Holly Smithson
All right. I want to open up tonight's program.
And invite all of you to enter, the into it, conference center here and take your seats so we can launch this rocket in effect.
I want to welcome everyone.
My name is Holly Smithson. I am Athena's president and CEO, and we're really excited to see great faces, new faces, and, an old, And we're going to be talking tonight about the future of work through AI. And it's really interesting because, we have arguably the wealthiest man on the planet, who has warned that the two greatest threats to humanity are climate change and AI.
And so as we as we contemplate what to make, as we contemplate what to make of AI algorithms,
Thank you. As we make as we.
Look to contemplate.
Industry. You're like, what else is new, Holly? So we certainly appreciate that we are very comfortable, for those of us that are joining in the room and certainly remotely around the world, we are very, very comfortable with the unknowns. It's part and parcel of why we are in this industry. And I want to tell you that we're excited to partner with, into it.
And before I hand over for, sponsor remarks to my colleague and very good friend, Kimbra Brookstein, I just want to do a little bit of a shout out, today into it was recognized as America's top three best companies in the country.
And I just I just got a little bit of chill. I just got a little bit chill.
Down the spine, because when I think about Athena, we're we're going global. We announced it at our, lifting wall climbing summit a couple of months ago and taking all of this rich content that has been curated and developed, from our Stem community. And we want to make that available to all the women around the world.
And if they're not fortunate enough to have an Athena in Southern California or down the street, where removing all geography and all boundaries and all language barriers so that the women around the world and their male allies now have access to all of this really rich content that's designed to help these women advance in their careers, to build their confidence, and to see other women that are leading in Stem.
And so we're just super excited. My, my, being able to have a partner like, into it that has been a long standing supporter and sponsor of Athena. And then to be recognized as one of the top best countries, companies in the nation, it just, it really reminds us of how special we are when we have companies, that put their money where their mouth is.
So with that, I just want to say thank you, Kimbra for your partnership. And Kimbra has been the face of our partnership for, for many years. I've been with the company for the last seven years, and, I think it was day one that I got to to meet Canberra and the team at Tech Women at into it.
And before I hand over, the microphone to Kimbra I also want to do a little bit of an embarrassment, because it's just something that I enjoy doing. So I want to bring attention to somebody who's probably going to be very uncomfortable. But when she comes up here, I want you to notice, if you can't see her and you, you're getting, like, blinded.
I just want you to know that's an engagement ring.
So congratulations, Kimbra Very, very excited for you.
Kimbra Brookstein
Leave it to Holly to embarrass me. Hi, everyone. As I mentioned, my name is Kevin Bergstein. I am a principal, DEA partner and leader of our Tech Women at Intuit initiative here. Globally. It into it. I'm thrilled to host. Gosh. Now I want to say maybe it's our sixth annual ai ml event. We launched it at Intuit probably six years ago, where we continuously put a focus on AI and the importance and into it being an AI, driven platform.
And now all the new, work going on that you'll hear about later. But I just want to thank, Athena for, like I said, being a longstanding partner. We partner and align with companies that share the same values and do the same work, that we want to see in, in the representation and parity. And so we're thrilled to continue to partner with Athena and to welcome everyone and host you tonight.
Thank you for all those that are here. There's food, drinks and swag. Don't forget swag. And thanks for, joining virtually. We hope you find value. And with that I'm going to introduce Cheryl who's our moderator. And she will bring up this appearance. So thank you.
Cheryl K Goodman
Okay. Thank you Kimbra. Thank you Holly. What an amazing group of women. As per usual with Athena events, I got to tell you, Holly indicated early that you really do have an amazing panel assembled. And you're going to see that tonight. And I'm going to announce each of our guests, who are all phenomenal leaders in AI. And you're going to I promise you, you're going to learn not only going to learn, but you're going to be inspired.
And I think you're going to build and create right along with into it. So yes, setting the bar high, but hey, that's what we do here at Athena event. So it's at that bar. Hi. So with, further delay, I'd like to call up Sheryl Anjanette. She is the CEO of parsley360. And Sheryl and I like to joke that we have the same mother who named us both Cheryl.
So, you know, she likes the name, but Cheryl is amazing. Leader. She and I have worked on, many projects, in the past. And so you're going to really be delighted about the promise that she has with, parsley 360. And so I'm going to call up our next panelist, Yi Ng, who is a director of product management here at Into Intuit
And I cannot wait to dive into all the goodies that you have, intellectually and otherwise, on the front of AI. And then our third and final panelist is a lovely Miss Lena Skilarova-Mordvinova, and she is the CTO of a fantastic company, The Good Face Project. I write about it in my book, which is on the table there, How to Win Friends and Influence Robots.
And I myself, have been a robot wrangler at Sony, a decade at Qualcomm and a few IPO startups in between. And so. Well, we know that I certainly isn't new since the 1950s. AI has been around. It's really that consumer experience in November of 2022, really, this month is the first time that we had generative AI as a platform that we can engage with.
So it's really nice two years later to have three stellar leaders to talk about the implications of what that means from a workforce standpoint, what that means from an anxiety fear work standpoint, what that means from a learning standpoint, and then some projection on what that means for the future. But before we dive in into all of those questions, what I'm going to have each of our panelists do is tell them a little bit, about themselves and their company and their mission.
And we'll start with you, Sheryl.
Sheryl Anjanette
Did I there I am, did I hear fear and anxiety in there.
Cheryl K Goodman
A little bit.
Sheryl Anjanette
Just a little.
Cheryl K Goodman
Bit of a scared image in it, but yes.
Sheryl Anjanette
Oh, gosh. So again, my name is Sheryl, engineered. I'm the founder and CEO of parsley. 360 and we're kind of more on the human side of the AI equation. So we are an AI driven, performance optimization company. So we're interested in corporate wellness and corporate performance. And we are a holistic solution. So the experience with parsley 360 is you have a, coach in your pocket or a coach in your hand, 24 seven for every employee of the organization.
And the organization itself is getting root cause analysis. So it's very data driven but private, de-identified.
Cheryl K Goodman
There's a great summary. And I'm sure, there's so much that into it can talk about. So we'd love to hear not only what your role is, but what are the big projects and platforms that Intuit is embracing and leading with AI?
Yi Ng
All right. My name is Ng. So nice to meet everyone. Very excited to be here. The fact that you guys are here means, it's a big deal. And into it, our mission is to power prosperity around the world. And it's money in people's pocket at the end of the day. And it's really important for us to work through, whether it's a consumer side, to help people manage their finances or on the business side, to be able to help their they run their business.
I work in the Customer Success team. Because today when we think about where I can be the most, leveraged in terms of applied AI, it's really about how I can help you answer questions when you're you're in the product, don't know what to do. And so we have a system, an AI system that can help people answer the questions.
And I'm leading that effort across, into it, for all of our products. So very excited to be here and excited to talk to each one of you.
Cheryl K Goodman
Awesome, excellent. And Lena, not last Lena, but Lena, your last. Go ahead.
Lena Skiliarova-Mordvinova
Thank you. Cheryl. So my name is Lena Nova. I'm a CTO, a cofounder in the good Faith projects, our companies, over six years old. And, we developed. I for chemistry. So, we have created a system that knows over 160 millions of molecules. It knows there are features, functions, toxicity can predict, their qualities that are yet unknown.
It also knows how to use them in different products. And, we productized that for, creating a SaaS solution used by chemists, from manufacturers, brands and other companies that are developing personal care. So, for example, they can create, anything that you put on your body now a platform scan for regulations in one second instead of spending months on that work, develop a product that would be best suited for specific skin types.
So our car also knows about how skin works, what is best for the skin. You can create nontoxic products. You can, find their doppelgangers on the market, you can replace raw materials in those products. So basically, AI is the core. And based on that core, any vertical can be standing. And in our case, it's the personal care vertical.
Cheryl K Goodman
Thank you. Amazing amazing okay. And so you know about our panelists. Now I want to know a little bit about you. So raise your hand if you're using AI on a daily basis as you do it. Everybody okay I would love for folks to shout out what they're using. All right. Very good. Any more.
Okay. There you go. All right. Great. Well, so we know that ChatGPT certainly is the default. You know, it's probably the, the biggest brand value, but there's so many layers and so many limbs and so much power behind all of these systems and the diversity in these systems. And so I want to talk my first question really about, the education.
Right. How do we get, society to a world beyond ChatGPT? And more pointed question is I had mentioned we've we're two years in. What do you think will be in two years with the workforce? What do you think the workforce impact will be? And maybe let's start, with you on the Intuit front. I'd love to hear your thoughts and then we'll go to you.
Sheryl.
Sheryl Anjanette
So I think there's two questions. Maybe one is where are we with the technology? ChatGPT today and where is it going next? The second question is what's the impact to the workforce? So, I think technology is really hard to predict, honestly. I remember, I like to tell stories, five years when I, I worked in AI for a long time, I started in data and happened in life or maybe almost 6 or 7 years when I was, working on the foundational, infrastructure of AI platform.
One of the things, I talked about is it would be awesome if one day everybody walks around into it and they're all talking about AI, maybe 50% of them will talk about AI. And that was actually on the our goals, as a leadership team. And we are at a point where everybody 100% of people are talking about AI.
We did not predict that. So, I think, Suzanne, which is who's our CEO, I've been with into it for about 13 years, worked on all the different products at into it. And one of the things Suzanne talked about is I like electricity. However, I think one difference is what we if you think back in the electricity days, you have to build infrastructure for electricity to be used in the masses.
We are now building infrastructure. Infrastructure as fast as we're building the, the actual algorithm and all the data behind it. And this is a time where infrastructure is coming, like in speed. So potentially two years from the time ChatGPT launched to where we are now, we're still figuring out how to apply it in a way that create efficiency and business outcome.
But in two years time frame, I bet you will see the business outcome and the business efficiencies, not just in one industry, but in multiple industries. And from a technology perspective, it's very much a, infrastructure. Sure will probably be set up, in a way that's pretty widespread, so that everybody can be leveraged not just for, chatting on you know, on story building or generating images, but being able to actually use it in applied AI.
And, it's actually a little hard to imagine what the future could be. I don't think very few of us can think that far, especially with the speed in which it's going. But we know it's changing fast.
Cheryl K Goodman
Yeah. And and then on the workforce part, just to kind of double click on that a bit, it's those roles are going to change. And so now we're in the era of generative AI, a lot of generative AI with these large language models. But I heard a phrase recently, and I think I'd love to hear your reaction to this phrase is that generative AI is great, are good for almost everything.
There's so many general good purposes, but it's not great for everything. So when we think about attorneys or we think about, you know, where you have Sox compliance or you have these nuances or, verticals or near industry that are very specific, it's kind of hard to be able to use a general AI to achieve an outcome. Any thoughts about that and how that affects the worker?
Yi Ng
I would say so a month ago I would say, yeah, that sounds bad, right? But we have made so much, so much progress just in the last couple of months. And I think the way to think about it is generative AI. You guys kids, you have kids. We fell out when we were little. Remember the fill in the blank?
There's the fill in the blank finish. Right. Some sentences fill in the blank. Images, videos, text. It's just filling in the blank, you know. However, there, large language model is starting to have reasoning capabilities. It because it understands language and understand how we communicate. It's starting to understand why we're saying something and what are the consequences in which we're saying something.
And, what's related to what we're saying? And therefore that reasoning and able to assess what's connected to what we're saying is starting to happen. We're also building technology, infrastructure that's starting to understand how we retrieve the information and how we present the information in such a way that does not hallucinate. So 18 month ago, you know, even few like 6 or 3 month ago, we see a still a lot of companies trying to work through the hallucination.
But Wragg, has come a long way. Function calling has come a long way. And so there's a lot of technology out there that's starting to look at how do we govern, large language model in such a way that it's starting to tell truth. One example is multi-source. So, I source one source of information, I get another source and I get another source.
Now I know, oh, this is truth. I'm not just making it up. And you're, you're if you go out there and play and everybody seems to be playing, you'll start to see you're starting to get information that are closer to truth than the hallucination starting to retrieve in the background. And then let's talk about the workforce a little bit.
At into it, we have, one of the largest, network of experts and especially working in customer success, we work on connecting people, which is our consumers to the experts. And, that connection is really important, but our experts are people who are using their current skills like product support skills or accounting skills or tax skills to be able to support our customers.
And those skills are starting to be, in some sense, democratized or starting to take over by some of the AI and machine learning capabilities. And there is definitely a path in terms of how machines automating some of those things. But there's also a path in which machine is helping our experts to do a better job at their job and upskill their job, allowing them to do more.
So again, it's not a simple question of how it affects workforce, at least from a Intuit perspective. We're looking at, both creating automation and efficiency. But at the same time, how do we help people get the better skills, uplevel their skills, and helping them learn more with AI so that they can be even more productive? Because at the end of the day, we have to solve for human to human, and the technology is there to help us become a community, and become a society, not to become less of it.
Right? Right.
Cheryl K Goodman
So as, as the, the work product evolves up with the use of generative AI, we're actually operating on a higher level of problem solving, right? Yeah. Yeah. In theory. Right. So okay. Fantastic. Thank you for that deep dive on that. Phenomenal insights. And so, Lena and and Cheryl, both your, your your leading companies, you're hiring people.
Really the same question. But from your perspective, what does this mean for the skills that you'll hire for what you're looking for, what you think the outlook of the job, is changing? And let's start with you, Cheryl.
Sheryl Anjanette
Yeah, sure. So I think the first part of this question is really about what we'll call adoption. And it really has to do with human behavior. I think when you see these large language models, we have to realize that there are the really large language models like Claude and like ChatGPT. But now there are also smaller language models, smaller large language models that have fewer parameters that are.
And then there's another level which are expert systems. What I feel about adoption is that it's all about someone's experience and also that whatever that technology is in this case, I that it's matched by a need that person has. So the more closely we're solving a problem, matching a need and having a good experience and a good result, the faster the adoption will happen.
So I do think to your first question, Cheryl, that we're going to slowly see people move away from ChatGPT not totally. I think ChatGPT can always be in the mix, but I think as they start to have more defined goals, more clear, specific needs, they're going to want system that match those needs. And so I do think it's a process.
It's an experience. It's experiential. We have to have good outcomes. As far as my company and the future of work, I think I think less about how I'm hiring for my company as much as the companies we serve and how they're hiring, and what future of work means for them, because we really help companies with this human side of not just this, but anything that has to do with human behavior, emotion.
How am I dealing with that fear and anxiety about the real clear threats that Holly mentioned in the beginning of this? They really there really are threats. But how do we deal with this and how do we look at the opportunities in this? So we prepare for the threats, but we take advantage of the opportunities. So I think, future of AI is going to really expand depending on how those guardrails are set up, what the experience are, if we can keep the threat level down and into tact, and we can be really clear and honest about what those are, we can't just kind of push them away, steamroll over them.
But at the same time, we solve in a way where we're watching out for people, individuals and organizations, security and privacy.
Cheryl K Goodman
Right. And we're still human. So like, it's still got to work in a human system. So same question for you, Lena.
Lena Skiliarova-Mordvinova
Thank you. Cheryl. So I think you already heard me saying this. Large language models are not the products. And so, I agree with Cheryl that, people will be using ChatGPT less and less simply because, the thing that was trained on all the knowledge from the internet might lack some quality, and we'll see that the quality will actually go down and down.
And so, another thing about LMS is that the way they're designed is, there is always a probability of getting around the answer, just by the nature of the structure of the data and how they work. So I'm pretty sure that we'll see, one more iteration on the architecture of large language models. And that will be more interesting and more, based on reasoning.
But for now, they're still lacking that.
Cheryl K Goodman
Yeah.
Lena Skiliarova-Mordvinova
And so, for example, in, our system, LMS and smaller language models, there as an existing part, but definitely not the course we're developing and using many proprietary, models that are allowing us to getting actually correct answers in 100%, which is not something that you can get without alums. And alums definitely give us a lot of, speed up.
We are capable of doing more. But from the perspective of the chemistry, they actually are not answering the question. You can never make a mistake in chemistry. That's not something that you are allowed to do. But, you know, you can change words in, in a sentence. So LMS are so very good for, something that does not require 100% precision, but they're not there.
And I frankly, from the technical point of view, I think that the structure is not built to be there.
Cheryl K Goodman
So is it the case? Is that industry specific? It is a challenge, right? Yeah. So it's you know, kind of back to that phrase. It's good for a lot of things but not great for industry. And I think that's the impact to the workforce that I'm trying to uncover. Because we weren't really having those conversations last year. Last year it was, oh my God, I is going to take my job.
All writers should just give up and go home very themselves. Which is why I wrote the book, right? As someone who is a prolific writer, like, I'm going to put these things stress test and I'll tell you what is a the reality is, is if ChatGPT writes a book, please don't buy it. It's not that good, right?
You know, you're still very much a part of the equation. But back to what does this mean for the workforce? And so I want to now ask a question to the audience, and online to, I think you guys have chat functionality. There is really are you in companies now that are embracing policies that say, hey, here's our, AI policy.
You can use this to have efficiency? Or are you in organizations that have not decided yet in, you know, I think this is an indicator of, you know, how industry is adopting AI. Any hands, any shout outs? Yeah. Let's do thumbs up. If you got you got a policy, you got a pro AI policy in your organization.
Or you can do this if you don't know. All right. Are you could just wave at me or you could tell me you could use your middle finger, like, maybe not do that on the video, but like Cheryl I clean it up, girl, clean it up.
Yi Ng
And share it. Maybe I just wanted to add something, you know, into it. It's all about tax and accounting. And we keep, your financial data very, security to heart. We're a trusted company. And for us, at the end of the day, it's about the trust, and it's about the data of our customers, and it's about the answer that we give.
You know, in chemistry, we can't make mistakes and tax, you know, way. Right? Either there's a lot of industries in which making mistakes is costly. And, and also, I think even in industries where creativity, you know, also matters, right? Responsible being responsible AI, there are a lot of ways in which the infrastructure in which we're setting up can start creating guardrails in which, for example, auto evaluation process.
I don't know if many of you heard about it largely, which model is not just good at spinning out, you know, fill in the blanks, but they're actually good at evaluating themselves. So you could, before you could ask a human to say, is this accurate? Give me, 400, you know, question and answer pair and then go evaluate it.
And we have people, you know, sitting there and going, this is accurate, this is accurate, this is not accurate. You could ask a large language model to evaluate itself and say, hey, what's the answer that you're about to give accurate. And once it actually ask itself it's actually pretty good at going, oh, that was not good. Or I was hallucinating, you know.
And so those are some of the techniques. So we're starting to implement internally to say, hey, can you auto eval, but we still need human in the loop at the end of the day to adopt to that, accuracy of the right accuracy in which when we respond to our customers, we are promising our customers that, we we have the content, we have the information, but we also are promising you that we're giving the information that you you need to know and that you you should know in terms of the amount of accuracy that into a put behind its brand.
And so when when we think about, you know, how we're implementing these, sometimes we use hybrid approaches in which large language model is part of it. And you had some curated, responses. And sometimes it's about using auto evaluation and other processes. But there are many ways in which this, you're starting to see in the last 18 months, the industry has developed a lot of different algorithms in which we're starting to practice and test and learn and lean into.
So I wanted to add to that because, you know, you're asking it seems like everybody's doing, different ways of quality checking, writing. The quality checking is super important, right.
Cheryl K Goodman
And it's industry specific. Right. As a financial data versus, you know, medical data, HIPAA compliance and all of these things. And so I think, from, from the research in the data is like, this is where, we're going to see the changes evolve, where it's going to get better and more bespoke for companies. But back to the root of this panel is what does that mean for the worker?
What does that mean for you? How do you rise to meet that challenge. And what what should, you be doing? And so let me take this back to the panelists. What would you recommend as your go to plan? And yes, let's start there.
Lena Skiliarova-Mordvinova
Right. So well actually responding to what you said, that's also part of the future of work workforce, right. Yeah. You still need industry experts. You still need people who know very well what they're doing. They will be evaluating what, different very specific systems returned. And there will be either training or using the data. But it's still if there is no human in the loop, the value of the system itself is very low.
Right. Because, the trust will still not be there.
Cheryl K Goodman
And and we know business moves at the speed of trust, which I think is what you're really getting to from an Intuit perspective. So yeah.
Lena Skiliarova-Mordvinova
So, from the perspective of, people who are in this field and who are using this on a daily basis, it's mostly now about iterating, you know, being fast. You need to know very well what you are doing, your processes, which part. And that was today. So made or which part of your daily schedule can be automated, like can you make an algorithm out of that?
If you can say there are no unpredictable things in that, line up of actions, this means that this can be automated.
If there is something unpredictable, if like, some emergency can happen on everything, then like, you actually need a human in the loop and watch out for that. And, so from the perspective of learning, this means that you need to constantly check new systems. New, smaller alarms are actually quite good. Sometimes they're giving better results than bigger ones.
Like we are frankly seeing better results from smaller ones. And, from that you would basically like every week you will see annual. I'm coming up and your task, is actually take that alarm and check on the base, list of questions. Right. And see if you are actually getting better results. And if you are capable of doing that iteration and constantly being in the loop, then you probably will end up with very, very good products at some point.
But, just basically monitoring what's happening and pretty much be open minded and not stopping on one system and keep keep checking.
Cheryl K Goodman
What I write about in the book is that, you know, we know a lot of the universities when we talk about training and getting ready to rise to meet the needs of the new workforce that's powered by AI is there's not a lot of coursework out there. And I think this is the challenge that we're getting to it.
So I want to just get really, really practical. I want to just say to that level, if someone is here tonight or is there they're watching tonight, they're like, I just don't know where to start. Can we get to that basic, fundamental level? What would you start? In my book, I recommend community. I recommend being in places like Athena where you have experts like yourself.
Because I think if anyone says that they really, truly know all of the answers. I don't believe them. Right. I think technology is moving incredibly fast, and we know this. We don't know the trajectory of that. So can we talk just a bit about what would be a first step for someone that says, you know what, I'm ready to do a career shift or I was recently laid off, or I'm afraid I'm going to be laid off.
What would you say their next step, next right step to participate in the AI economy would be a show. Why don't we start here?
Sheryl Anjanette
All right. So and we should all always be thinking about this because whether we're laid off or we're just reskilling and upskilling, it's really important that we stay nimble. And so the very first thing I would say is I would do a self-assessment on what you are passionate about and what I mean by that is not just what you think you're passionate about, but when you say it, do you feel it in your heart and does it light you up to someone?
Start talking about something and all of a sudden you're leaning in and you're like, yeah, that's cool. So that's your passion. And then look at your competencies, your experience, but also, even if you don't have experience, if you're really good in certain skill sets, that would lead to something and then try to see where those intersect in some of the newer jobs.
The second thing I would encourage everyone to do is if you're not in a leadership role, if you've never led people or you haven't studied leadership, study leadership, become a leader because we need the skill set of leadership. No matter where you are in an organization, because with AI you have your own team. Now you can get AI to do things for you, but you have to prompt it in the right way.
You don't have to worry as much about its feelings.
This is a great in the art. The solopreneur is really the power of ten people that you know. It was last year, right? You can have the same efficiency. And that's, I think, where the fear is, right? Yes. Is, okay. Now at scale companies need a lot less people. So leadership great advice knowing yourself on AI specifically, what would you recommend that they.
Cheryl K Goodman
What would be the next step there. Yeah. And I want to say that, there have been reports that the reskilling and upskilling industry is going to go up to something like $9 trillion, but it's not new, and it's something like 82% of skills that we needed have like a third of them have, dropped off our requirements anyways.
Sheryl Anjanette
So to some degree we always need to be reskilling and upskilling. So and then I would get a mentor, I would get a coach, I would continue like, you're so it's so great that you have these network environments and ask people and network and just always be ready for the the next thing.
Cheryl K Goodman
I think that's fair. And I also say there's a tremendous amount of I have just throw this in. There is a lot of free resources out there for intro to ML. And, you know, I think, you know, these jobs are being disrupted. You hear, there's no, you know, don't study coding or don't study this or don't study that.
I think the best way is within your industry. Get with your industry group and then really understand, okay, what is my industry group doing. And this is why Athena is so important. But I also, lead at Startup San Diego and we do various bespoke specific entrepreneur for I, I for entrepreneurs. It's like, how are you being disrupted and what tools do you need to know?
And it's not just ChatGPT. So but I'd love to just get real practical there. So can I have the.
Sheryl Anjanette
I would say, you know, first let's look at human as a humanity. What are we good at? We're curious. We're good at adapting to change. Or at least some of us are. I have two, two kids, two daughters. And I'm always asking them, just be curious and then willing to learn new things, you know, and go, go explore.
And so what are the things that we should be exploring? So maybe a little contrary to what people believe. So. Well, large language models can do language. Well, why do we need to write? Right? Large language models can code. Why do we need to code? Right. But I would say on the flip side, go be a great communicator.
I think going back to the leadership community, action is the foundations of humanity in human, and how we relate to each other. If you understand how to communicate and how well you can communicate, I we go all the way, all the way in because large language model is one where at least like AI is starting to be able to starting to communicate like us.
And if we want to have a relationship with AI, which is coming, then we have to study it. We have to study it, understand it, and then build on top of it. And the stuff that we're really good at, we're curious, we're adaptable to change. We can communicate. We're all about relationship building. So go and go and really hone in on your communication skills.
That's my first ask. My second ask is so you so what? They can code, right? Go learn coding and go learn computer science and go learn data science and understand stats and understand math and Stem and Steam and those are all really important because at the end of the day, who's going to be building those computers? We are.
And that's the skill that you will need. And our next generation is going to need. So, because we're curious, because we're adaptable to change and because we can and also because we may not necessarily see what's coming. The best thing we can prep ourself are the things that we know what we can do. And I would say go to the base of what, you know, what we're all about, how we communicate.
And then go into tech in the sense that everybody in this room. Because I said, and when, when we first opened, you're all here. And that's great because you are here to learn and to figure it out. And that's already the first step. And therefore take the time to learn the tech. Everybody raise their hands and they're playing WeChat, GPT three or some some version of JNI tools.
That's great. And that's what's really important that you're learning. You're leaning in and that's what we need from, you know, humans, because humans will be building a relationship with AI. And it's it's just a matter of time. And so we need to go figure it out what that means. And therefore everybody should lean in.
Cheryl K Goodman
Yeah I love that. And so so we'll you'll answer that. We'll close that out. And then we're going to shift to business in use cases. But you're final thoughts on work work readiness. And I.
Yi Ng
Thank you. Well of course it's first of all great that you are all here and you can be prepared. You are already thinking about it, right? You are already there. You started your path and you probably will continue. You will be prepared for the change. You will explore the tools that are coming on the market. Not necessarily you are working in coding like everything connected, specifically to creating the AI, but maybe you are in some industry that already has some tools that are emerging.
If you, view yourself as a startup. One of the things you should always know is who are your competitors? So, if you are working in some specific field, look at, what system is are emerging, what are the new technologies, solutions, SaaS platforms that actually are, very close to what you are doing.
Explore their capabilities and see if, you may and hey, maybe you can speed up what your typical, actions on daily basis are. So if you are getting there and you know what is on the market, you are already getting a huge advantage comparing to like the whole planet. At the end of the day, everybody who is using the tools and you are not necessarily going to be developing them, just using them, you will be already way faster and more productive than people who are not knowing or not using those tools.
It's not going to be one big thing that will shift how we work. It's going to be more like an avalanche of tiny things that will pop up in, different aspects of your routines. Like we already are very much used to auto correction, right? Like nobody was concerned about, the correction, like in our when we're typing, we're just taking it as, like a given.
But that was like basic pre-history of what's happening right now. That was one of the building breaks of where we are. And, it's going to be like all those tiny things that will speed you up. You will be, pretty much in a very good place if you know where those bricks are and you know where you're stepping and where are you going.
So my advice is, whatever your industry is, just give yourself as a startup and check out what's outside. Like, maybe you're not building it, but just see what's happening.
Cheryl K Goodman
Then I would also say what I've observed over the last couple of years is that, if you're in an organization where maybe it's not pro AI, you know, raise your hand to be that advocate, say, listen, I'm willing to research in my off time to be able to curate these tools. I'm willing to, put together a task force.
I'm willing to be that person. And I think that's really that curiosity that all of you have spoken about, about Lean in learning is like, just be that person that says, hey, I'm willing to sort and figure this out and build together that Tiger team, because then you become the default expert. And we do know that it isn't going away.
May as well get in front of it, may as well lead it and just show your willingness and expertise and you'll learn a tremendous amount. Okay, so with that, let's switch now to emerging use cases and business outcomes. So I love this analogy. And in Microsoft, last year actually I was at the world Economic Forum.
Satya Nadella was talking about how they had used their, proprietary large language model to come up with a new battery chemistry. Right. Battery tech has been such a difficult problem to solve for many, many years. But they've come up with this very novel, composition, this battery chemistry. And these are the types of things that we just blow your mind about the power.
So it's not just the power of generating a poem or a goofy, material or an article, but really changing the world, with tough problems. So that's an interesting use case. Are there interesting use cases? You're all living and building use cases in your day to day life, but are there use cases that you think demonstrate and illustrate how significant the power of generative AI gen AI can be in anyone?
Start wherever you wish.
Sheryl Anjanette
Well, I can start or I can finish because my use cases are not around efficiency. Unless we are talking about human efficiency, which is also often subverted by our own fear, anxiety, self-sabotage, our, you know, feeling out of emotional alignment. So what we are solving for, I'm not seeing written about a lot yet. But it's really around corporate health and individual health and wellness, and it's this idea of, I don't know, maybe you're a working mom, maybe you manage a team, maybe you're feeling a little burned out and you're trying to balance all things.
Look, professional all the time. Maybe you're a woman, so you're held to a higher standard, and talking about having to take care of the kids might set you back. There's a use case for you. Maybe it's three in the morning and you can't sleep because you're worried your sleep is being affected. Or maybe you have aging parents, or maybe you're stepping into a difficult conversation with your manager and you're not quite sure how to handle it.
There are a lot of moving parts, and you're feeling a little emotional, and you just need somebody to listen to help you get centered. We need to feel seen, heard and understood as human beings. And a lot of times the answer to that is simple, but it's not there. It's not given to us even by well-meaning individuals. So what we are developing is really emotionally intelligent AI.
They can deliver empathy at scale. This idea that you can pull out your phone, literally click on your coach and have a human like conversation. Now it's not a human on the other end. It's an AI, but it has a human tone. It will listen generously. It is non-judgmental. It will help you get back to your grounded self, help your prefrontal cortex critical thinking, get back on track with your emotional brain.
Cheryl K Goodman
And what do you call I forget what you call your un partially that platform. What's your eyes name?
Sheryl Anjanette
Tulsi. So Tulsi is the pulse.
He's going to walk you back from the edge basically. So your use case is really being that counterpart, that extension, of you in a digital, AI powered wellness coach. Is that a great way to say.
That's a great way to say it. Exactly. So that's a use case. But the other use cases look all right. I'm starting to get it together as an individual. My organization, my team is doing better. But the organization, we still have some problems with our culture or a values mismatch or they're, you know, whatever that is. We need our leadership to also maybe make some adjustments, see their blind spot.
So we kind of connect the two and we're dealing with the organization as a whole, like in a consulting role and at the same time with the individuals. So that's a use case. And I think what it does is it helps all these other use cases. We I think we talked about change management, right. And being adaptable and flexible and curious.
And we're really good at that when we're good at it and we're really bad at it when we're not.
Cheryl K Goodman
See where everyone could definitely use, Tulsi, just segue to our other panelists. I would just say this is, I subscribe to a newsletter called eye of the day, and this is where I see all these really radically fun use cases. So I have 100 in my head. But I want to hear your favorites. What what do you think is, a great use case that demonstrates the power that we have ahead?
Yi Ng
So I'll talk a little bit about, applications in our system. So basically, the core of the system that knows, chemistry and can predict, ingredients toxicity is actually capable of filling the data gaps that, will probably not even be ever covered by scientists, because, there is not enough research to prove that some ingredient is toxic.
It can be a carcinogen or anything else, and you need to actually a lot of research to be, making some statements about that. So AI is actually great for that because on mass scale it can basically predict, the quality of ingredients and, actually help you with your health eventually. So what my favorite example is aspartame, which came up a couple years ago.
It's an artificial sweetener quite often used in, different diets, drinks and, unfortunately, that ingredient at some point showed up, in articles with some opinions actually stating that it is a carcinogen and, since that, ingredient is very widely used in the industry, that was very much toned down. And once we, the, the day one, the article appeared, we actually went to Kara and ask her what she thinks about that ingredient, and she was very much confident that it's in casino gin, that it's, mutagen, like, over 95%.
It's a carcinogen. Over 90%. It's a mutagen. And it's toxic to specific organs. Which means that, after you accumulate some of that ingredient, of course not. Like, after a drop, but but in general, if you are taking that ingredient all the time, it might be, actually bad for your health. And, just at the same day we went and asked, okay, so what what about other, sweeteners?
And she was able to find other ingredients that are very much similar to that. And, you know, there is no article about them, but it is like the knowledge the public.
Cheryl K Goodman
Health issue know it would have taken years to compile that data to get to that result. And you did this in moments.
Yi Ng
Yeah, unfortunately, we did this in seconds.
Cheryl K Goodman
Oh, and silence. Go get a girl.
Yi Ng
But unfortunately, you know, God will not push FDA to say, okay, we need to stop something using using that ingredient. Right. We, we will not see that change happening. But actually, if you look up there, for example, even chewing gum without aspartame, you will see, like, they're stating their are aspartame free. And so you see that this change is happening.
People are more conscious about what we're putting in your body and on your body. And we can help with that on scale, like for millions and hundreds of millions of ingredients.
Cheryl K Goodman
And that's significant because that really does speak to the power and truly check out good faith. I mean, it really is. When you get to the website and you see, you can see very clearly how powerful this system is. I swear, I'm not an investor or anything. I just love it. So, all right, that we we have about five more minutes and we're going to break it into question.
So we'd love to hear your insights.
Sheryl Anjanette
So remember I talked about powering prosperity, which is our mission. Add into it. I think the question the key question is are there business outcomes that AI is driving and is that happening in the industry? I would say pretty mixed and then pretty nascent still. In that sense, a lot of companies are experimenting, trying to figure it out and, trying to figure out if there's a business driver behind it.
In general. And into it. I'll speak to a couple of different use cases that, we're working through and experimenting. But I'll also talk about one that we know are already driving business outcome. So the I think the one that's driving business outcome has scale and also has a specific,
Specific solve for a specific problem. And that's usually when it actually hit, hit the go go post. The, the work that we're experimenting is one on the small business side. I don't know how many people are small businesses, but, we have two here. As you start a small business, it's really hard to do that job that you don't have.
You gotta get your own marketing person. You got to get your, financial CFO and going to get your own accountant and tax person. And one of the things that we're trying to figure is, how can we provide all of those expertise to a small business owner who doesn't, may not have that network to be able to reach out to those experts, but we have the data to have all those experts.
How can we provide that data? We know that could potentially have drive huge, business outcome. But again, we're starting to test into small pieces like, CMO in your pocket, CFO in your pocket, accountant in your pocket, tax advisor in your pocket, through all of the different, Credit Karma, MailChimp, QuickBooks and TurboTax, business lines. So that's like one big problem.
We know the small businesses need it and they don't have all the resources. So how do we provide that, you know, from a, holistic perspective? The one that is actually driving business outcome is really in the space where, many of you probably call customer support. When you pick up the phone, it says if for for one dial up for if you're interested in x, y, z, eta one.
If the interested in TurboTax, press two. Where experimenting in how a human like, agent tier one agent as a bot can actually start responding to you, right? And then be personalized to you. So we have contextual, bots that's already, in, in the, if you guys use TurboTax or QuickBooks so you can start trying our help systems is digital help.
Digital help is launched to over 60 million users, across all of, internationally. And mostly through QuickBooks and TurboTax interfaces, where you're going in and trying to figure out, well, how do I use QuickBooks, how do I use TurboTax, and what what's happening and why is this not working? And we have all of the transcripts and all the content from years and years of people calling customer support to be able to answer those questions.
And we're answering them accurately enough that we can, one of our key drivers is about contact rate. If people are not contacting human, we're able to save money, to be able to use leverage that to and.
Cheryl K Goodman
Better customer service because you're getting that answer just like that.
Sheryl Anjanette
After, you know, I talk to people, nobody wants to call, nobody wants to call. Everybody wants the problem just solved. And we're able to actually answer people's questions. So it's about speed to resolution. It's not about whether it's answered by a bot or human or voice or video or, you know, avatar.
Cheryl K Goodman
It's efficiency.
Sheryl Anjanette
It's efficiency. Yeah. So we're starting to drive efficiency. And that is, really important at scale. And so when you think about the what's the first thing, machines can do now, is driving efficiency. What are things they can do. Soon I would say maybe two years ish is starting to help people make decisions, but people are still making decisions, and machines are starting to making the right recommendations to help people make those decisions.
So there are going to be your little helper. And then in maybe 3 or 4 years time frame, we're going to start to see machine doing things for you, because they have seeing how you make decisions, they're starting to know how to make those decisions on behalf of you. But those take time. I would say if you're looking at your own industry and your own startup, I would drive for efficiency first, cost savings, efficiency automation, and then then look at making helping them, you know, helping human make decisions.
Then a couple of years from now, probably starting to look at automating that decision and do a for them type of, path. So, we're starting to see business outcome in the efficiency space. Like very clearly. But for the other ones, I think we still need to test the tool.
Cheryl K Goodman
Very well said. And I think it's data drives insight. Insight drives good action. And that's good business. Thank you so much for let's give it up for our panelists okay. They did a fantastic job I, I think we have a few minutes for questions. Right. So yes we do I I'm sure there's tons of questions. So who's got them now.
Cheryl K Goodman
Let's not fight. All right.
Right here I'll come over to you. Here you go.
Unknown
Hello. My name is Kyle Berry. I work at Thermo Fisher Scientific, leading market development for them. We are using AI and I have a question about AI. Maybe it's a little bit more of a theoretical question, but do you think it favors the generalist or the specialist, like early care, career professional, or the more experienced I'm I'm default to so it's definitely the way that it's operate is more for a generalist.
Cheryl K Goodman
I think the challenges to get it more specialized. I'd love to hear the panelists reactions. However, I.
Sheryl Anjanette
Think it depends on the use case. So in for example, in customer support today we are using AI for summarization. It depends on the content of the data that you have. It's if the content is there for the specialists, it can help them do from tier one to escalate to tier two. And just I think it depends on the use case and the data.
So I would definitely look at what they know. Do you have what is the use case. And then what is the problem you're solving and can solve that problem. Tests into it likely. If as you were testing to it, you'll get a sense of, oh, this doesn't solve my problem. So either I have to move here or there.
And so, I think you mentioned that the speed of iteration is important. So I would just go test the shit out of it next slide.
Kimbra Brookstein
And I think we had a question back here. Oh were you going to comment. Sorry sorry sorry.
Yi Ng
Yeah. So basically by default the ones that are available like ChatGPT are general realistic and all that. They can answer your questions. If you really are into some specific vertical, you need to look into something that was trained on the very clean data. Otherwise the data that is going to GPT might have, first of all, been added randomly by somebody on the internet.
Right. And then that system itself is training on that data and producing more data that is based on the wrong input. So it could just better use something that was designed for your industry.
Cheryl K Goodman
Yeah, yeah. And I would just add, actually, I was very close to what I was going to add. But if you think about the large language model up here that we're thinking of like a ChatGPT, it has something like, what is it, 80 billion parameters. And so the, the percentage of times that it's going to what we call hallucinate or not get the right answer just goes up.
And to get that smarter and more fine tuned is exponentially more costly in terms of cost and energy. And so it doesn't make sense. So now the next tier are large language models, but they're smaller. They're more precise. And then under that are the children, the expert systems. And you can have more than one expert system with one of these kind of medium large language models.
So it's fewer parameters. Maybe it's 8 billion instead of 80 billion. But then these expert systems are really expert. That's where in healthcare chemistry you know, the things we're doing we cannot hallucinate. We're, you know, hippies and GDPR compliant. And so you can have multiple expert systems that are speaking to that fine tuned large language model. It doesn't need to be a generalist because you don't need to know, you know, what the distances from the Earth to Mars is, but you need to know what you need to know things.
Kimbra Brookstein
Well said. Okay, we had another hand over here, and I don't want to ignore this side of the room and how much time do we have on our Q&A? A couple months, 8 to 8 minutes. I think we can get four questions. And what do you think? All right, so someone over here had their hand raised.
Donna Cruz
Hi. Donna Cruz, I'm the CEO of Savvy Technologies, and we help clients evaluate, disruptive technologies, like, I, so for the woman, at Intuit, especially, you mentioned that, CSS, right, is an area that's very mature in AI and agree. I'm and you mentioned you were, experimenting with bots. So I'm curious, are you writing your own or are you leveraging, an industry tool that's already developed?
And in either instance, why?
Sheryl Anjanette
The tools out there are pretty good. We use existing language models. We have, GBP 4.0 mini right now. We tried many different ones in the past 3.5, haiku and, different versions. And really, I think it's not the large language model, it's the infrastructure you build around it. And how especially in the customer success space, how we help people answer questions, especially in tax and accounting questions.
You have to have the content. So if you're helping your client first, the first question is do you have the content. Because large language model are generalists and then they can help with some of the generalist language like a language model. Fill in the blank. Right. But do they know tax in the way we know tax the tax rules, the changes, the changes in the tax rules, the accounting rules, the nuances of year after year, the specificity of one business versus another business.
No. So, what, you know, what the profit margins should be, how to think about it as a business and what is the right strategy? None of those exist. And then so when people start asking domain specific questions and or product usage questions, you want to make sure you have the data to help your client. Either fine tune the model or build the right things.
So what we do is we leverage the model we actually don't find to where, leveraging a lot of the rag, a retrieval, capabilities. And so think of it as a library, like you have a library of books and these are your content, right? And you're a librarian, you're coming in and then you need an index card or some sort of an AI system that can tell you if I there's a sentence inside a book, that's going to talk about a story about a girl who does x, y, z, go find it.
And so you're like, okay, now I need to find the book, open the page, find the page, find the paragraph and find the the paragraph of the sentence. And that's the job is not to just regurgitate something or say something that's not untrue. But go find the sources in which the sources exist. And that's the retrieval system that you're building that you could help clients build, and leverage large language model to say it back, like to the client.
Here's the summary, right of, how in this book, this girl found something. And then here's how she explored. And she here's how she prevails. And and then if you want, go to this page, this, book on this shelf, and then you'll find the book. And that's the retrieval system that you need to go built. And not every company has the resources to be able to do that.
So I think a lot of startups are starting to also look at how to help integrate. So integrations not about just integrating with a large language model or fine tuning language model, but building the structure around it so that you can get the information that you need. But first you have to have the content. If you don't have the content, you don't have the library.
It's full of empty books. Right? And then so what am I supposed to retrieve?
So and then the large language models aren't really focusing on that. So this is the opportunity ahead. Yeah.
Sheryl Anjanette
Right. Yeah. And I think like I said it's fill in the blank. It can summarize. It can and generate a summary. So once you find the book then they can help you generate the information, generate a summary. But first you got to find the book.
It's like Dewey Decimal System but better. So all right. Yeah. We're gonna pass.
The mic to Jensen, who is attending.
Virtually.
Jensen
Oh. Hi, everyone.
Yi Ng
Nice to meet you. And also nice to meet you again, because I was a summer intern into a system.
So one of my Question. Is that, as you mentioned, that. So I can. Provide the data. The market data for small business owners. That probably the experts don't have. So, so actually, I think AI is changing the roles of many. Businesses and. Also the roles in the team, that probably previously asked for it don't need to have this data. They don't only need to rely on. Their own experiences, but now I. Probably offer better, information for the customers. So, what I want to ask is that. So what. Do you think that, how. AI is reshaping, the business and also reshaping the AI team? And how can the, the teams that are working on AI and startups working on the AI can remain. Competitive in market. Or, like, be more competitive?
Thank you.
Sheryl Anjanette
One of the things that Suzanne talks about, our CEO talks about is that, is not only changing how we work, it's also changing our business models and that is very true. We're still, of course, figuring it out in this, changing the business models of all of us, our small businesses and large businesses. And, if you think of it today, here's how I think about it.
Today we build products, software, we charge for software. This is how we monetize. And, our experts, our services on top of the software to provide, like, customer support. Right. In the future, we're starting to move to a services model in that, AI is starting to replace the software automation. It's starting to be the, instead of building software, AI is starting to automate what software does for our people.
And therefore, AI is starting to be the kind of the front facing interaction with our, human and what catches them. And just always talks about, our, our customers are never alone. We are always catching them with an expert and therefore expert expert TS is starting to become a huge deal, across the industries, and this services model in which we are starting to think virtual expert is the AI, right?
That's helping to automate some of the software, workflow. And then the human expertise and the human expert are the one we're starting to monetize. So there there is, we're changing the business model. It's changing the way team operate, changing the way we operate as a company. And those things are happening not just at a, I think, big company level, but potentially, small company as well.
So it's basically the vertical ization of SAS is what's happening. I mean, there's a tremendous opportunity. But what a great question. The other panelists want to comment on that. I think we're probably than running up against our timeline.
Cheryl K Goodman
I would just like to add one caveat to all of this is and this is meant to get everyone excited, let fear go out the window and to be excited. If we look at most people's work life today, there's a tremendous amount of burnout, a lack of wellbeing, a lack of work life balance, ace efficiency. It's magic that we know isn't really magic.
Right about that is it's numbers.
It's explained about. But, I think one of the superpowers that we can embrace is that it will give us some of our time back. And I think Irwin Jacobs talked about this at the Startup San Diego event. It will allow us it will give us some of our time back so that we can spend it in more creative ways in our business life.
We can become more innovative. We're not. So our plates aren't overflowing to the point where we don't have time to actually innovate. Being creatively have the relationship with others and with ourselves that lead to a healthy life and well-being.
Yeah, I think that's great. Let's close it out.
Yi Ng
So just from the practical point of view and speeding up the workflows, AI is basically about making things faster. And, what will we'll be observing is actually you have more free time made. In our platform, chemists actually create formulas of, personal care products in the typical, life cycle of that creation would be for six months.
You developing a formula at the end of the six month period, you take that formula to regulatory consultant. And that consultant tells you that some ingredient is wrong, and you need to take it out and you start again. So instead of having that six month cycle in our platform, chemists are actually capable to do this. Like instantaneously.
They create the formula. The system advises them on what ingredients they should be using, and they check regulatory for hundreds of countries and, retail and other agencies, just like in one click. And so that saves months of work. And actually, I think some of our clients were feeling some anxiety first when they were seeing the system, and they were literally asking to, what do we do now?
And, the answer was, you're getting more time and you can be more creative. And so they're now actually creating better lip oils and wash risers and more healthy things in.
Their solving new problems, because they now have the time to do that.
What what I can advise some very good one.
Cheryl K Goodman
Fantastic world. Well, I think that puts us, really right up against our time. I'll just make one final comment. You mentioned Erwin Jacobs. I did interview him last week. He had amazing insights about the future of AI. And I you know, if you look around this town in the Qualcomm zone, everything. And as an innovator, and that, that interview is available on my podcast called Mind the Machine.
So check that out. That drops next week and you can watch that interview. But huge thanks to Holly, Athena, all of the members that we can have this conversation, that we can learn together and learn from each other. So again, give it up for our panelists. They are epic.
Thank you. Thank you Cheryl. You know, when I think about, every one of us, we we took it we made tonight a priority. And, as Kimbra reminded us at the beginning of today's program, we've been doing this for six years. So I'm really excited that there continues to be a devotion and a commitment to this big, big behemoth called I.
Sheryl Anjanette
What does this mean? How do I make sense of this and how do I drive value for myself, my career, and certainly my company? So I commend I just, I obviously am super, super thrilled to continue to learn about the topic and to know that there's so many of these experts that we can call on. But I also want to shine a bright light on all of you.
I think Cheryl, said it. Or maybe it's what I'm talking about. Really glad that you all came here. And making an investment in yourself. And make sure that you're continually bringing that lifelong student mentality, to this awesome, awesome, wild, wild West. The other thing I just want to remind people, as you're thinking about things to take away, like to take away tomorrow into the company, like really thinking about significant takeaways outside of just sort of the big picture that that we got to hear from these expert panelists.
But also, if you're if you're still challenged with what to do with this and you want to take it to the next level, I invite you to continue to ask these questions in your company. Right? This whole this whole notion about disrupting the business model, I mean, that blew me away. It's like, yeah, okay. The we used to have software.
That's what we do it into it. Now the AI is replacing that. So now we have to get in front of that and have a say. So the same thing that into it as doing as the third greatest company in the in the nation.
Be thinking about those things.
To inside your company.
And it's not necessarily that you have the answers that you're continually asking and probing and bringing that level of curiosity into your companies. So that's my first, suggestion or invitation for you to take back into your companies, your team. The second one is I want everybody to know that Sheryl has a company called Find Good Tech. When you are looking for somebody who's an expert who has worked with Qualcomm, who's worked for obviously the co-chair of the startup, we have resources here.
So please don't think you have to leave tonight and go, oh my God, how am I going to implement this? This is why we chew well, this is why we chose this expert panel, because we want people to know that there are resources and there are experts that can help us do this together. And ultimately drive value for our companies.
And obviously, our, our, stakeholder in the audience as well. So I just want to give a big round of applause, for our, viewers that joined us, globally. Thank you so much for your engagement and participation. And again, thank you so much for our speakers. And I think, sure, I'd like to give a book away tonight.
So if someone wants to win a book, sign up for my newsletter. And I'll look up my sign ups tonight. And then I will randomly have my husband pick. Or maybe an I, I'll just pick up a I bought, but I'd like to, I'd like to give away the book so. Or. Tulsi. There you go.
Thank you. Every year.
All these spots, all these hot babes. Thank you everyone. And really enjoy. Enjoyed, learning with all of you tonight.
01:15:15:23 - 01:15:23:16
Holly Smithson
Should. Oh, yeah. We got to get a good group photo. Hey, can I invite everybody?
Comments