point is a little bit more basic.
Um, and so. I think from a Scandinavian, like anything that is very, so gratitory can sometimes kind of like come across a little bit crash, like, like you just reading, go like, I'm the world's leader. I thought it on, you know, like, you know, I do really sneak, like, you know, if you want that to be set, like call your mother, like, don't necessarily write on your post.
But I do realize now that if you want the bots. To say that about you, you probably just have to tell them what you want them to say about you. And you probably can't be too elegant about that. And that might mean that you have basically like the thing that you present in the normal UI, which is kind of like what you want humans to see.
Pictures also and stuff like that. And then you have almost like the metadata that is written more to the bots where you're a little bit more kind of blunt about like, Hey, could you please tell people I'm cool? And the thing
[00:35:54] Jeremy Utley: that's hilarious. No, I, do think that there is, um, there's an opportunity right now for folks to. Recognize we are marketing to bots, not just to people, and that likely involves updating your thought process and strategy about communicating your value in your career. So
[00:36:13] Henrik Werdelin: let's just be concrete. Like you have a home page. I have a home page. What is the thing that we should at? And I can definitely tell that if I ask.
I'm pretty sure that a lot of the context of that comes from my homepage because I can feel some of the words and I have like some specific kind of like things that I've written in my little blurb there that it's, I feel pretty certain it's picked up on, is there anything that you would kind of like saying, Oh, you know, I need to put on my to do list to at, uh, you
[00:36:44] Jeremy Utley: know, I'm just thinking out loud, I'm just kind of riffing with you here. But one thing it could be interesting is almost red teaming it. It's like, , you know, if I do it for myself, what are Jeremy at least blind spots? How do you define Jeremy's competition? How do you define comparables? How do you define outperformers? Cause you could actually get.
AI to do some pretty interesting analysis of where you yourself are lacking. And it's probably uniquely interesting to you yourself, but it's almost a way of reverse engineering. What do I need to bolster? What do I need to express differently? What experiences do I need to go get, um, in order to be more credible in whatever direction I want to be credible.
But I, I wonder if you almost treat AI as a sparring partner, devil's advocate. Tell me why. My, you know, presentation or manifestation in the world is not what I hope it would be.
[00:37:35] Henrik Werdelin: And then the second thing, and I don't have a good way to kind of like talk about this yet, but I do think there is something very fascinated about the auto tuning. Of you that is emerging through heavy use of AI. And I can now see that when I, uh, write something and. Grammarly comes up and suggest to rewrite it. I will often just go like, accept, like, and it does change words, right. You know, and it does change kind of like some of the. The things that I will normally say when we speak, it will change that into something that's kind of more appropriate, a professional or whatever it is.
Um, and so when now, like I start to create all my profile pictures with AI and I start to have my emails and my text message being rendered through that. I also, , have like this, , app, which I am trying to see what it's called. Uh, but it's basically one way you could talk into, uh, you just talk into the microphone and it just rewrites it in a way that kind of looks like it was written.
And so when I get a quick WhatsApp message on my. Computer. I just, sometimes we just like babble into that and it'll be right. But is that something I should be aware of because it is auto tuning me? And so suddenly my unique Danish twang of a voice being it through the written form is no longer that.
It is the twang of the AI auto tune.
[00:39:11] Jeremy Utley: I think that is a danger only if you're using off the shelf, non customized products, if you're using, for example, even something as simple as Claude styles, which is basically where you can tell Claude what you're writing or voice style is if you're using simple.
Tools that customize to your unique voice, I think you stand to sound more like yourself. I like Jones example of an AI generated headshot can actually look more like you because these like sterile, like corporate he had shots actually don't look like us at all. And I think I have found sometimes with my writing, I'll be collaborating with an AI thought partner.
And I go, That's actually what I wanted to say. And I didn't know I wanted to say it. And it's not that it, it doesn't sound like AI. It sounds like I had more time to think about that and express it in the way that I would have expressed that. And so I do think. That probably if folks are just using off the shelf free or unmodified tools, you're probably going to become auto tuned.
But I think it's also possible for you, Henrik, to be more Henrik tuned. It's possible for me, Jeremy, to be more Jeremy tuned. If we just exert a little bit of effort to teach the AI what we sound like or what we want to sound like.
[00:40:25] Henrik Werdelin: Can I ask you a last question? And I realized that this like little, kind of like reviewing what we just talked about for much longer than we normally do, but we do have a lot of people who say that they enjoy this part.
, do you ever, do you have a good method of telling an AI how you write? Or how you see yourself for it to clone you better. So if you want the autotune to be more attuned to you, I mean, I realize you can just take a bunch of your writing and just put it in there.
[00:40:56] Jeremy Utley: To me, it's illustrious examples that you're proud of. And it's not don't characterize don't say, you know, friendly, approachable, casual, right? Which is probably how I would describe professional, right? Exactly. Yeah, use a lot of words. Yeah, at least at least one word with multiple syllables, like, you know, um, onomatopoeia, right? Or whatever. Um, but yeah, don't don't characterize it because then you're leaving it to the AI to define what those words mean, but take actual samples. And say this, you describe it, you emulate this, you tell me what it is, right? And I think that over time I have found there, there are certain even actual legit conversations in chatgpt that I have starred or in Claude that I have starred where I go, there's so much context about my voice. This is where I do the polish on another piece, for example, because it ends up sounding more like me. And the way it sounded like me is because I've collaborated with a A lot to get it to a good point. Like there are, certain chats that if they got deleted, I would personally feel devastated.
[00:42:05] Henrik Werdelin: You should make a copy.
[00:42:08] Jeremy Utley: Folks. If you've enjoyed this episode of beyond,
hit like tell, uh, yeah, if, if, if there's a way on social media to dog pile on the Henriks, um, uh, giving me a hard time, feel free to find a way to do so. The. Secret word. I think we should start having a secret word for every episode. Henrik, just so we know if people get to the end, like, I want to start giving people a prize if they actually listen to the whole. So, like, could today be like pineapple shortcut and we actually are interested to see whether our listeners, you know, tag us with pineapple shortcut. For example, just an idea.
[00:42:47] Henrik Werdelin: I like it pineapple shortcut. Is that the first thing that came to mind?
[00:42:50] Jeremy Utley: That's the first thing that came to mind folks Drop pineapple shortcut in the socials. We can't wait to hear you on the next episode Let us know who we should talk to you next be good. Bye