Ep 468: AI 101: How to Use AI to Save Time, Make More Money, and Simplify Your Life
35:14 Share

I like to think that if you can put your little imagination cap on, talking with these tools is really very similar to talking with a person. Imagine you have an expert assistant that can do pretty much everything for you as long as you tell them what you need them to do clearly.

Hey, everyone. Thanks so much for joining us today on Her Money. I'm Jean Chatzky. Very glad to have you along because today we're going to talk about AI. And at least for this episode, bear with me, we are going to set aside those job security concerns. We're also going to set aside the fact that AI just can't seem to get people's fingers right. And instead, we're going to focus on AI.

Well, some of the positive, actually. We're going to dive into the ways AI tools can actually make our lives better. Because the truth is, those robots, they are here to stay, and they're only going to improve over time. So I

I say it's time that we try to make the best of it. And that's why I wanted to bring an AI expert onto the show. And I could think of no one better than Celia Quillian. Celia is an expert in artificial intelligence. She is creator of the popular TikTok channel Smart Work AI and author of the book,

AI for Life, 100 plus ways to use artificial intelligence to make your life easier, more productive and more fun. We are here for that. We're going to take a very quick break.

You know that moment when you're standing in the kitchen staring at your glass of tap water and wondering what's actually in this? Yeah, same. I've always been big on staying hydrated, especially when I'm training for a race, but after learning that nearly every home in America has harmful contaminants in its water, I knew I had to do better. That's why I love AquaTrue.

Y'all know that I love to have coffee brewed all day. I'll admit it. And I can honestly say it tastes so much better now. Plus, with filters that last up to two years, I am not constantly worried about replacements. AquaTrue comes with a 30-day money-back guarantee and even makes a great gift.

Today, my listeners receive 20% off any AquaTrue purifier. Just go to AquaTrue.com, that's A-Q-U-A-T-R-U.com and enter code HERMONEY at checkout. That's 20% off any AquaTrue water purifier when you go to AquaTrue.com and use promo code HERMONEY, H-E-R-M-O-N-E-Y.

And we are back. We're talking with Celia Quillian. Celia, welcome. Thank you. I'm very excited to be here. When I think about generative AI and the speed at which things are changing, I'm just overwhelmed. I got to admit, I am overwhelmed. I think a lot of people feel that way. So how do we instead look at it as you do, which is as a positive?

Yeah, I think one of the things to start with is approaching it with curiosity and not fear. Something that I like to remind people is that AI has actually been around for about 75 years as a field. And it's been helping us in our daily lives for a while, even before these chat GPTs and other generative AI tools have kind of emerged into the pop culture existence the past couple of years. A great example are your Netflix recommendations. If you see a show that's recommended for you, that's artificial intelligence working in the background.

Now, when it comes to generative AI, when you approach it with curiosity, you can find that there are so many ways that it can outsource the tedium of some of your daily life tasks and some of your work tasks so you can focus on the things that you love to do or the things that you're especially good at. And they can be great collaborators in that regard as well.

I know that from that perspective, you talk a lot about how AI can make you more efficient at your job. It can help you stand out in a job interview. It can help you create a budgeting plan that is tailored to you or plan a vacation with a lot less effort and work. How do you know exactly where and when to plug in?

Yeah, so for me, the more that I use these tools, the more I find what they're good at. But the thing to keep in mind is that the way that these work is these are trained off of a massive amount of information. Like imagine every book in a library, or every piece of information on the internet. And because they've learned to recognize patterns across all of this information, they can generate responses that are excellent for you.

So generally speaking, when you approach using these tools for any tasks that you may not be an expert in yourself, they're going to be very good at helping you with anything that can be generated back in a text format. So when it comes to planning, they're excellent.

They are also great creativity boosting tools as well. They are intentionally creative, which is why sometimes they can create things, and I think people have pointed this out in the past, that may not be 100% totally accurate, but they can be a great jumping off point. So when it comes to travel planning, that's something that takes a lot of time, can be stressful. I always say, start with an AI and see how far it gets you, recognizing that you should always make sure that everything that it gives you back is accurate.

the thing that you need it to be. And you can always collaborate with it further from there by continuing to have a conversation with it. So it can be excellent at planning, outsourcing tasks like travel planning, meal prep, meal planning, because of all that information that exists within its training data.

I'm realizing we are getting a little bit into the weeds without giving people some basic definitions. And for those people who really haven't gotten started using ChatGPT or other generative AI tools in their daily lives, let's talk about it and let's sort of lay out why.

what it is. When we're talking about generative AI and large language models, what exactly is that? Yeah. Generative AI is high level, a tool that is designed to generate new content for you based off the inputs and requests you put into it using natural language. So you can effectively, with these tools, speak many things into existence.

And how it's able to do that is because of the large language model. These are basically, again, a huge massive amount of information sped into a computer at some point in time. That computer

was able to notice patterns by running through, I'm getting very technical here for a second, a capability called transformers. This new introduced technology is able to allow the computer to rapidly recognize different patterns in language and effectively create associations and weights between different words. So for example, taking it back more simply, the computer is able to learn that the word kitten is more commonly associated with cat and never associated with dog.

Additionally, it's able to understand the context of our language. So if you look at your autocomplete on your phone when you're writing a text, it'll suggest three words maybe and sometimes the word makes sense and sometimes it doesn't. For a long time computers were able to sort of predict human language. They often associated the last word in a sentence as being the most important, but we know that's not always true.

This new technology allows the computer to understand the context of a language that you are putting into it, not necessarily just focusing on one part of the sentence. So they're able to effectively, again, taking a big step back, communicate with us very, very effectively because they understand human language.

and they are just again predicting the best possible answer for you based off of their quests you put into it and those answers are pretty darn accurate a lot of the time it's kind of the

kind of amazing. I mean, for people who just Google a lot, and I still, I use ChatGPT, and I have a subscription that I've got to a paid service, which we will talk about. But even just Googling, you can tell that it has gotten so much smarter because of these generative large language models. Yesterday, I was searching, this is going to sound random, but

I write a column for AARP magazine and I source things for that column. People send me

information at a particular address at the end of my column, if things go to rescue, because it's Chatsky to the rescue. And I wasn't sure if it was at aarp.com or at aarp.org. And I just sort of Googled and it brought up this whole block of text that said, rescue at aarp.org is associated with the column that Gene Chatsky writes.

And my mind was kind of blown because that would not have happened a year ago. Absolutely. Yeah, those AI overviews are great. And that's the interesting thing, the difference between the Google search of the years ago versus using an AI tool that can search the web for you is Google search is going and crawling all the information on the web and retrieving the things that best match your search.

Generative AI tools are responding and communicating to you about those sources when they're using search. So they're going out, browsing the web, finding the things that are most relevant in that deep context you provided, and then summarizing those results back to you, which could just make things so much faster. I mean, what do you say to people who are resisting because they have privacy concerns or copyright or ethics or even environmental concerns?

Yeah, there's so that you just touched on a lot of different things all at once there. So I'll try and take them one at a time. Copyright and ethics concerns, you know, one of the big things with these tools is that we don't know what they are trained on as consumers. We just know they're trained on a lot of information. And there's a lot of, you know, has been a lot of things happening where it's like, you

I'm pretty sure that this was trained off of some copyrighted material. I mean, we pretty much know that for sure, that there's no way that it wasn't. It knows too much about too many things. So there is that question about, did they have the free use rights to use them? Was it fair use?

So that's one concern. Another concern that you mentioned was the environment. I will say to that point, these tools are getting more and more efficient and they're actually as they're evolving, this is the worst they'll be from like an efficiency standpoint, from an energy standpoint. And if you can compare, there was this great analysis done, I wish I could remember the exact source, where you compare the use of chat GPT, a query to one of its lesser efficient models.

and compare it to using a dryer at home when you do your laundry. Using a dryer is way worse for the environment technically than you doing thousands and thousands of chat-to-be-critique queries a year. And with all new technologies, they can seem scary. I'm sure the internet seems scary 20, 30 years ago. Cars were scary 100 years ago. But generally, when new technologies come about,

The good outweighs the bad, and that's the way I like to think about it. When we talk to ChatGPT, I would imagine, I know I do, I reveal a lot about what I'm thinking about, what I'm researching, what I'm working on. And if I'm talking to ChatGPT about my budget or my financial life, there could be personal information in there.

Do we have to be guarded in the way that we talk to these AI tools? To a degree, yes. We want to avoid putting any hypersensitive information like social security numbers along with your name or social security numbers at all, bank account numbers. That's only from more of a cybersecurity risk standpoint, just like you want to post your social security number in a Facebook message with someone else. You wouldn't want to put that in chat, TPT, because there's always a risk of a cyber attack.

When it comes to more personal information, it's up to you. I find the more personal you get, the more effective these tools can be. But the risk becomes these tools often train on the chats that are put into it just to make the models more efficient and understanding that human language. When they are trained, when new models, large language models are trained off of those chats, though, that data is generally anonymized.

fed into the computer to kind of detect those patterns. And then the patterns are established for how people communicate. And then that training data is effectively wiped away. So it just has the patterns that remain. Now, there's always a risk that theoretically, if you put super proprietary information, like a business information,

into that tool that one day someone might ask a question, that's just the perfect question to ask, and it'll generate an idea back that might be yours. So it's also best practice to not put any proprietary information. And if you're using the tool for work to make sure that your company is clear on what you are allowed to do and what you are allowed to input into these

tools. Yeah, that's very, very important. I want to get into how we most effectively communicate with these tools. But before we do that, I think it's

And maybe it's common knowledge, but it wasn't common knowledge to me that there are within chat GPT these custom GPTs. There's a resume GPT. There's an Expedia GPT for travel. There's a website generator GPT. How do we suss out which are the ones that we should be using and using?

whether or not we should be using these versus ChatGPT. And while you're at it, how do we find them in the first place? Sure. So custom GPTs are effectively like an app store that exists within ChatGPT. They are customized versions of the chat experience that different individuals or companies have made to do a specific task.

So if you are stuck on a prompt or really need help with something, you can go to explore the GPT store, which by logging into chat GPT, you'll see it usually on your left sidebar. The little thing that says explore GPTs. From there, there's a search bar. You can kind of type in roughly what it is you're looking for and some results will come up. Now there are millions and millions of these GPTs. You can see kind of user ratings and reviews for how well they work. But from there, you can choose to engage with them or not.

But what I find to be most helpful with custom GPDs personally is building them myself.

So you can't do that on the free version. I do pay for the $20 a month version of ChatGPT. And there are versions of what custom GPTs are with all these different leading AI tools. Google has Google Gems, Claude has projects, Perplexity has a version as well. But what these effectively are is you have a field where you input your prompt, much like you would imprint a prompt, and we'll talk about this later when talking with these tools.

These prompts tend to be a lot longer, a lot more in-depth with clear instructions and clear knowledge about what you're trying to accomplish. And you can additionally attach additional information to train them on. So if you have a repeatable task that requires a lot of context, custom GPTs can help you regularly return to doing that task without having to reprompt it every single time. Can you give me an example?

Yeah, one simple example, like I made one a while back, I wanted to practice my French. And I gave it very thorough instructions on you are a French instructor specialized in speaking with someone that is intermediate level hasn't taken French in a couple years, but did progress through college, I gave them that deep context.

This is how I best like to learn. Gave it a lot of context about that. This is how I'd like you to engage with me. If I say go into conversation mode, we'll practice having a conversation. If I say go into lecture mode, I'll have you explain different things. So I've given it all of these instructions for how I can engage with it a little bit faster. I can attach a recent list of vocabulary words I'm working on and have it incorporate those into its conversation.

Then whenever I start and open that custom GPT, it's already pre-trained on all of that information on what it should do, shouldn't do, should avoid, what my preferences are. And then whenever I open it up, I can get right into it. I don't have to retrain it constantly on that information. Fascinating. Fascinating. And I just want to point out to people what an incredible time saver.

this can be. I'm sure there are some of my listeners right now who are scratching their heads saying, why are you doing the show? And we're doing this show because time is money. And if you are engaging with these tools in a way that helps you in your life, then you quite frankly are likely saving a whole bunch of time, at least once you get

past the initial learning curve. I'm working on a new book and one suggestion that I got early on was to open a perplexity account because I want to be able to know that I've accessed the best academic research on the top

that I'm writing on. And I also want to know that I'm not missing any of the best academic research. And so I was told that's where I should go. You said you pay for the $20 version of ChatGPT. Who should be paying for a service and who shouldn't?

That's a great, great question. The good news is the free versions of a lot of these tools, Chat, GPT, Claude, Perplexity, Google Gemini, offer a lot. I don't think that it's worth paying for them until you feel like it's giving you the value back that you're contributing and then some. If you're just getting started with any of these tools, you're going to have to pay for them.

practice first with the free version. If you find yourself hitting the limit, or if it says, okay, you can only do this many times a day, you can only do so many attachments, you have reached the limit of the chats for the day. And if you find that what you have done is providing you value, that's the time to say, you know what, $20 a month, if it saves me 10 hours a week, it's worth it to me.

But I think a lot of these tools out there, they're getting better and better. And you can switch, you can learn to switch between those as well. If you get hit with a limit on one, you might just pop over to another one. So if you're really on a tight budget, you can do a lot with the free versions today. We are going to take a very quick break. When we come back, we're going to talk about prompting and how to use ChatGPT and these other tools to manage and to make more money. Be right back.

I spend a lot of time in my apartment, working, recording, unwinding, cooking at the end of the day, and since I'm home so much, the quality of the air I breathe, it really matters. But here's the shocking thing. Indoor air can be up to 100 times more polluted than outdoor air. That means the place where I spend most of my days and you spend most of yours could actually be exposing us to allergens, smoke,

mold spores and other yucks that's why I trust air doctor this award-winning air purifier removes 99.99% of harmful contaminants and unlike standard HEPA filters it captures ultra fine particles that most purifiers miss right now air doctor is giving my listeners up to $300 off

Just go to airdoctorpro.com, that's A-I-R-D-O-C-T-O-R-P-R-O.com and use promo code HERMONEY, H-E-R-M-O-N-E-Y. You'll also get a 30-day money back guarantee and a free three-year warranty and $84 value. Go to airdoctorpro.com and use promo code HERMONEY.

her money. It's no surprise that your career is one of your biggest financial assets. The more you earn, the more you can save and invest and build the kind of financial future you want. But earning more doesn't just happen. You have to take action. I remember early in my career, I was doing all the right things, staying late, proving myself in the magazine industry, but I didn't see it reflected in my paycheck.

It wasn't until I learned how to advocate for myself, how to negotiate, how to own my worth that I started to see real financial progress. And that's exactly what Strawberry.me Career Coaching helps you do. They match students and professionals with certified career coaches who help you. So if you're ready to make your next career move, go to Strawberry.me slash Career.

HerMoney, and claim your $50 credit. That's strawberry.me slash HerMoney, because taking control of your career is one of the smartest money moves you can make. We are back with Celia Quillian, author of AI for Life. So the key to using chat GPT or these generative AI tools successfully is prompting. It's

basically telling you what you need in a way that makes the most sense and elicits the best results. So what's the difference between a good prompt and a bad prompt? Context, and as much context as you can give it as possible in a clear, effective way. I like to think that

If you can put your little imagination cap on, talking with these tools is really very similar to talking with a person. Imagine you have an expert assistant that can do pretty much everything for you as long as you tell them what you need them to do clearly. If you leave anything out, they're probably not going to know that that's what you want. They might guess that that's what you want.

But you won't be as well off if you don't provide that extra information. So that's what makes it very different from a Google search, right? Google, you probably put in a couple words at a time. With chat GPT and all these other generative AI tools, you can put as much of your desire as you would like into your prompt.

And we can talk about some structured frameworks as well. Let's talk about the structured frameworks and maybe talk a few examples, including the fact that I've heard when you say please and thank you to these bots that you get better results. Yeah.

Yes, I think the jury is still out on whether that is true. But I think what makes saying please and thank you so great with these tools is it gets you again in that mindset of, oh, I'm just talking to a person here. Obviously, you're not talking to a person, but it gets you in that conversational mode. Whereas if you're kind of trying to think of it as a robot, it's a little bit harder to get yourself there into that conversational approach.

Now when it comes to a framework, my framework that I use is SAGE plus. That's S-A-G-E plus sign. What that stands for is specific context. So what's everything that you need to know that intern would need to know to understand what you're trying to accomplish? What's the background scenario that they need to understand? If you're, for example, applying for a new job and you're trying to update your resume,

What is the job that you're applying for? What's the company you're applying for? What's the role and the title? What's your background? Can you attach your resume as it currently stands and can you attach that job description? There's a lot of information you can provide.

The next is A for action. So action, what is the thing that you would like it to do for you with all of that context? Be specific about the format as well. Do you want it to rewrite the bullets of that resume for that one particular job description optimized for some keywords? Do you want it to critique your resume and give you advice? Do you want it to outline that to you in a table format? Do you want it to provide it in a paragraph format?

Do you want it to play the role of a career coach and walk you through how you might approach the interview?

Next is goal. What's the ultimate outcome you're looking for? The goal in this scenario is I want to get this job. Help me get there. And then E is examples. And this is optional, but do you have an example formatting that you would want to leverage? So if you found a resume that someone else did, the Harvard example that's already online everywhere else, do you want to attach that and say, I'd like my resume to ultimately look like this. How can I adjust it there?

Can you paste in an example of your writing style for your cover letter to say, I'd like you to try and emulate this while we're working through this together?

And then finally, the most important part, I think, is that plus, which is people get really wrapped up in, I've got to have the perfect prompt. I've got to include all of this when I first start engaging with this tool. But the reality is these tools act conversationally. So if it doesn't get it right the first time, you can provide feedback in that follow-up prompt, ask it to adjust its output, say, nevermind, let's restart, or that part of what you said was good. This part was bad. Let's continue from here.

Amazing. As I said earlier, we are a money show, first and foremost. So I want to talk just a little bit, and you acknowledged at the top of the show that you're into personal finance as well, which always warms my heart. I want to talk about how we can use generative AI to better manage our money. How can we use it to...

help build a better budget, for example. Absolutely. Generative AI can be a very effective financial coach, not as good as a professional financial advisor, of course, but it can help you understand where your knowledge gaps are and where your habits are, where your money mindset is.

So one example from the budgeting standpoint, if you go to an AI tool like chat GPT and say, you know, specific context, here's where I'm at, I'm in this much debt, if that's applicable to you, I make this much money a month, or I think I make this much money a month, I'm not even sure exactly how much this is roughly what I spend.

I'm trying to find a way to budget better and then ask it to act as a financial coach and ask you 10 additional questions so it can better help you. That can even help in the prompting process. If you don't know what to say, if you're feeling very lost, it will ask you 10 questions. You can respond back and then say, now that you know this context, help me come up with a better budget. Help me think about how I should approach this. Introduce me to budgeting frameworks that have worked for people in the past and

There are so many ways you can go if you just approach it as a, I'm pretending like I'm talking with an expert that can help me out with my problem. When you describe that prompt, right? Somebody who is, they lay out their salary, they lay out their spending, they lay out the amount of debt that they're carrying, right?

It's a fairly, I don't want to say simple, but it's basic budgeting. Are there limitations? I mean, you said it's not as good as a person. Are there limitations as your finances become more complicated? Can it still work for us if we have retirement concerns or other big picture investing concerns?

It can. I think where the nuance comes is getting in that specific, if you don't know what to ask for, then it's going to be difficult to give you back what it wants sometimes. But if you said, you know, this is where I am, can you create me a starter guide to progressing my personal finance towards a better state in retirement, outlining all of these specific things?

elements that I should be aware of that I may not be. It can create an outline of topics and you can ask more about that from there. These tools can also, if you upload financial statements about a company, they can kind of summarize what the analysts are saying and what their reports are saying. But again, they're merely summarizing. So you might trust or feel more comfortable talking with the human advisor about those things. However, yes, they are trained off a massive amount of information. They can help.

in a great way with a variety of topics. Some main things to be aware of, I do have a use case in the book about tax preparation. It can help you answer questions about, you know, what deductions am I eligible for? Is there any way to reduce my tax liability based off of all this information about where I make income from? It can provide you with some suggestions.

Some of the suggestions it may provide you might be not completely accurate to that tax year. So it's always best to double check and ask where to go or go to irs.gov to be sure that the things that it gives you is exactly specifically correct.

When you, just to go back to your investing example for a second, when you ask it to summarize what the analysts are saying, this is something that we talk about all the time in our investing fix club, which is our investing club for women that we teach live on Monday nights on Zoom. And some of those analyst reports are proprietary. Some of them you can only get if you've got a Bloomberg terminal. Does

chat gpt and the other services do they have access to those to things that are not googleable

It should not be able to, no. So you would in those situations, if you do have those subscriptions, you could download and attach the PDF that you have and ask for questions about it. But what these tools are able to do is when they're searching the web for you, they're doing what you have the ability to do, which is go and browse the web and summarize what it finds. But they're not able to generally get beyond paywalls or any restrictions that the websites have themselves.

One of the great examples that you have in the book is about negotiating. And as we start to wrap up this conversation, I'm wondering if you can break that one down for us because negotiating for that next raise, negotiating for an increase in your fees, negotiating in general is something that a lot of people have trouble wrapping their hands around.

Absolutely. Yes, these tools, again, as creative tools, they can be very effective and helping understand what you're going through, helping encouraging you and also providing tips on how to approach negotiation process, both at work and if you're trying to make a large purchase like a home or a car.

So if you ask it, give that context on this is what I'm working on negotiating. In a race scenario, you might say, I want to increase my title, increase my salary. Here's some background. How do I best approach this? Here's the industry I work in. Here's what I've tried in the past. Or maybe this is my first time negotiating. What are some tactics?

What's a reasonable amount to ask for based off of my role and title and where I live? Providing all of that context will allow it to kind of guide you through coaching how you might approach it.

And from there, you can even say, okay, now I want to role play the scenario. I want you to be my boss or the car salesman or whatever. And I want to have a conversation where I am practicing with you. Now, please don't give in too soon. We want this to be like a real negotiation. Here are your goals. Here are my goals.

Let's go. And that can help build up your confidence as well. Additionally, these tools are great if you're doing a negotiation that's more over email and you're not sure exactly what to say. You can paste in, this is what I'm thinking about saying. Do you have any critiques or suggestions? You can ask it to make you a starting off point and you can edit it from there and collaborate together through that process.

But it can be excellent at just giving you basic and advanced negotiation tactics, plugging into human psychology and things that are known to work better than others, as well as helping you practice to get there. We are having this conversation in the beginning of 2025, the almost spring of 2025. If we were having it a year from now, would it be markedly different?

I think in some ways it would, and in many ways it wouldn't. The interesting thing about generative AI is a lot of companies are starting to incorporate that large language model capability into their softwares and technologies. So, you know, you've got different companies that have an AI chatbot within them that's specialized on that tool. So that's going to be an increasing thing that we see more of. The other new thing that's kind of

The year of 2025 is, I think, the year of agentic AI. What is that? Yes. Agentic AI is when you're using natural language to command an AI to go accomplish things for you, but it's actually using your computer and growing off and doing things for you and checking in occasionally. So, for example, you would be able to say, I want you to plan a full trip and book the hotels and everything for me and just check in with me, and it will go off and...

create the itinerary, browse the website, look up the hotels for you, and ask what's your budget. You can give it commands as to how often it checks in. But agentic AI is when you're literally outsourcing a lot of the manual tasks that would otherwise take you a lot of time clicking around through different tabs in your computer.

So that is something that a lot of companies are investing in. A lot of these general purpose AI tools are starting to introduce. Right now they're on some of the more expensive tiers of plans, but I assume by this time next year, they're going to be a lot more available for most people. - I was gonna say, I have to imagine that

asking somebody to completely book my vacation, asking a tool to completely book my vacation is going to cost me more than 20 bucks. It certainly would. But you know, these tools, again, they're getting more and more efficient and effective. And even the evolution of the free tier of chat GPT has gotten so much more robust in the past year. So it'll be interesting to see what happens in the future with the pricing of all these things.

Celia, what an education. Thank you so much. I've got to say my mind is blown in a lot of different ways, but I'm excited about this. I'm hopeful about this. I'm not, I'm a little scared, honestly, but not so scared that I'm not going to use it more tomorrow than I'm using it today. And you are coming back to answer some mailbag questions for our listeners. We will drop that later in the week. But for now, let me just say thank you so much.

And thank you. It's been a pleasure. If you love this episode, please give us a five-star review on Apple Podcasts. We always value your feedback. And if you want to keep the financial conversations going, join me for a deeper dive.

Her Money has two incredible programs, Finance Fix, which is designed to give you the ultimate money makeover, and Investing Fix, which is our investing club for women that meets biweekly on Zoom. With both programs, we are leveling the playing fields for women's financial confidence and power. I would love to see you there.

Her Money is produced by Haley Pascalides. Our music is provided by Video Helper and our show comes to you through Megaphone. Thanks for joining us and we'll talk soon.