Buck Woody is an Applied Data Scientist working on the Azure Data Services team at Microsoft, and uses data and technology to solve business and science problems. With over 39 years of professional and practical experience in computer technology, he is also a popular speaker at conferences around the world; author of over 700 articles and nine books on databases, machine learning, and R, he also sits on various Data Science Boards at two US Universities and specializes in advanced data analysis techniques. He is passionate about mentoring and growing the next generation of data professionals.
You can also learn more about Buck on LinkedIn here.
Here’s a quick summary of key takeaways:
Buck Woody, an expert in technology and AI, shares his journey and insights into the world of artificial intelligence. Buck’s fascination with space and technology during his childhood, combined with his early experiences building computers, laid the foundation for his love of AI. He takes us through the evolution of AI from symbolic AI to data mining, machine learning, and deep learning. Buck highlights the potential of AI in various industries, particularly healthcare and accessibility. He provides valuable advice for those interested in AI, emphasizing the importance of understanding statistics, programming, and continuous learning. Buck’s passion for making complex concepts simple shines through as he encourages individuals to stand out by solving their boss’s problems and staying updated with the latest advancements.
- Buck’s childhood fascination with space and technology inspired his love for AI.
- He built his own computer and started exploring AI in the 1980s during the symbolic AI era.
- The resurgence of AI came with the introduction of data mining and the ability to work with larger datasets.
- Buck explains the concepts of machine learning and deep learning, drawing parallels to human learning processes.
- He highlights the potential of AI in healthcare and improving accessibility for people with disabilities.
- Advice for aspiring AI professionals includes understanding statistics, programming, and domain knowledge.
- Continuous learning and staying updated with the latest developments are crucial for success in the AI field.
- Buck emphasizes the importance of prompt engineering in effectively communicating with AI models.
- Standing out in an AI career involves solving the problems that matter to one’s boss and consistently upgrading skills.
Listen to the podcast here, or find if wherever you get your podcasts:
Ryan Atkinson: [00:00:00] Welcome Buck. I am super pumped to have you on. You’re going to be awesome.
Buck Woody: Yeah. Glad to be here, man.
Really am. Uh, getting people started in their career. I do a lot of mentoring work. In fact, I am a formal mentor. I’ve done it for the University of Washington and here at Microsoft. So this is my wheelhouse.
Ryan Atkinson: And you’re gonna be the perfect person to talk with us about these gigs. Because as you said, we, you have been with Microsoft for 17 years.
You started in June 2006 as a senior technical professional, and now you are a principal. Applied data science in this role. What has changed like the most since you’ve been with Microsoft? It’s a huge company, trillion dollar company. Some of the most prestigious CEOs ever have come through here. So I’m just curious, what has changed the most?
Buck Woody: Yeah, it’s interesting. When I started back then we were transitioning out of the, the Bill Gates era, which was sort of the embrace and extend, they were fighting all these different other companies and really wanted to get a niche in the marketplace and I go all the way back to DOS 1. 0, I’m embarrassed to say.
But they, they really kind of had this scrappy underdog kind of look. And then we moved into the Balmer era, Steve Balmer era. And it was more of an attack on every front. You know, we’re going to attack iPhone. We’re going to attack everybody. It’s a different approach. Uh, not one I would have taken. And we’ve now moved into the Satya Nadella era.
And he’s gone back a little bit. If you think about it to the way Bill Gates used to do things. We’ve gone through some changes recently, as to be a little bit more scrappy, a little bit more lean to attack the market and so on. So I’m really excited to be here. I’ve had five roles at Microsoft.
I’ve done five different jobs, everything from consulting to products and AI and Microsoft research. And now, uh, in, in the data science space.
Ryan Atkinson: That is so cool. And your career extends like well beyond that because like you’ve had over like 39 years of experience in computer technology, which like kind of mind blowing to think is like you said, the iPhone just came out what, like 16 years ago or so rough rough estimate.[00:02:00]
I guess facts like the early buck days, like, where does this love for technology comes from? It’s not like you’re playing around with an Xbox one. So like, where does this love come from?
Buck Woody: Yeah, believe it or not, when I was a little kid, I was very, very poor and I lived on the space coast of Florida and I lived around people who worked at NASA and this was during the space race.
This is the 60s. So I idolized these guys and it was this program that came out. Some of your viewers may have heard of it. It’s called Star Trek and this little show had the guy on it named Mr. Spock. And Mr. Spock was a scientist, and he got to play with all these cool things. And I was just fascinated with the knowledge side and computers were, mainframes back in that day.
My mom used to bring home punch cards from work and I would use those as bookmarks. I still have them. Uh, so that’s kind of where I started the love of technology in general. I went down the electronics and robotics road, which was very primitive. This has been the late seventies, right? But we didn’t have home computers, so I built one.
I got a book, a magazine back in the day about that thick, and it showed you how to take a Zilog processor. And make a breadboard and then hook up this bizarre keyboard, which was this chunky thing that I had mounted on wood and you hooked up to a black and white television, little tiny thing and the storage, the memory, well, I had one K of memory that was external to the chip and the storage for the programs you would write which were in assembly.
I believe it or not was on a tape recorder. So cool. So that’s kind of where I got started and just absolutely fell in love with it and got into the AI space fairly quickly after that. During the eighties, I worked with a language called prologue, which dealt with logical constructs the different logical things you need inside AI.
Ryan Atkinson: Interesting. So you got started with AI in like the 80s, which is like, take us to that time. Like, what was AI like in the 80s?
Buck Woody: Yeah, so this is the era of what was called [00:04:00] symbolic AI. So in 1956, there was a big get together where some folks got together and said, Hey, you can reduce human tasks.
Down to sort of a program, if you will to where it’s, it’s fairly ritualistic and maybe you could have a machine do it. And this was the first term of sort of artificial intelligence, which was built primarily on sort of linear algebra. Some of the constructs there a lot on predictive statistics, some calculus to calculate long loss functions and so on.
So it was very early on. So by the seventies, prologue had kind of started making its appearance and in the eighties we had what they now call, which I’m not really sure I like the term, but it’s good old fashioned. A. I interested about being called old fashioned, right? But They had these programs, basically some fuzzy math, fuzzy logic was a way we would do things.
This was all based on symbolic AI and expert systems, which you can think of as a bunch of nested if then statements. It’s a bit more than that, but that’s what it really did. Well, of course, it over promised, as all tech does, and it under delivered, as all tech does. And we went into something called the AI winter.
To where everybody’s like, this is not real and so on. And I was heavily based in databases anyway. So I just stayed with databases, but I kept my hand in
Ryan Atkinson: AI. Interesting. And then can you talk to us a little bit more? So how did we come out of like this AI winter type of thing? Was there a technology that really prompted us like out of it?
Buck Woody: Yeah, there was in the nineties, late eighties, nineties, we started to get into something called data mining. Mm. And data mining did inferential and predictive statistics. So, being able to predict a probability and so on. The, the sort of magic sauce that was there was, it could work with larger sets of [00:06:00] data.
So data mining was an interesting thing, but it did not incorporate some of the constructs, uh, of sort of ai. After data mining became a thing and data mining is a thing and it’s still used if you’re doing a clustering algorithm that was back in the in the data mining days. Interesting. , but basically data mining and AI got together and they had a baby and that was the new sort of we came out of that winter because you could actually prescriptive Analytics, which is not just here’s what will happen, but here’s what you should do because of what will happen.
That’s a new world. And the other sort of big thing was Hadoop and larger sets of data, relaxing the acid compliancy of databases and you could walk across much larger sets of data. This opens something called machine learning, where you could, um, I always tell people you, you’re a perfect example.
So Ryan, when you learn to read in school, how’d you do that? You went to school, little Ryan went to school and the first day, what did they do when they were teaching you to read? What did you?
Ryan Atkinson: I feel like I just opened up a book and like read .
Buck Woody: They gave you really simple books. And they said, that picture is an R, and that picture is an E, and that picture is a D, and that means the word red.
Here’s red. So they, they gave you an example of something, and then they gave you a trained example, which was, that’s red. That is red. This down here is red. And you began to see it. Then you saw it in a different font. It was italics or it was bold. But your mind went, that’s probably is this red as well?
And they went, yes. And so that’s a positive classification. And you said, is this red? And they said, no, that’s blue. And you’re like, no, it looks weird. So you got to learn that way. And then they gave you another book. And they gave you another book and they gave you, and how many books did you read throughout your education?
Do you think, take a guess?
Ryan Atkinson: Oh gosh. I, I could even know from like [00:08:00] first grade to graduating. Yeah. A lot, probably like over 300, maybe. I don’t know. Yeah. I don’t know.
Buck Woody: It could be in the thousands. Right. And um, the interesting thing though, is you don’t currently have all those books. Do you? No, definitely not.
But you do know how to read. So how is that possible? Well, what happened is through reading all those books and getting the positive and the negative classifications, your brain now has a model. When you look at the little pictures of letters on a page, your brain doesn’t see the pictures. It’s got a model that it interprets them with and you break it down into pieces, trigrams and engrams.
And your brain goes, that’s the word red. And I got it cause I know how to do that. You’ve got a model of how to read. That’s exactly. The way not the process, but the way that a machine learning model works, we have, we have a couple of those. One is the supervised where we take a bunch of data and we say to the computer, see what those things are like with various pieces of math.
And the math says this is like that. I predict this is the stats part that this is like this. And I have a loss function that says I’m pretty close to that real answer. And that’s from calculus. And I, I have a state vector model, and that’s from algebra, and I put all these maths together, and it looks like the computer’s learning.
It’s not really learning, but what it’s doing is making these predictions that these words belong to red, and these words belong to blue, and it creates… a model just like what we have in our head. The other thing we have is sometimes we don’t know something. Yeah. Like we see a group of things and if I were to give you a whole bunch of things you’ve never seen and I were to tell you, Ryan put these things into three groups and you’re like, groups of what?
And I’m like, I don’t know, three groups. You might say, okay, well, let me put these together. And no, that’s not quite right. I guess there’s three colors in this group. Let me put it together by [00:10:00] color. So you would create this clustering algorithm in your head by taking salient points. That we use, we call those centroids inside data science and you would separate them out.
And if I were to ask you, well, why are they in those groups? Well, these are mostly red and those are mostly blue and those are mostly green. I figured that was the most salient thing to do and you’re not right or wrong. Yeah. And if I were to say. Put them into two groups. Then you’d be like, Oh, now I’m going to do them by size because now the colors don’t matter.
And so interesting. So you get the general idea. So those two things, then we got our data mining from our fuzzy logic and so on. And then we’ve got our big data and then we’ve got machine learning. Now we can do something called deep learning and the deep learning simply takes those same things, but it turns them into layers.
This is very similar to how your eye sort of works. You have broads and cones in your eye and one does color and one does edges and so on. Yeah. And it passes through those things and then it hits your brain, another model and says, that’s a dog or whatever. So deep learning is similar to that. And of course now.
We’re in the age where we’ve taken deep learning and even push that further into large language models.
Ryan Atkinson: I think just, I don’t even come from like a data science background and just hearing this type of stuff, like just energize me because I think it’s so cool and you can apply it to like so many different use cases.
And I’m curious, like, what are like some, like your favorite usage cases for like a deep learning model, like industry wise or? Yeah. Yeah.
Buck Woody: So our brains are arranged in what are called narrow intelligences, your eyes do a thing, your hearing does a thing, different parts of your brain do a thing, and then the corpus callosum between the halves of our brain shuttle that information, kind of like a nerve center or the corridor that gets the data all around so that when you see something, it puts it together with the language of your brain and says, [00:12:00] that’s a dog.
People who have epilepsy, where they’ve had to actually cut the corpus callosum. If they cover one eye, this eye actually transmits to this side of the brain, and this eye goes to this side. And language is on one of them, and vision interpretation is on the other. If they cut that, they can cover an eye, show them a dog, and they know, they see the dog.
They can describe it, and they’ll say, what is that? And they don’t know. They can cover the other eye and look at the dog and say, what is that? They’re like, I have no idea. Then they take the picture away and they say, write down the first thing that comes into your head, dog. And so they saw it, but they weren’t able to interpret it because the Colossum’s not together.
But if they have both eyes, they can do it because they’re each getting a different transmission. It’s kind of an interesting way of thinking about that. And if we think about the way. Artificial intelligence to your question. What’s the most interesting? I think the most interesting and what’s coming in the future is sort of the joining we call these things narrow AI when they only do one thing.
Like if they can recognize a red light, that’s a narrow AI. It’s a vision problem. We also have audio language models that can do things with audio. We can also do things with text and so on. Putting those all together, creating sort of our Generic. I’m very nervous when we start talking about A. I like the brain because it’s not, but In a similar fashion, um, putting those things together and coming up with a Corpus Colossum will be pretty amazing when we have something that can see a house and then can use a large language model to describe the house to you, uh, imagine that you’re sight impaired and you’re walking down the street and something is telling you 30 feet to your right.
There is a blue house that is two story and so on. There are birds singing as I mean , imagine these kinds of narrow AIs coming together to form a more complete general AI.
Ryan Atkinson: I think what excites me so much about AI is basically like the usage case you just said. [00:14:00] Like, say like, if Just like talking about like products that you can make out of this would be like a deaf person had that in their ear and like, here’s like a house coming up like lights or roads coming up ahead of you with five cars here.
That’s what I think is so exciting about AI, because when I think about AI, like, correct me if I’m wrong in this, I just think for like, the health, like, I’m a type 1 diabetic. So I just think there’s so many usage cases for like, diabetics, people that are. Deaf or blind or anything like that. Um, so that’s what really jazzes me up about AI.
I think it’s another, a new world that will be unlocked once AI is.
Buck Woody: A friend of mine, Hank Bowman, who’s in the Netherlands, he works at Microsoft and, and AI he and I did a conference last year together and on stage, what we did live and on stage. We took a grocery store shelf, uh, and we brought some cans of things and we put them on the little shelf.
And then he walked up with his phone and we wrote the code on stage to where, as he walked by, it was explaining coke, 16 ounce, a dollar 39 for six pack. Here’s the nutritional information for somebody that might be sight impaired, could go shopping alone, but then see the shells, just see the shells, just like you and I.
Not only that, we trained it in his voice, so he was talking to himself, um, to where it was his own voice saying, this is what this is, and so on. That, to me, AI is a tool. Computing is a tool. Everything is a tool. So I’m not really impressed with the tool. I’m impressed with what the tool can do. I don’t like the hammer.
I like the house that the hammer builds, if that makes sense.
Ryan Atkinson: That is like super fascinating and like anyone listening has to be like super jazzed up about like the future of AI. Like I, I honestly having this conversation makes me want to like pivot to AI in a way because I think it is so cool that like the usage cases for it.
But I do want to talk about someone that does want to pivot into AI or just get into the field or machine learning. Let’s talk about like a traditional path first and let’s talk about like a non traditional path and ways they can do
Buck Woody: that. Sure. I would say a couple things. First of all, I would pivot on, [00:16:00] well, first of all, you need to figure out what you’re good at.
I’m not talking about skills. Like I can write Python. Who cares? That’s a skill. You can pick it up. You can put it down. It’s like riding a bicycle or tying your shoe. It’s not, That impressive to have a skill, so I would not base. Okay, I need I need a Python job. No, you need to solve problems and you happen to use programming to do it.
And maybe Python or maybe you learn Java or whatever, right? So, I worry less about that than what you’re good at. And, maybe you’re somebody , and what I always bring to the table is I’m actually quite simple. I, I, it’s hard for me to learn. It’s very hard for me to learn math. I was not good at it and I do it all day now.
But it was hard for me to do. But here’s what I do. Here’s my strength. I make hard things simple. I do that because once I understand it, and I’m simple, I can re explain it to you. So when I was first learning AI, I had all the brainiacs from the planet Tron that were telling me all the math and stuff, and I’m like, So how does, how does a machine learn?
And they kept explaining all of the gradient loss functions, and I’m like, I, I don’t know what you’re saying right now. And so finally I said, is it like how I learned to read? What are you talking about? Well, I had a bunch of books and I read that. That’s kind of training, right? Well, yeah, I guess. And then now I know how to read and I know that’s the model, right?
Well, you could say that. I mean, can I say that? You can, you could say that. Okay. So then when people ask me, what’s AI? I’m like, you know how you learn to read? It’s like that, that’s it. And so I just make hard things simple because I’m simple. So that’s my strength. Interesting. And when you’re getting ready to get into any career, figure out what you’re good at.
I’m good at following problems through, I’m good at detail, I’m good at broad strokes, I’m good at talking, I’m good at listening. Figure out what you’re good at, and then use those [00:18:00] strengths to do the job you want to do. So, if you’re interested in getting AI, there’s probably two forks that you need to figure out.
The first is, do you want to use AI, or do you want to create AI? So, if we’re talking about using AI, there’s a bunch of ways to do that. Um, you could be someone who teaches seniors. How to use A. I. To not be scammed by telemarketers or whatever that could be using all the way out that or, we’ve put it Microsoft A.
I. Inside word documents and power points and all kind. You could teach people or use A. I. In your Excel spreadsheet to be smarter than all the other Excel spreadsheet users. There’s also another part of that. I am an applied data scientist, so And what that means is I don’t create a I I have worked in that area, but that’s not what I do.
There are way smarter people that do that than me. But I do work in taking things like hooking up. I wrote a blog post yesterday for Microsoft on how to use chat GPT inside a database. So in other words, you could select your product name from a table like you do in a database, but then tell chat GPT, write me some ad copy for my 13 inch bicycle.
And it came back with ad copy. Now I don’t need to go directly to an ad person. I just need to edit what it comes back with. And I’ve got my ad copy done for me. So I teach people how to do that. Now I happen to use code to do that. But I didn’t use code to write AI. It was to use AI. So that’s the first big area.
To do that, what I would suggest, if you’re wanting to get into that area, is just immerse yourself in YouTubes and build conference from Microsoft and Ignite and all these other conferences from whoever and just immerse yourself in whatever’s free and read, read, read, read, read. It’s the most important skill you have is to be able to read.
If you want to go into the, you know, I want to build the AI [00:20:00] and you’re going to need the math, you’re going to need linear algebra, calculus, and statistics, and both descriptive and inferential statistics, it’s just the way it works, no way around that, because you’ll never do this correctly, even if the math is done for you, you can load up a library in Python and say, give me the logistic regression, but you need to know what logistic regression even means, should you be using it, The danger inside machine learning is that you always get an answer.
If I write something incorrectly in Transact SQL or in Python or C Sharp, it fails. I get an error. But if I average four numbers, I will get a number back. Every. Single. Time. Now, what does the average mean? No idea. That’s why you have to understand statistics. Should you be averaging that? So, where do you live, Ryan?
Where are you at?
Ryan Atkinson: I’m in Austin.
Buck Woody: Okay, you’re in Texas. I’m in Tampa, Florida. And we have Tampa Bay. And if you take the way that the bay goes, it goes like this, right? It’s a curve, and there’s a shore, and then there’s a deep end. If you were to take the average of water depth, it’d probably be about four feet.
Which includes the one inch here and the one inch over there. But it’s probably 30 or 40 feet deep in the middle. So if I told you, oh, it’s only four feet deep, you can walk across. That’s, that’s wrong, right? Right? So you have to know when to use a certain statistic or math algorithm and when not. So that’s number one.
The next thing you do need to know how to program. And you also need to have what’s called domain knowledge. So you need to know what area you’re doing. Like if you’re writing for engineering, you need to know that. If you’re writing for airplanes, you need to know what airplanes do and so on. You can’t just make stuff up.
So it’s a far heavier path to, to create AI than it is to use AI.
Ryan Atkinson: Yeah, when you’re talking about this, I was like, Oh, I’m glad you broke this out into two uses cases. I’m definitely more the usage [00:22:00] case, not the statistics linear that just that that terminology just goes to over my head. Yeah, beautiful that you broke that out into.
Buck Woody: So it’s, it’s you still should I agree. I’m always curious. So I I, I was literally, before we got on the call, I was fighting a particular loss function that I don’t really understand well. And my friend who does build a AI said, why are you, why are you bothering with that? That’s done by the library.
Stop it. Yeah. And I’m like, I don wanna know why it does that. So, but the neat thing is I don’t have to, I can, I can use it
Ryan Atkinson: right. Yeah, I’m curious. So let’s talk a little bit more about like this applied, like data science, like basically what you’re in, like, what are some of the projects you said, like YouTube, read, read, read, but like, are there any projects like someone like me, like, Oh, hey, that sounds awesome.
I want to do that. Like, what are some projects that like I should work on?
Buck Woody: So I, I actually start with, with exactly what you said, what, what are you trying to solve? So you go find a problem and, and grab it, grab the problem and then break the problem down before you ever, who cares what the tech is. Who cares what the tech is?
I had a grocery store exec one time that was came to me and he said, I’m sitting in this big meeting and he’s got all these execs there and he’s CTO and all that. And he said, we need to get AI in our stores. And I’m like, you do? And he said, yeah. And I said, what kind of, you want a large box of AI or a medium?
I may have a small box in the car. And he’s like, what? And I’m like, I don’t. Who cares what, what are you trying to do? Well, as we talked through the problem, it turned out all he needed was a regression, which you can do inside Excel. We did it on his laptop while he was in the meeting, which didn’t make my salesperson very happy because they didn’t buy any AI because they already had Excel, but they came back to us later.
But the point was find out what the problem is. That’s where you start to answer your question, find out what problem you’d like to solve. And it could be something like, why are people getting diabetes? And that could be, you may want to know why, you may want to know how to prevent diabetes, you may want to know how to [00:24:00] better treat diabetes.
So you have to get your problem really small, like, How do I stop type 2 diabetes in African American infants under the age of 3? That’s a real problem you could attack. Um, that might be a place to start. Then you would look at what’s been tried already. Because you don’t want to go down a path somebody’s already gone.
Yeah. And then say, what tools do I have that could help do this? And let’s say you could do some principal component analysis, PCA, to find out what’s common and what are the main reasons? What are the coexisting features that happen? When african american infants under the age of three get diabetes, oh, they, they come from this area or they have these kinds of foods or this kind of water from that county or whatever it is.
Now you can start doing it. So I would, I would recommend finding a problem that interests you for sure, because it’s going to get hard. You’re going to have to read and you have to do math and you’re not going to want to, but if you like the problem enough, you’ll solve it.
Ryan Atkinson: That’s interesting. And then how do you get like act as like, is there a piece?
Someone’s like, listen, there’s like, yes, like I projects problems that I want to work on. Like, how do I access all this data though? Like, is there a central place to do that? Or is it a lot like decentralized?
Buck Woody: I mean, it’s very decentralized, and that’s what we’re still good at, and AI is not yet. We’re good at synthesis.
We’re good at grabbing things from disparate places. If you’re using Edge you can use the Bing chat feature, which is GPT 4, or you can use Google’s BARD, and it is fascinating the kind of things you can do. I would highly recommend to anybody who’s listening to, to learn prompt engineering. It’s called some people go overboard with it.
Some people don’t believe in it at all. I think there’s a we have to learn how to talk to the machine right now. It’s not, it’ll be good enough someday where we won’t have to do that. But at some point, you need to know how to talk to the machine and it. It can give you some dramatic answers. You can find prompt guides [00:26:00] everywhere for free.
The one I like is from OpenAI. They have a full free course. If you type in prompt engineering, OpenAI, it’ll come back. And again, you’ve got GPT 4 if you’re using the Edge browser and you’ve got BARD. If you use a different model than OpenAI does, very similar concepts though. That’s really
That is so cool. And yeah, it jazzes me up having these types of conversations. I think there’s so many usage cases. Yeah, we are winding down on time here. We got a few more minutes. I just want to ask like a few more questions for you. I’m kind of jumping into let’s just say like someone works on a project, they get into AI, they land a job at like a company like Microsoft.
And they’re like, I’m ready to do really, really, really good at this job. I mean, how would you advise them to like stand out in a role like similar to yours early in their career?
Buck Woody: I think there’s, uh, two things that I’ve always found that have helped me stand out, regardless if it’s Microsoft or a tiny company or wherever it happens to be, um, the, the number one way you stand out is find out what problems your boss has and make sure you were solving those.
If they’re like, we don’t have enough revenue, figure out how your part is to help them make revenue. So, if she says to you, we’re having a problem retaining people, then figure out how to retain people and so on. Always work on something your boss cares about. That’s number one. And number two is, is.
Rather constant skills updating. So the first, I get up at five in the morning or so, and my first hour at work, I go to work about 6. 30 in the morning, um, my first hour is nothing more than reading RSS feeds on what’s out now. I’ve really refined it over the, the edge, and if I find something, in fact, behind you on my browser.
Is a Delta Lake thing from Databricks is the name of the company. I’m reading what they say about a Delta Lake. I know what a Delta Lake is and so on, but I’m like, do I know everything there is to know about a Delta Lake? So this morning I was reading up on what’s a Delta Lake. So you just…
[00:28:00] It’s just constant and make your own self syllabus. You’re, you’re never out of school. I’ve been out of college. I’ve taught college for eight years and I’m still learning every day. Something new comes out.
Ryan Atkinson: That’s amazing. And let’s end it with that book because that is a great, great answer. Um, so I want to thank you so, so much for coming on.
You were phenomenal. So thank you so, so much for being here.
Buck Woody: Thank you for having me and happy to pop on anytime and, and chat about anything.