MLOps Community
+00:00 GMT
Sign in or Join the community to continue

From A Coding Startup to AI Development in the Enterprise

Posted May 10, 2024 | Views 242
# Coding
# Startup
# Intel
Share
SPEAKER
Ryan Carson
Ryan Carson
Ryan Carson
Senior AI Dev Community Lead @ Intel

Ryan has been a founder, entrepreneur, and CEO for 20 years, successfully building, scaling, and selling three companies. He's passionate about empowering people to become developers and then connecting them together in a global community.

After earning a degree in Computer Science in Colorado, Ryan moved to the UK and worked as a web developer. He then organized global tech conferences, hosting thousands of attendees and influential speakers such as Mark Zuckerberg, the founders of Android, Instagram, and Twitter, among others. His company also produced Twitter’s and Stack Overflow’s developer conferences.

Following that, Ryan started an online Computer Science school. Under his leadership, the team grew to over 100 employees, educating more than 1,000,000 students. During this period, he secured $23 million in venture capital and earned recognition as Entrepreneur of the Year.

Over the last two years Ryan dove deep into AI and LLMs. He built an educational proof-of-concept called maple.coach, which focuses on teaching Sales. The platform is built using technologies like Next.js, TypeScript, gpt-4, and Vercel.

Outside of work, Ryan shares his life with his wife of 20 years and their two teenagers in Connecticut. They enjoy spending their free time sailing and taking walks with their Sheltie, Brinkley.

+ Read More

Ryan has been a founder, entrepreneur, and CEO for 20 years, successfully building, scaling, and selling three companies. He's passionate about empowering people to become developers and then connecting them together in a global community.

After earning a degree in Computer Science in Colorado, Ryan moved to the UK and worked as a web developer. He then organized global tech conferences, hosting thousands of attendees and influential speakers such as Mark Zuckerberg, the founders of Android, Instagram, and Twitter, among others. His company also produced Twitter’s and Stack Overflow’s developer conferences.

Following that, Ryan started an online Computer Science school. Under his leadership, the team grew to over 100 employees, educating more than 1,000,000 students. During this period, he secured $23 million in venture capital and earned recognition as Entrepreneur of the Year.

Over the last two years Ryan dove deep into AI and LLMs. He built an educational proof-of-concept called maple.coach, which focuses on teaching Sales. The platform is built using technologies like Next.js, TypeScript, gpt-4, and Vercel.

Outside of work, Ryan shares his life with his wife of 20 years and their two teenagers in Connecticut. They enjoy spending their free time sailing and taking walks with their Sheltie, Brinkley.

+ Read More
SUMMARY

Ryan shares his professional journey, tracing his transition from building Treehouse to joining Intel. The conversation evolves into a deep dive into Carson's aspiration to democratize access to AI development. Furthermore, he expounds on the exciting prospects of new technology like Gaudi three, a new ASIC for AI workloads. Ryan emphasizes the need for driving competition in compute to lower prices and increase access, underlining the importance of associating individual work with company-based OKRs or KPIs. There is also a reflection on the essentiality of forging quality relationships in professional settings and aligning work with top-level OKRs. Discussion on the potential benefits of AI in constructing and maintaining professional interactions is explored. Touching upon practical applications of AI, they also delve into smaller projects, the possibility of one-person companies, and the role of AI for daily interactions. The episode concludes with an expression of optimism about technological advances shaping the future and an appreciation for the enlightening conversation.

+ Read More
TRANSCRIPT

Join us at our first in-person conference on June 25 all about AI Quality: https://www.aiqualityconference.com/(url)

Ryan Carson 00:00:00: All right. I am Ryan Carson, and I am helping intel build a global community of AI developers. And I love coffee black as often as I can drink it, which is at least like six times a day.

Demetrios 00:00:20: Welcome back to another ML Ops Community podcast. I am your host, Dimitrios. And this episode was very special for me. I got to interview someone who made an impact on my life so long ago, and it turns out they are a listener of the show. And because of Twitter X, we got in touch and we are here talking with Ryan Carson, the one and only Ryan Carson, who created Treehouse, which is how I started Dibble dabbling in the tech world. I saw him on the interview series foundation with Kevin Rose, and I was literally just scrolling through all the different interviews that Kevin Rose did back in the day on his foundation series. It is incredible. He was talking with Ben Horowitz.

Demetrios 00:01:09: He also talked with Biz Stone. And there was another one, so wild how early he was and everything. He was talking with the Coinbase CEO Brian Armstrong. Oh, Malik. And each one of these different episodes that I would watch and I would see when they come out, I would go down the rabbit hole of the person because I felt like, oh, well, if Kevin's interviewing them, then there's probably something worthwhile there and I should check out what these people are doing. So ohm Malik comes on and I started reading Giga. Ohm the first one ever was with Jack Dorsey, and then they did another one with Chad Hurley, the YouTube founder. So I would just go down the rabbit holes like, who are these guys? What do they do? Because I would listen to them for an hour or half hour.

Demetrios 00:01:58: This was the first time that I saw an interview with Elon Musk. They're on the factory floor of Tesla back in like 2012, 2011, and Kevin's asking him about the hyperloop that he put out drawings for. So needless to say, this had a big impact on my life, especially when I was in my twenties living in Spain, playing a lot of guitar. And I got to talk with Ryan, who was on there, and he talked about what he did with Treehouse. And in that time, he was in the thick of it, creating treehouse. He ended up staying with Treehouse until, as he says in this episode, until 2021. He sold it. And he's since recently joined intel because he's so passionate about AI development and empowering people to leverage AI.

Demetrios 00:02:48: This conversation took all kinds of fun and quirky turns and twists, but I will say there were two pieces that I took away from it, one being that he gave a lot of sound advice on if you're just playing with stuff, if you have these skills, there's never been a better time to ship something. The speed at which you can ship a product and a company has never been easier. And that is very cool to see how he has that entrepreneur in his DNA and it was almost contagious. And the second thing was now he's at this big company, he is living the life of an intel employee and he just said something that is said in many ways. The truth is one expressed in many ways. The wise know it in different words that proverb and it's tie yourself to the top level okrs that the company has and figure out how your okrs can affect those. So we've heard it said a million different times. It is find a number that the company cares about and see if you can move the needle on that number.

Demetrios 00:04:08: Let's get into it with Ryan Carson again. This is very special for me and as always, if you enjoyed it, you like it, give us a, like, give us a share, subscribe or just give us some feedback on Spotify because I've been reading all the feedback that comes through. See you on the other side. I have to come out the gates and come clean with everyone and say that this is a bit of a. I had to pinch myself being on here talking to you because twelve years ago I was really into this podcast. I was one of those weird ones that liked podcasts. Twelve years ago I still absolutely devour podcast.

Ryan Carson 00:04:53: Were early as always.

Demetrios 00:04:55: Yep. I was very early. And you were even earlier because you were the guest on the podcast that I was listening to.

Ryan Carson 00:05:02: Yes.

Demetrios 00:05:03: And so it was, it was the foundation series by Kevin Rose and you had been running treehouse, the. It's like a website or a coding learning institution. I watched that episode and I went and I immediately was like, I think I should learn how to code. It seems like, you know, Ryan's making it sound like it's possible.

Ryan Carson 00:05:27: It's a thing, you know, all the.

Demetrios 00:05:29: Mental blocks in my head of like, oh yeah, that's for really smart people. And I don't know if I could do it. Then I saw some of the preview stuff that you had on Treehouse and I just dove in and now I get to actually talk with you. So it's full circle, man.

Ryan Carson 00:05:46: This is amazing. Demetrios. Like, and it's. So you are the reason why I started the company. Folks like you to open that door to say, maybe I could do this. Like maybe I don't need to be a super computer nerd or a super, you know, electrical engineering guy or. Or really good at math. Like, my whole goal is say, you know what? Anybody can do this.

Ryan Carson 00:06:11: And so I got goosebumps when he told me that. And it's. It's so fun to be on the show. I listen to it all the time while I'm working out and sweating. Your guests are strong, are way smarter than me, so I'm a little nervous, but I'm trying to offer value.

Demetrios 00:06:24: Exactly, exactly. I will tell you what the. I remember the thoughts that were going through my head, like that were impeding me from thinking that I could code. And it was almost like this idea because I had, when I was in middle school, some friends that would. They knew how to build websites and so I would watch them. Yeah, exactly. The middle school kids. And this was like, I remember in middle school we were using, like, ask Jeeves.

Demetrios 00:06:51: That was the cool thing that was going on. And we thought it was amazing with the little logo, the Jeeves guy.

Ryan Carson 00:06:56: That was the first.

Demetrios 00:06:58: That was the original ask Jeeves GPT. Somebody should bring that back, actually.

Ryan Carson 00:07:04: That would be classic.

Demetrios 00:07:05: So I saw them doing that and I was like, well, I guess I missed the boat. I can't code. I don't know how to put together websites since middle school. I'm not going to be able to do it. And it just so happened I was in Spain and I had a lot to do. My fingers were sore from playing too much guitar, and I was like, well, let's see if I can actually do this coding stuff.

Ryan Carson 00:07:26: I love it. And ironically, because you are musical, it probably was actually easier than you thought. There's a lot of beautiful connections between math and coding, and I actually think it bleeds into the beauty of the universe that there is this universal language of math that connects everything and it ends up. And then this is why I love AI. I just can't believe you can teach machines to think with math. It's just. It gives me goosebumps.

Demetrios 00:07:57: It's super to abstract it from math to language. Yeah. Wow. That was very special. Yeah, that. Well, another thing that I loved about Treehouse. And then we can. We can move on and I can ask you what you've been up to since then.

Demetrios 00:08:10: But the. The thing about it was that I was always appreciative at how highly produced and, like, how serious you took the online course, because this was. I don't know, this was probably like 2012 around that time.

Ryan Carson 00:08:29: Around 2011. Yep.

Demetrios 00:08:31: 2011. And there weren't people that were like, I'm going to put a lot of time and effort into these videos that I'm putting on the Internet.

Ryan Carson 00:08:41: Yep. Yeah, we. So I want to give massive credit to Nick and Jim, who are our first teachers. I fundamentally believed in video as a way to learn. I just love watching video. It helps me connect all the dots and I. And I struggle with reading, you know, just walls of text. And I thought video would be wonderful, but it needs to be taught by really engaging teachers.

Ryan Carson 00:09:05: And it. We really wanted to hold our students hand and say, we're going to walk you from video to video so that you have a coherent experience. And I tweeted out, hey, is anybody out there that loves front end and back ends and is really good on camera? And Nick and Jim were like us, and they literally had, like, a show that they ran and they were just wonderful. And they wrote the content, they shot the content, they edited the content. That's really how we booted up treehouse. And I just wanted it to be a source of education that you could trust, that you felt was friendly and welcoming. And that was really the way we approached the product for a decade. So, you know, and you could always go to YouTube and get videos for free, you could dig around, but everything was kind of disconnected and there was different qualities.

Ryan Carson 00:09:56: So we wanted just one great place to learn how to code.

Demetrios 00:09:59: Exactly. That value prop still is very strong. Like, can I go to someone who I trust? And I know it's going to be quality every time? Or am I going to spend the time to dig through the Internet and try and find something that is going to take me down the path that I need to go down?

Ryan Carson 00:10:19: Yeah. And the reason why I ended up there was actually, I'll tell you the quick story that leads to it. So it was 1987, and my dad started in 1987.

Demetrios 00:10:30: It's going to be the quick story.

Ryan Carson 00:10:33: It's like a medium story. Okay. I'm just going to be real.

Demetrios 00:10:37: So the year I was born. Exactly. Let's keep going.

Ryan Carson 00:10:40: So I was ten at that point, and my dad brought home an Apple Iie, and I was like, what is that? And. And sort of messed around with it and fell in love with it. And then I went into high school and my calculus teacher grabbed me in the hallway and said, ryan, you should take our computer programming class. And I was like, I don't even know what programming is. What is that, Miss Bruce? She said, it's whether you tell the computer what to do and it does it. And I was like, what? I want to do that. So learned Pascal, and then I ended up in computer science, right? And I was very privileged, like very, like, at the top of the privilege chain, right? Had a dad who brought home a computer. I had a teacher who said I should learn how to code.

Ryan Carson 00:11:25: You know, my parents could afford college. I chose computer science, like, and then, but I didn't get all that, like, at the time, really. And then I walked into my first job interview for a web developer job, and they said, do you know cold fusion? And I was like, what do you mean? And they said, it's a programming language. And I was like, no, I know C. And they said, it doesn't matter, we're going to hire you. Just read a book. And I'm like, what? I thought I got this job because, you know, I could code for you. And it turns out the computer science degree was the thing that got me the job.

Ryan Carson 00:12:06: It really threw me because I thought, wait a minute, the world isn't working like I thought it should work. Like, I thought you'd get a job because you earned the skills to do the job and you worked hard. But it turns out I'm in this secret little club called computer science, and you assume I can do the job and you let me in and I decided the world shouldn't work like that. And I feel like there's millions of people that are never going to figure out what computer science is and get this job. And this job is where you make all the money and you create all the value. And so that moment was when that seed was planted, and that was in 2000, 110 years later, grew into treehouse.

Demetrios 00:12:48: Wow.

Ryan Carson 00:12:49: And I realized everybody should learn how to code at a fraction of the cost of a computer science degree so they can get a job that then will change their life and change their family's life. But it took me ten years to get from there to starting the company so well.

Demetrios 00:13:03: And more topical, the stuff I still remember you saying on that interview, like, we can have a react component, a video up, like a day after a new release of reaction. Tell me a computer science course or degree that does that, it just doesn't happen, right? And so that is, I, I think you can, you can look at it in many different ways. Like, you can go the full, you don't actually need any formal education. You can just get everything you need from the Internet these days, which is great, or you can kind of do hybrid and you can supplement it and still have those experiences of the university or whatever and get that because it is good. Some people still value it when they're looking at a resume or just, yeah, I think those are the two. I was going to say just do nothing. But I don't think you're getting many jobs.

Ryan Carson 00:14:01: That's probably not going to work. I mean, I would say to everybody, now, if you're going to go to university, just go all in and try to get at the huge brands like a Stanford or Harvard or Carnegie Mellon. Like, just go for that, because that is truly valuable. That's only because it's a brand like the computer science you learn. That is probably no better. But then you get connected to the community there, like so that great. If you can do that, awesome. But I went to Colorado State University and no one cares about that.

Ryan Carson 00:14:32: So you might as well avoid all that and just go straight into building. Ship something real, impress people with something that you built with your creativity, and work your way into a job that way. But, you know, everything's changed since Treehouse was acquired in 2021, though. I mean, you know, large language models just mean education now will forever be different. And so that's a whole bag of cats that we could open up at some point if you want.

Demetrios 00:15:00: All right, real quick, some words from our sponsor, Zliz, and then we'll be right back into the program. Are you struggling with building your retrieval? Augmented generation, aka Rag applications? You are not alone. We all know tracking unstructured data conversion and retrieval is no small feat. But there's a solution available now introducing Zliz cloud pipelines from the team behind the open source vector database, Milvis. This robust tool transforms unstructured data into searchable vectors in an easy manner. It's designed for performance. Zliz Cloud platforms pipelines Zliz cloud pipelines also offer scalability and reliability. They simplify your workflow, eliminating the need for complex customizations or infrastructure changes.

Demetrios 00:15:55: Elevate your rag with high quality vector pipelines. Try Zliz cloud pipelines for [email protected]. You can find the link in the description. Let's get back into this podcast. And are you thinking about getting back into the education space? Like, where's your head at right now?

Ryan Carson 00:16:15: I spent a number of months relearning how to code, so I essentially missed it. I missed building, like from the ground up, you know, because I was fortunate to hire engineers and then hire engineering managers and then VP of engineers and ctos, and I got abstracted away from the coal face of actually writing code, and I missed it badly. So after Trios was acquired and had some time to think, llms appeared and I just thought, whoa. And then when I started to ask, I basically paid for chat GPT. And then I started asking questions like, well, how would I write a Python app that could use the OpenAI API? And was like, oh my goodness, this is amazing. This is better than anything we could have built at Treehouse. Then I retaught myself in front to back engineering. I taught myself typescript and I taught myself next js and how to deploy in Vercel, and I shipped an app built on OpenAI's APIs and I was like, what happened? I did that in like a month.

Ryan Carson 00:17:28: And so I think once I realized that, wow, this is everything we hoped and dreamed for at Treehouse, times ten, and it's $20 a month for a chat GPT plus account, wow, it's all different. So I constantly think about education. It's like in my blood. But I ended up realizing, you know, what compute is, is going to be the world's most valuable resource. And there's really only a couple companies that actually can create compute, which will create intelligence. And I realized Intel's one of them. So I was at a point where I was like, you know what? I want to join the company that I think is going to be creating intelligence in the future and be inside of that. And inside of that, I want to empower developers to use AI to build the world's most amazing thing.

Ryan Carson 00:18:19: So I chose to join intel and sort of, I'm doing educational on Dev rel and all sorts of exciting things inside of that framework. So it's just, what a time to be alive.

Demetrios 00:18:33: Yeah.

Ryan Carson 00:18:34: Can you believe it? Like, we're the ones that are going through this. We should have been born it, you know, hundreds of thousands of different points in history. Right. But here we are.

Demetrios 00:18:45: Yep. Or not made it.

Ryan Carson 00:18:47: Or could it?

Demetrios 00:18:48: Yeah, that's another little bit more pessimistic or different spin on it that I was gonna take. But it is good to think about and be appreciative. It brings that gratitude towards it 100%. And we get to see, you know, like I. Yeah, I often think about how my daughters, like, I have two daughters right now and how their world will look. And especially now, you've probably seen just as much as I have on the Internet and out there how there's a lot of companies making plays towards wearables.

Ryan Carson 00:19:24: Yeah.

Demetrios 00:19:25: Right now it seems like not going.

Ryan Carson 00:19:26: So well right now, though, is it?

Demetrios 00:19:29: They've unanimously flopped. Poor folks out there. But. But they're taking some swings, which is cool.

Ryan Carson 00:19:35: It's v one. Give me a break. Like v ten. It'll be in your bibles or in your brain.

Demetrios 00:19:40: So, yeah, yeah, so, so it is fascinating to think about that. And that brings me back to this other piece that I was really reflecting on today about me feeling like we're in the AI phase of, like, the field of dreams. And I say it's like the field of dreams because anybody that's seen that movie knows probably the most famous quote from that movie is that if you build it, they will come. And so it kind of feels like a lot of the stuff that we're doing with models, whether that is talking about, oh, well, we're going to make like this big model, but smaller, and then you can put it on the phone and it's like, great, but what am I going to do with it on the phone? And it's like, well, then you've got agents that are going to be able to book out a flight for you. And I'm like, but I have Google flights. So what do I need these agents for? Like, please. I'm really struggling. I'm trying.

Demetrios 00:20:40: And there's probably somebody out there that's just like, this dude has no vision. What is going on with this guy? He cannot see the future at all. And I, I like to think that, like, I do, but maybe I'm about to get, I'm so deep in it and I might just get like, you know, blindsided by the whole revolution of agents coming and what they're able to do. And I hope so. But a part of me is a little bit like, are we overengineering a problem that nobody has?

Ryan Carson 00:21:11: Right. Are we trying to hammer everything because we think it's all a nail? Yeah, but first of all, let's go back to that movie, and you have to whisper it. If you build it, they will come. And by the way, if anybody hasn't seen field dreams, please watch it. It's like watching Kevin costume. So I think you're absolutely right and absolutely wrong at the same time, so. Absolutely right. There's just a bunch of solutions that everyone is building, including me with Maple, which was just my stab at a fun, quick AI app, is a solution to not really a problem.

Ryan Carson 00:21:51: And a lot of we're all seeing are that. But I want to back up and say that is part of the startup journey. I mean, it is trying to figure out what is the problem that people really have. And that's just the nature of startups. So I think it's natural and good for everybody to be going through that process. That's thing one. And the thing, too, is it's never been so unbelievably cheap to ship something. I mean, it's mind boggling.

Ryan Carson 00:22:24: You can do it yourself. This is a whole nother category of excitement. And obviously, we all heard Sam Altman say, yep, we're going to see the rise of the billion dollar one person company, and we all believe it. But once you experience the fact you can ship a company by yourself, it's mind blowing. That's why I built maple. I was like, I want to proof a concept. I'm going to start an LLC, and I'm going to write the code from front to back, and I'm going to ship it and I'm going to maintain it and I'm going to market it. I am basically a beginner.

Ryan Carson 00:22:59: Dev, let's just be blunt. When you look at my skillset, and I did it, there's that. Where is this all going? I think the thing I'm most excited about is, is. Is a pretty practical use case that I think we're getting close to. So if you look at a limitless AI, which is a pendant, I think this genre is going to change the world. So think about, you know, the. The, you know, the iPhone. Gosh, what are we on now? Is it the 16 that's next?

Demetrios 00:23:30: Yeah, like 15 or 16.

Ryan Carson 00:23:32: I lost track of that fancy pants iPhone that's coming out. You know, it looks like it's going to have a local model. That's really good. Right. You know, I find it fascinating as well that Apple is now open sourcing some of this stuff.

Demetrios 00:23:44: Yeah. What Apple, of all people? That's because they're very known for how they love open source and letting other people tamper with what they've created. Are you.

Ryan Carson 00:23:56: Are you being. Are you being honest? No, it's just shocking. Right? So they really said, so we're going to see an amazing model on your iPhone, right? And now that thing will be able to listen to you if you let it, and it will either fine tune or use rag on everything you say all the time. And then you actually can start to live this augmented life where you're like, hey, you know, when I was talking to my kids today, I feel like they didn't go very well. Cause I have a 16 year old and a 13 year old, two boys. So you have two girls, you said?

Demetrios 00:24:29: Yeah.

Ryan Carson 00:24:30: Yeah. So sometimes we should compare notes. Yeah, well, you know, I'm. I sometimes many days I'm like, I wish I did that better. You know, I had a conversation, I lost my temper or I feel like I wasn't kind enough, or I feel like I wasn't firm enough. So I'd love to say to my phone, you know, just give me some coaching on how I could have done that differently today. Because the amount of tokens that I say today or that I listen to are probably pretty small per day. Very ingestible for rag or fine tuning.

Ryan Carson 00:24:58: So I think that, wow, that was a really long answer. I think thats a cool idea. I want to see.

Demetrios 00:25:03: That is a very cool idea. Well, especially if its funny because last week we had one of the founders of FedMl on here and he was talking about how were not there yet for federated learning on your phones and that type of thing. That still needs, we still need a few iteration cycles, but we're what you're talking about feels like. Yeah. Then that's when we're coming of age and we're getting into that phase where I say, yeah, I, I love Doctor Becky and I want you to coach me in the style of Doctor Becky. You know, like, give me that.

Ryan Carson 00:25:38: Wouldn't that be great?

Demetrios 00:25:39: Yeah, yeah. Tell me how I could be better in that regard. And then we're living. Yeah, but then we're living. Then it's becoming dangerously close to an episode of Black Mirror.

Ryan Carson 00:25:51: Yeah, but that's just what we all like to say because it freaks us out. I'm a technological optimist, so I tend to think, yeah, of course this would go really sideways, but I can't really control that. And what is is going to be what is. So I choose to believe that everybody could benefit from, you know, their own health professional that's customized to you and everything you say and think and do is recorded and that's just kind of where we're going, you know, and so why not see how that could be great instead of terrible?

Demetrios 00:26:26: Yeah, I appreciate that viewpoint, too. Like, let's explore the potential for upside.

Ryan Carson 00:26:36: Yeah.

Demetrios 00:26:36: Not this potential for the. That it needs to be bad all the time.

Ryan Carson 00:26:40: Right? This is the power of everybody listening to this show. You can create, like, you can change the future. Like, that's amazing. Right? And now you can change it for less capital than has ever been possible in human history. And you're alive right now listening to this. That means you have the opportunity. And I'm just so excited about that. When have we ever had a being by our side that was available 100%, that was wanted us to win that had more iq than the average person and more knowledge than 85 of us for $20.

Ryan Carson 00:27:25: So I just can't wait to see what gets built. I am nervous. You hear again, Altman say things like, if you're trying to build rappers around an LM, we're going to steamroll you. Nobody wants to hear that. Maybe that's why the dev rel team and things seem to be weird around OpenAI right now because I was a part of, we actually produced Twitter's one and only dev conference. It's called chirp. And everyone was like, yay, we're going to build on Twitter's APIs. And then at the show they're like, nevermind, we're actually, we're going to cut off all the monetization and we're going to win.

Ryan Carson 00:28:08: And everyone's like, oh, that's a bummer.

Demetrios 00:28:11: Yeah, thanks for getting us here and making us pay for that news.

Ryan Carson 00:28:16: Yeah. And this kind of goes into this whole fun conversation on open source, closed source and why I think all of us are so bullish on what's happening with open source models and meta basically, you know, just like putting billions of dollars into giving us all access to really good foundational models.

Demetrios 00:28:37: Yeah. And setting the tone. Right. Because then you've got Microsoft saying, all right, well, these small models, we'll open source those. And like you said, even Apple, if they're coming out of the woodwork and open source and models, something just feels weird about that.

Ryan Carson 00:28:52: Yeah, it feels weirdly good, I'd say.

Demetrios 00:28:54: Yeah, I like it.

Ryan Carson 00:28:57: So weird.

Demetrios 00:28:58: I'm confused.

Ryan Carson 00:29:01: I love that.

Demetrios 00:29:02: Very true. And you said something earlier where you were like, yeah, I'm a junior dev at best. I'm paraphrasing here, but do you feel like your developer skills are one thing, but then as you mentioned, you basically took everything, you packaged it up, and you were able to ship something. You're a seasoned entrepreneur. So all of that, those other pieces that aren't the code really were playing into, like your ease at being able to shit.

Ryan Carson 00:29:36: Absolutely. Yeah. I'm not afraid, right. To try and be wrong. You know, I have enough, you know, financial sort of understanding and capital understanding to move these needles around. I don't think it's as hard as folks think. If there's anybody listening to this show that hasn't shipped a company, just try it. Use stripes product atlas.

Ryan Carson 00:30:06: Create an LLC with a couple clicks and ship something real. It sounds like I'm being paid by stripe. I'm not. And then stripe does all the billing and all the reporting on churn retention. Like, there's all these things that just didn't exist when we built treehouse that is like literally one click, just try it and it'll probably be dumb and it probably won't work for your first thing, but you got your rep in and then you're like, okay, now I'm in the ring. Like, it's not so scary here. That's all you gotta do is just get over that first one and start trying.

Demetrios 00:30:39: And being out there and putting a price on a product that you create is so valuable too, having to go and sell something. Maybe it's not necessarily like you get people on the phone and you're selling, but saying, hey, this thing has value, and let's see if other people feel that same way.

Ryan Carson 00:30:58: Absolutely. And one thing as well, folks listening probably have wildly different skill sets or specialties. I think there is a tendency to think everybody now needs to be an ML engineer to win. And that's absolutely not the case, I believe. I think you can take your domain expertise, you can teach yourself to be a junior dev, and then you build on amazing infrastructure like Forcell, like an API from either a closed or open source or use grok. You can now take your domain expertise and use very beginner skills to then build a solution on top of that. So I was just talking to my friend who is this amazing musician, and he understands the science of music and he can take that deep, deep, deep, deep knowledge and then think about, okay, how could I use AI to solve problems I have or offer solutions? And I think that's where that kind of thought process should work. It goes back to your question of, like, are people just building a bunch of stuff that's kind of dumb.

Ryan Carson 00:32:04: Yeah, I think they might be if they're not building on their own domain expertise.

Demetrios 00:32:09: Yeah. Some of the coolest stuff that I've seen is not complicated at all. And it's just, it's exactly what you're talking about. It's scratching the itch of recognizing someone has a lot of pain when they do this. And for example, I was talking to this guy who started this at, well, company, I would say called feathery. I don't know if you've seen that, but they help people fill out forms. It's like a typescript, but with like on steroids. Right.

Demetrios 00:32:42: And so you as a company, you can create a form, but then they have these capabilities where. And it's a lot of these really in depth forms. Let's imagine you have a gigantic tax document and you're going to do your taxes. But then there's one form that you need to fill out, and you need to add, like, three different tax documents to that form. And all of the information that you need to fill out in the new form is actually contained in the old form. So they just grab everything, they fill it out for you, they like, pre populate it, and then you get to say, yeah, that looks good.

Ryan Carson 00:33:19: That's great. See that? I think that's a good product. I don't know how they're doing financially or that is a pain point, so why not try to solve it and see if you build a business out of it? Or maybe it's a huge business.

Demetrios 00:33:31: Yeah, yeah. I think they're doing really well from when I talked about, yeah, shout out to them. That's awesome what they're doing. And they came from, you know, the complicated, like, financial institutions. Just imagine the last time that you had to, like, do stuff with your bank, fill out those kind of forms, or if you're a professional investor and you need to do all those types of forms. I think we also had someone on here from Angellist a while back, and it was similar type of thing where they were, they were kind of bringing up these resources that you would need and helping you fill in the gaps of, okay, I'm going to make this investment, but pretty much everything in the investment is, is already buttoned up.

Ryan Carson 00:34:17: Right.

Demetrios 00:34:18: You don't have to have questions about it.

Ryan Carson 00:34:20: Oh, love it. Feathery. Sponsor the show if you're listening.

Demetrios 00:34:23: Yeah, this is not sponsored by them, but I would love it. Yeah, they were, they were awesome. So. But, but, you know, like, one thing, I want to talk. I want to continue along this road of starting stuff. And you being someone who has started multiple different projects and ventures and companies, all different stages of life, how do you feel like people out there can learn from what you've learned?

Ryan Carson 00:34:57: I would say you have a unique advantage in large language models, and I would use it. So, you know, the mental battle, I think, is the hardest part to solve on all this. And I think just having a being that you can bounce things off of as you're thinking about startup, starting something is a huge unlock. So I would say just grab your favorite model and start talking to it about what you're thinking, what you're afraid of, what you're not sure about, and it will progress you. Right. So I would say that's the number one thing because I didn't have that. And so I was. I tended to be a solo founder or solo founder, plus my wife.

Ryan Carson 00:35:45: And, you know, I'm super fortunate to have an amazing wife who've been married 20 years now, and she has really, like, been the supporting through everything I've done. But I was always out in front, kind of alone. And at some point, you just can't talk to your friends anymore about your idea or your parents or your kids or your partner. They just. I don't know, at some point, they get tired of it or they don't have expertise in it, and so you need that. So I would say that's thing one, which you have access to now, which is a major, major unlock. And of course, the LM is going to, you know, not be the perfect thought partner on that, but I think it gets you, you know, a long way through. That's thing one, I think thing two now that you can ship companies so much faster and cheaper, you know, building on things like stripe, on things like Vercel.

Ryan Carson 00:36:40: It's good. Just get it out, see what happens. You know, if you work for a company, do it on your own laptop. Don't do it on your work laptop, everybody. You know, there's these lines you got to draw in your life. Another tip, you should talk to your company. If you work for them and let them know you've got a side project, it'll be terrible if you don't, because then they'll just say, we own that. Thank you very much.

Ryan Carson 00:37:04: It's things to know about that. So a couple of things like that. I think the other thing is this, and I think you've done a very good job of this. So I'll give you credit. Build a community, and in a non egotistical way, put yourself in the middle of that community. Here's why. So, and this is. I didn't know I was executing this playbook, but it turned out to work very well for treehouse.

Ryan Carson 00:37:25: So I started off by doing these fun meetups. It was called by designers, for designers. We would meet at a bar. We would talk about web design. And this is like 2000. I mean, this is early and pre social networks. And so people would show up. We had fun, and, like, we should do another one.

Ryan Carson 00:37:41: So we did one in London, then we did one in Manchester, and then, well, that was cool. A bunch of people came, and then. And then let's make a website and a community and do news. Wow. Okay, well, let's do a conference, and I'll be on stage being the emcee. And what happens is you end up being the connector. Right. And you get to know everybody, and then you can eventually leverage that as needed to tell people what you're doing or ask for help or get connections.

Ryan Carson 00:38:12: And so that's. So you saw me on Kevin's show. The reason why I know Kevin is because I invited him to speak at my conference in San Francisco, and we became friends. And then he was like, well, what's this thing you're doing? And I was like, it's called Treehouse, and it's working. He's like, I wanna invest. I'm like, you do? And then he's like, I'll pull on my friends to invest. And it all came about because I had started this little community in a bar called by designers for designers in 2000. And, you know, so build a community and, and offer value and grind at that and don't expect a return.

Ryan Carson 00:38:53: But you, but you will get a wild return on it eventually if you put in the time and you're doing a fabulous job of this and everybody, you need to go to the AI quality conference. For real. Like, it's going to be a good show. I'm not being paid to say that, but community IRL and then wrapping it, you know, with an online community, a podcast, like, it's a genius playbook. And if you, if you can, you should.

Demetrios 00:39:19: Yeah. And it is. It's funny you mentioned that. First of all, thank you for saying that, because it makes me feel good about where we're at and what I'm doing with my life.

Ryan Carson 00:39:28: Smart.

Demetrios 00:39:29: Doing it well, that, like you're, like you were saying, sometimes you're out and being the person who is at, in the company or the founder of the company. Like, my wife has heard too much of it. My daughter just goes, oh, my God. Well, if you don't have money, just like, go get more. I'm like, brilliant advice. I love that you are hitting. Yeah, like, exactly. Hired and help me hire more people like you.

Demetrios 00:39:56: So. But the thing that I was going to say is starting a community is incredibly hard. Right. And you have to really focus a lot of time and attention and effort into it. And even then, I kind of liken it to catching lightning in a bottle because you can do everything the exact same way that I did it, or you can do everything, like, follow the playbook on how to do it, but sometimes it just doesn't catch.

Ryan Carson 00:40:27: Yeah.

Demetrios 00:40:28: And I know because I've told people that I'm going to help them start their communities, and it's a lot of effort and a lot of time and you need. Yeah, you got to put that work in. And what I was thinking is that if you are someone who already has the idea and the product and you're already like, well, I've got this code and I've got this piece of whatever shipped. You can also go into existing communities or go into existing. Like, I recommend quite a bit. If there are influencers in the space that you're trying to get into, just go hang out in their comments section or in their replies section.

Ryan Carson 00:41:12: 100% put in the time and connect.

Demetrios 00:41:16: Yeah, exactly. Cause that's where it's all happening and that's where you're gonna hear if there are these problems that you think people have. You get to see what language people are using. You get to actually, like, be empathetic to other people. Not just like, if you're building to scratch your own itch, that's one thing. But then when you want to bring it onto the market, then if you are out there adding value, people notice that.

Ryan Carson 00:41:42: Yeah, that's great advice. Gosh, I'm going to, I'm just going to two x that and say, yeah, just do it. Demiro trio says, everybody smart.

Demetrios 00:41:52: Especially, especially the AI quality conference that we.

Ryan Carson 00:41:57: I mean, like, we just got back from an off site with our team at intel. And the value you get from IRL connection, you just can never, you can never replace that with online connection. Online connection is important and good, but yeah, the quality relationships you build is just next level IRL.

Demetrios 00:42:19: And now, speaking of intel, I think that's a perfect segue into, like, you being at intel, first of all. Can you explain a little bit about. I know they launched some cpu's and what is it? It's Gaudi three.

Ryan Carson 00:42:35: Yeah.

Demetrios 00:42:36: Is that what it is? Can you break that down? Because I saw that and it looked super fast. It looked really cool. I want to know more. Thank you. This again, I did not expect to talk to you about this on here. You did not pay me to have you talk about this. I think it's just perfect timing that it came out like a week ago.

Ryan Carson 00:42:56: Yeah. So we talked a little bit about this. You know, why did I join intel? It's a, you know, we have 125,000 employees. It's a huge, huge company. Right. And the reason why is because I believe that compute is going to be the most valuable resource, if it isn't already. And there's very few companies that actually create compute that actually know how to make the semiconductors. Right.

Ryan Carson 00:43:26: Because it is wildly hard to do. Like, I just don't think anybody understands how actually hard it is to make chips. And so there's only a few. And I wanted to join a company that was going to create compute because I believe it's going to literally create intelligence in the future. So it's like, all right, I want to be in the middle of that. And then the thing that I'm deeply passionate about is community and empowering people to become developers and become better developers because I believe that developers are going to continue to change the world. I know everyone's afraid that devs are going to be replaced by Devon or.

Demetrios 00:44:08: Insert another one that didn't age well.

Ryan Carson 00:44:10: Coding your agent source. But I absolutely don't think that's true. I think using it, like knowing how to use tools is going to mean you can make better things, like, not, you should not learn how to use the tool and then pretend you don't need to build anything. So. But that's my rant for the day. So I was like, gosh, if I could be at intel, which is literally creating compute, which is creating intelligence, and I could empower people globally to either become a devs or become better AI devs. Like, what a dream job. And so I joined.

Ryan Carson 00:44:40: I'm like, let's go. I want to help you create a global AI developer community, and I want to empower them. And so that's why I'm there now. And I'm never going to promote our products or other products or hype our own stuff. I think the most important thing is to tell people what's out there. Like, hey, here's the tools available to you. Here's what works, here's what doesn't work. Obviously, everybody wants h 100 so they can train their models, right? And obviously everybody does not want just one company to create computer.

Ryan Carson 00:45:12: The more competitors we have in this space, the better for AI devs, the better for humans. I'll mention it briefly because I really don't want your listeners to think I'm here to talk about intel, but we did release a brand new ASIC. Now, it's not a GPU, which is interesting. It's designed for one workload, which is AI workloads. It's called Gowdy. We just released the latest one called Gaudi three, and it is faster than an H 100. And I'm just really excited that people have more options now to train because I think every company, every country, eventually every family, every city, every village is going to be either training models, fine tuning models, running inference this is the way that intelligence is going to work. We need massive access to compute and the price needs to come down.

Ryan Carson 00:46:09: And I'm super excited that we can be a part of that.

Demetrios 00:46:13: So that vision is also really compelling. The idea of it is going to be just as decentralized as us all having phones. Right. It's all going to be this. Yeah.

Ryan Carson 00:46:27: Right. And so we can't have one organization, you know, creating compute and really dominating. We need many. And there's a lot of, it's exciting to see the competition heat up. I think it's good for humanity.

Demetrios 00:46:41: Yeah. Yeah. Because as you mentioned, it'll drive the price down.

Ryan Carson 00:46:44: Yep.

Demetrios 00:46:46: You've been through the startups, you've been on one side of the table at startups, basically like Boss man, and taken that to selling a startup. Yeah, that's your official title, I think, on LinkedIn and everything. And then you come to a company where there's thousands of people. How do you think about adding value and bringing your best self to that company?

Ryan Carson 00:47:18: Everyone, if you're in a startup or in a company, you're probably, if you're at a good one, you have OKRs or you have KPI's. If you don't, white red flag. Right? Okay, so. Or exciting opportunity. Right. So. So here's the way I would if you're at a company and you're thinking about advancing your career and becoming more valuable and more helpful and winning in general, I would. The first thing you should do is figure out what is the top level OKR for the entire company or a set of okrs, and how does it flow down to me and then how can I impact that number? Right.

Ryan Carson 00:47:59: Because you need to align yourself to the strategy of the CEO and the company. Right. So that's thing one. If you are trying to work really hard and do something, but it's not a part of the company's core OKR is like you're never going to really add value, you're never going to be appreciated, you're never going to get a raise. Like you're never going to get more power. So that should be thing one. And if the company doesn't have clear OKRs or KPI's, then you should be the one to say, I would love to align our work to a top level strategy. There's a framework called okrs.

Ryan Carson 00:48:36: It's pretty simple, let's figure that out. And it's beautiful because it goes from the top down. It's like, okay, if you look twelve months out, what is the one thing that we need to say we got done. And you can sort of imagine it being an announcement. Like, what if there was a press announcement? What would it say? And then you work backwards. Okay, well, if we say that, then what are the five things that need to happen in order to do that? Okay, and then you break each one of those down into, what are the five things that need to do that? And then it ends up being, oh, I need to do x, y, z today so that I can move my number. And there's a bunch of great tools like, I love Asana. I'm not being paid to say that.

Ryan Carson 00:49:14: That Asana has okrs that break down into tasks. And it's just, ux is great.

Demetrios 00:49:19: Wait, I'm just laughing because we gotta, we gotta count how many times we've said we are not getting paid because one of these times we better be getting paid. Yeah, we're not getting paid so much.

Ryan Carson 00:49:32: That's right, everyone, we're trying to offer you value here and tell you about tools that actually work.

Demetrios 00:49:36: Affiliate links in the description, in case anybody's wondering.

Ryan Carson 00:49:40: Oh, my God, I've never done affiliate links. That's one of those things. I'm like, that's a line. Maybe I should feel bad. I don't know. So I think that. I think often technical folks, we think we just need to ship this thing that we care about. But if that thing doesn't align to an OKR, somehow you just can't really advance and you can't really offer value or change the needle.

Ryan Carson 00:50:07: I would say that's thing one. Thing two, this is kind of a very simple thing, is buy a good mic and a good camera, because you're probably remote and, you know, I know we both have good mics and good cameras, but I feel like I'm the 1% of people in meetings that actually sound good and I'm like, makes such a difference. I mean, so this pod mic from rode, I think it's like a $100, right? This stand, what was it? Maybe like $200? Yeah, I don't know. And then I've got a rodecaster pro, which is, I think, like maybe three, $400. So that kind of, you sound amazing. Forever package is maybe $500 and your work will probably pay for it. So just like do that.

Demetrios 00:50:51: It's very true that you do get respected and I could see how being in a big company, you can get more respect, especially if all of your calls are online and you're hardly seeing people.

Ryan Carson 00:51:02: I mean, often intel, I'm often the only person with a good mic in these meetings. And that's just because people are kind of normal and they don't have this stuff and they're like, wow, do you have a podcast? And I go, I do actually. But you don't need a mic like this. You don't have to have a podcast to have a mic like this. You too. And we're not being sponsored by road either, people. But I do feel like, I think most of the conversation is around like modern, cutting edge AI development is happening on x. Maybe I'm in my own bubble, but it feels like it's all happening there and I don't know how.

Ryan Carson 00:51:41: And this is, yeah, one of the things I'm working on is bringing our conversation to X. And so I host a space on X three times a week, but it's not sponsored by intel. But it's like, hey, yeah, we want to be here because for sure it's important to cover the news and be a part of the conversation, you know? So.

Demetrios 00:52:01: Exactly. There was one last thing I wanted to mention on the OKRs piece because I really appreciate that view of understanding the top level OKRs and then how you can affect that. Right. Like how that comes down to me.

Ryan Carson 00:52:19: Right.

Demetrios 00:52:20: And what things can I contribute to those top level or that are going to help those top level okrs? Do you feel like you're always going to be in a position where you understand or you know, what those top level okrs are? Is that another red flag if you don't know what those are, it is a red flag.

Ryan Carson 00:52:38: Everybody in the whole company should know the top level okrs. Absolutely. And so what I love about intel is it's very transparent. I was literally, we were at an event where Pat RCO was speaking and so, you know, and he just pounds this home. Like, these are the top jobs and everybody knows it. And then you flow down from that. Now coordinated that among 125,000 people is not perfect. Right.

Ryan Carson 00:53:03: But, but it's very clear. So, yeah, you should know. And the beautiful thing is, gosh, if you go ask your CEO, she'll, and you say, I want to know what the top level okrs, what is the one thing that you need us to get done in the next twelve months? She'll say, I love that you're asking me that and let's talk about it because that's their whole job is trying to drive the entire team to do that thing. And as soon as you care about it and you work on it, then you, number one, make an impact but number two, you become more valuable the company. And the other thing is, it's like, if you can always be the person that's offering solutions, you know, not problems, people just will like you better. Right. You'll be more successful. So if you're like, hey, I know this is a tough challenge, but here's an idea I have to solve it.

Ryan Carson 00:53:54: You know, I always tell my kids this. Like, people want to work with people they like, you know? Yes. Just, not that you need to be pushover or be nice all the time, but. But if you are helpful, you have integrity.

Demetrios 00:54:08: You're.

Ryan Carson 00:54:08: You're trying to help other people win. Like, generally, people are gonna want to work with you. My kids never listen to me or care what I say, but if they did, they could be more successful in life.

Demetrios 00:54:21: Exactly. If they wouldn't know how good they could have it. Yeah, that's a lot.

Ryan Carson 00:54:27: How old are your girls?

Demetrios 00:54:28: One's five and the other's one.

Ryan Carson 00:54:30: Oh, my gosh. Okay, good. You have plenty of time for them not to listen to you, so.

Demetrios 00:54:36: Exactly. Oh, that's so funny. So. But it's fascinating what you're saying about the relationships piece, because. And just, like, how we like to work with other people that we like. I feel that so much. And we had Sol Rashidi on here about a week, two weeks ago. And one of the things that I asked her was like, oh, yeah.

Demetrios 00:55:00: So you get the, you can go out there and you can be learning about the data foundations, or you can be learning about the newest thing with AI, or you can be, how do you balance what is more important? Because then you can also be doing the, you know, the democracy thing or the relationship building thing and. Or diplomacy. I shouldn't say democracy. Diplomacy thing.

Ryan Carson 00:55:24: Be democratic, if you like.

Demetrios 00:55:26: Also good.

Ryan Carson 00:55:27: Yeah, I think you gotta do both. Right. So. And I think you can be disciplined about it, for instance, by saying, making it to do for yourself once a week. Reach out, build relationships. Right. And then that just reminds you, yeah, I should do that one thing this week. I'm gonna.

Ryan Carson 00:55:42: I'm gonna reach out to that new person. I haven't met the company yet, or I'm gonna make time to connect. And then, you know, over 50 weeks of the working year, you'll have made 50 actions that will advance your networking, your relationship building, your helpfulness. Right. And then I hope it doesn't sound, like cold to say, well, I should make a to do to build my network. But I do think if you don't make a to do, then maybe it doesn't matter to you, right? So that's probably a good thing to do.

Demetrios 00:56:17: Yeah. And it's top of mind, it reminds you, and so maybe you want to. It's just that you forget until you're like lying down in bed at night and you're like, ah, I wanted to write that guy. I forgot. Well, I'll remember tomorrow when I wake up. And then you never remember.

Ryan Carson 00:56:33: Yep. I wish. You know, I feel like I would pay for this. So if someone wants to build it using AI, I think the augmenting and the helping of you building relationships in a non creepy, terrible way, this idea of like, and this is where I really do want my phone to be able to listen to everything I say, because then as I talk to people, it will have context. And then I can say, help me remember to reach back out to Demetrios every once in a while because I care about them. And then when it reminds me, it would remind me we talked about. And there's that natural. I want these relationships to be built, but sometimes I get busy or something happens and I don't.

Ryan Carson 00:57:16: I think that AI would just be a killer for that.

Demetrios 00:57:19: You know what that makes me think is how special the relationships that we actually like, have in our lives are and the ones that we actually do spend time on, make sure they're good. So that's. This has been awesome, man. I really appreciate this, Ryan, and I'm excited for everything that you're doing. I'm really hopeful that we get to meet in person. And thank you for coming on here.

Ryan Carson 00:57:49: It's been an honor and I'm a big fan of yours and the podcast and what you're doing, and I think you're both kind and smart, and I respect that. And I appreciate you having me here. So thank you so much.

+ Read More

Watch More

54:32
Posted May 26, 2021 | Views 578
# Presentation
52:07
Posted Sep 26, 2021 | Views 450
# Yugen.ai
# Machine Learning Systems
# MLOps Startup