Subscribe
Share
Share
Embed
FutureProof You is a team comprised of: Career Pivot Expert Dan Yu, Recruiting Master John Lovig, and AI Consultant Aaron Makelky. Listeners will hear discussions of job trends, career advice, and actionable tips for making sure their careers are future proof!
EP 12 AI Rate of Change
===
Aaron: [00:00:00] Welcome to another episode of the Futureproof You podcast. I am one of your co-hosts, Aaron Makelky. I'm joined on this episode by Dan Yu and John Wig, the topic for today's episode, AI Models and the rate of change and the pace that things shifting in the AI world. What's this mean for job hunters and people looking to career pivot?
Aaron: How do you keep up with all these tools? How are we using 'em in our daily workflows? All of that and more in this episode. Thanks for joining. Let's get to it.
Dan Yu: just in the past few weeks, what we've seen, just the three of us have seen when we test out different engines, right? AI engines like Claude or the new Gemini, right? seen the acceleration of how good they've become just in the past few weeks.
John Lovig: Yeah.
Dan Yu: does that mean? What are the implications?
Dan Yu: I don't know, but what are the [00:01:00] implications going forward? Will it get even better? I, I mean, instead of by weeks, maybe by day, I mean how, like, how,
John Lovig: Um,
Dan Yu: it? Is it learning?
John Lovig: mean, Dan, we, we, you and I are directly pla we've directly placed people who are training these AI models, you know, like.
Dan Yu: That is true.
John Lovig: So I, I think given the, knowing the behind the scenes as to how that gets done, start to see pretty quickly that there are, there is a very enormous concerted effort to continue to make this more natural language.
Dan Yu: Yeah. I mean, just, just using our voices. It, it's, it's starting to be much more realistic, you know,
John Lovig: I,
Dan Yu: like, like we can't even tell right.
John Lovig: mean, you, you talk about, you know, Aaron's, Aaron's, you know, uh, quote that I love is the, you know, that we've got a eager intern, right?
Dan Yu: Right.
John Lovig: for my, for, [00:02:00] for the new model of chat, GPT, that eager intern, I barely even have to double check their work anymore. I'm scanning their work now versus like having to make serious edits. It's crazy.
Dan Yu: So what are the implications for work then, right? If that overeager intern is so good that you just have to check their work, you know, barely check the work, what happens to I. admin, what happens to the virtual assistant? What happens to the paralegal? What happens to all of those people that are tasked with in their everyday job to handle, just handle things?
John Lovig: They need to re-skill.
Aaron: Yeah, they need to learn to use those tools or they will be replaced. And it's cliche, but the person using the tool will replace them. So it's not ai, it's the AI powered legal [00:03:00] assistant is going to do 10 x the work and 10 x the quality in one 10th the time of the old school Microsoft Word only legal assistant.
Aaron: And I think John actually answered one of the other questions, Dan, which is, well, what's the moat that keeps people in with the large language model? There's. Two ways to look at this, and I live in both worlds. One is John's moat is all of his training data is in chat. GPT. When somebody goes, oh, Aaron posted on LinkedIn that Gemini 2.5 advanced is really awesome, John rolls his eyes and goes, yeah, but chat GPT has all these documents and all this history and all this context on me.
Aaron: Am I really gonna move to a new model that edged above for, let's be real, maybe a few days to get. Marginal returns on that. The moat for him is his data and the context that he has built in chat GPT. He would be stupid to, to jump the.
John Lovig: about it just with [00:04:00] training perplex. Complexity. But at the same point, the thing with perplexity that's been awesome is because I do all of these things in Google Docs already. I connected my drive. If I throw them in one folder, I use that for context. Now it has every time I put something in there, 'cause it's connected to that folder.
Dan Yu: right. So,
Aaron: As a,
Dan Yu: and returns.
Aaron: yeah, and as a team. That's the next level. We're currently, I don't know, 48 hours into this. I can connect my docs to chat GPT alone, but when you're working in a team. Say us three John's Docs and Dan's and mine are all connected, and that's a whole new le level of collaboration and context. So instead of, oh, I have to take Dan, can you send me that?
Aaron: I'll download it, re-upload it into my tool. Dan logs in and says, sync my drive with the whole workspace enterprise, not just with Dan's account, the, the other way that people. Like me who are weird, see, [00:05:00] the moat is I have to live enough in GPT and enough in perplexity and enough here, there, there, whatever, that I can pivot around to what has the feature and the model that I want for the task.
John Lovig: Mm-hmm.
Aaron: When you get out of your LinkedIn bubble and get off of your Slack channel of nerdy AI people, 99% of people don't care that Claude added this little feature last week and then GPT added image gen. It's like I'm not gonna move all of my stuff because I miss out on the context. And all of the data and history that I've put into that platform is actually what keeps me behind the moat, not the model's quality.
John Lovig: Yeah. Well look at where it's going to be more interesting what you, we've talked about this in the past too, right? So. Right now with all the embedded tools within Google, right? I, I used it yesterday. I used Gemini while I created a target company list for a search that I'm working on. [00:06:00] And I said, Hey, Gemini, gimme 20 more companies like these companies.
John Lovig: And then it just went boom, right? My little list there, I didn't have to leave or anything, now you look, apple is investing billions of dollars to create, um. data center for their AI models and their AI is already pretty good, but imagine now it's embedded in your hardware. Now it has all the context. That's the race to win is, is to be the device that that can do it all, because now it's embedded with everything that you do. It has your behavior as a person. It understands the websites you go to. It understands everything that you do. Um, scary in a way, but also really exciting for someone like me.
Dan Yu: Well, right, so then you, this is the, the whole 1984 Big brothers watching, you know, who do you trust then? Right? So if it's embedded in the hardware, then it's already two [00:07:00] levels up, right? So then who do you trust and. Do we trust the, the team that we've just happened to be, know, in the moat with,
Aaron: And he.
John Lovig: I mean, I don't trust any corporations, to be quite honest. I trust myself. you know, at the end of the day, the ones that speak the most about data privacy are the ones that. I'll trust more. And so far Apple uses supposedly all the computing on hardware. It's just doing calls to the data center. It's tokenized.
John Lovig: So it's not identifiable supposedly. Does that mean it's gonna be the case? Who knows? It's just like Alexa hears you and knows what you're saying and they can pull that up at Amazon. Maybe they can do the same thing at Apple.
Aaron: Yep. And one of the things people don't realize about the free versus paid models, they go, the use limits go up or you get access to a better model under the hood. For me, that's valuable, but a thing that a lot of people don't appreciate, you're buying with that $20 a month [00:08:00] subscription is control over your data.
Aaron: You, you can, let's say in perplexity. Most of the free versions of tools don't let you disable training the model. That's a default opt in chat. GPT is a good example. ROC does that as well. You actually have to pay to be able to not train the model on your data. And a cool thing for enterprise and small business teams, let's say with Perplexity Enterprise Pro.
Aaron: You can say, even though all these are my documents are synced, I don't wanna train the model. 'cause by default it won't. Like you can't enable that. So you're buying that privacy with your subscription. But the other thing it'll do is it'll tell you what happens to your docs if you cancel. And there's a lot of nerdy people who actually read the terms of service.
Aaron: Google retains your documents for three years and says humans can and will review your documents, not just AI models. Perplexity for [00:09:00] 30 bucks a month for enterprise accounts, your documents are gone. Seven days after you cancel, you're not training the model and humans won't look at them. So some people, let's say in legal or FinTech or something proprietary, it's not even about the model's power.
Aaron: It's about the privacy that I get when I buy that subscription
Thank you for joining us on this episode of The Future Proof You Podcast. We have a lot going on on our website. If you haven't checked it out, it's future proof Y u.com. We've always got resources for you to help take control of your career future, whether that's live classes or recordings of past ones.
If you're trying to find a job or make a career pivot into a new industry, got laid off. That's what we're here to do is help you future proof your career. Thanks for joining us on this episode. We look forward to seeing you on the next one.
Dan Yu: .
Dan Yu: So,
John Lovig: Yeah.[00:10:00]
Dan Yu: uh, I think the, uh, majority of people out there, especially people that might be tuning into our show for the first time, uh, they probably don't know that I. And, um, I think it'll be helpful for them if we have this in the show notes that we can just list out some of the things we've, we've said and, uh, we can pull the transcript for, for everybody as well.
Dan Yu: But, uh, I think this is invaluable information for whoever is watching, uh, what we're talking about here.
Thank you for joining us on this episode of The Future Proof You Podcast, where we talk about ai, the rate of change, and what that means for job seekers and career pivoters like you. If you haven't checked out our website, future proof y u.com, you can find upcoming live classes in May of 2025. We've got career pivot [00:11:00] classes for late stage career, pivoters, like a second career.
I'm retired, but I'm not done. I wanna do something different. Finding your first job out of college or getting your degree, and for teachers trying to make career pivots. So make, make sure you check out our website, future proof-ou.dot com and follow us on LinkedIn. Just search for future proof space YU and follow our page.
Thanks for joining us. I'm your host, Aaron Makelky, and we're excited to help you take control of your career today.