Breaking No-Code Barriers: The Simplicity of Lightyear

On this episode of Behind the Bots we interview Mark Hendriks, co-founder of Lightyear. Lightyear is an integration platform that enables developers to connect apps and services more easily using code instead of No-Code tools like Zapier. 

- Mark talks about how Lightyear leverages AI like GitHub Copilot to help generate integrations faster and even have AI write code based on documentation provided.

- After being in private beta for over a year, Lightyear has now launched publicly so anyone can sign up and start creating workflows by connecting their favorite apps and services. 

- Mark shares his vision for where Lightyear is headed long-term, which includes being able to describe an integration in plain language and having AI generate the necessary code automatically.

- Hendriks discusses the origins of Lightyear and how he and his co-founder came together over 7 years ago to eventually start building the tool.

- Mark provides his perspective on where AI is heading generally and how he sees it evolving when it comes to assisting with development and design work.


LIGHTYEAR

https://www.runlightyear.com/
https://twitter.com/runlightyear
https://twitter.com/workbymark


FRY-AI.COM

https://www.fry-ai.com/subscribe
https://twitter.com/lazukars
https://twitter.com/thefryai

Creators & Guests

Host

Ryan Lazuka

The lighthearted Artificial intelligence Journalist. Building the easiest to read AI Email Newsletter Daily Twitter Threads about AI

What is Behind The Bots?

Join us as we delve into the fascinating world of Artificial Intelligence (AI) by interviewing the brightest minds and exploring cutting-edge projects. From innovative ideas to groundbreaking individuals, we're here to uncover the latest developments and thought-provoking discussions in the AI space.

Mark Hendriks: So I live in Spain, I'm a founder of Lightyear, which is an integration platform. And this is all written in code. So one of the things I got into is, of course, having open AI on my platform. So writing code using the open API.

And then the other part would be like, as a designer, using AI to kind of like collaborate or get some initial ideas to to create the final design, for instance.

Ryan Lazuka: Cool. Yeah. And then just to clarify for everybody out there, if you want to go check it out, while you're watching this, it's runlightyear.com. So is it just you working on the project? Do you have a team with your team look like? How big is it? How did you get them together? All that stuff.

Mark Hendriks: It's still early stage. I co-found this with Eric. Eric is the real engineer behind all this. I can code, but not at the same level. And then we have a couple of contractors for testing and quality. So it's very early stage. We're actually going like, today is the first day people can sign up without going through a waitlist. So that's kind of the stage where we had a couple of people using it, got some good feedback. And now we're opening it up for everyone else.

Ryan Lazuka: Very nice. So can you just tell us a little bit like for the layman, what is the project and how does it work?

Mark Hendriks: Okay. So basically what we have created is like, there are a lot of like no codes tools out there where you can try to get different services like Slack or open AI or Google sheets. And what we realize is that you can do a lot of stuff with this like through code. If you like can program, or have some experience, sometimes code might be like a lot easier and more powerful to do it. So that's kind of what sort of the main idea behind it is. And then what we found is like with the whole AI revolution, if you will, it's all like based on text or code. So it's actually has become like a pretty big benefit to us also in like creating the platform.

Ryan Lazuka: Very nice. So who are you looking to use this project? Who the audience is? Yeah, yeah, who's the target for your project?

Mark Hendriks: The target would be people that have experienced programming and have gone through the whole sort of painful integration exercise, either they tried to do something on their own without any tools, or they felt somewhat limited by using no code tools. That would say this talk sort of the prime candidates to use our platform.

Ryan Lazuka: And is that kind of where you got the idea to? Are you a code of yourself? Some of the inspirations and then yes, yeah.

Mark Hendriks: Yeah, absolutely. There were like, you know, when you sort of start out with no code and it can work like if you do something very simple, you're not technical. It's great. Like you can do a lot of things you can do. But sometimes you want to do like more programry stuff where it's like, what is easy in code can be very hard in no code. That was like part of the frustration for like, okay, this can be so much easier if you can just write six lines of code instead of trying to like piece it together in this like UI or visual builder in the fix.

Ryan Lazuka: Yeah, one of the things I just ran into like for actually for these podcasts and I do some YouTube videos as well is I get the we have an editor and she gives us the edited video and then I transcribe it and get a summary of that video via chat.gbt and you can do that on Zapier or like make.com by Zapier. And just to do that was like really hard. You have a user interface. You think it'd be easy, but like how many options to you, so many options to pick, so many things you got to tweak.

Like just doing that three set process is very hard even though it's supposed to be easier. So I think like marks this light year can help with that if you have a little bit of programming experience.

Mark Hendriks: Yeah, and we do like you can write code, but we like made it like a lot easier. A lot of like sort of called like abstract functions or like simple things like our new Slack message around the following code. Like you don't have to go through the entire Slack API design. Like we make it pretty approachable.

Ryan Lazuka: Is it possible for someone to use chat.gbt to write the code for them for your tool as well?

Mark Hendriks: You think? Well, even better we use co-pilot to write the product or the integrations. So what we have done, like what we, the basic setup is like we have like automated a lot of things of course, but we just take the documentation from a company you want to integrate with, let's say Slack.

We just take all the documentation, throw it in a code editor and then format it in a certain way where co-pilot starts all the completing the integration for us, like the actual code, kind of looking at how we want to do things. So it does like all the boring stuff you don't want to do, basically. And then you still have the decisions on what to include and how you want to do this, but it saves like a lot of time and it keeps getting better. We keep getting better and this becomes a somewhat automated process. So, but you should take like super long, we can do like new tools in like one or two days, including like documentation on the website. So that's, I think like it's like the new GPT-4, like it seems to like accelerate for us. So that's like one of the reasons we are all going for code, basically.

Ryan Lazuka: That's awesome because like some of the, some of your competitors, not really your competitors, you guys have your own niche, but like the big boys make and Zapier, they've got like hundreds and thousands of integrations, like someone had to do all that, like integration work. Yeah. Back end. So.

Mark Hendriks: Right. And it's the same code because you also use code to write it. So it's not like this whole visual UI you have to build, right? There's nothing to design. It's just the same type script or JavaScript that we're using. So I think that's, that makes us like a little, little faster. I mean, if we don't have the amount of Zapier integrations, like, it's like 5000, but we'll get there to like a good amount pretty, pretty fast now.

Ryan Lazuka: Cool. Yeah. That's good. That's gotta be a gigantic help to have that be written for you. That's awesome.

Mark Hendriks: But it's also fun, like having someone to help you. It's not like another human, but it kind of feels like a collaboration where someone does the suggestions and you like sort of pleasantly surprised, like, oh, you got this, like right from the first start, like, okay, it's sort of fun in a way, aside from being sort of faster. And it gets around totally sometimes, right? It suggests all kinds of nonsense. You're like, okay, but it's generally, generally it has been pretty helpful.

Ryan Lazuka: Now this co-pilot, I haven't used it yet, even though I should, is it, does it have any kind of personality to it, like at all, or is it just spit out the code and that's it?

Mark Hendriks: Well, as I understand it, it also trains on your behavior, and it can look into like the code directory if you want to. And I think I like some plugins even, but basically it will like kind of learn from the way you write code, and it will, of course, recognize the language you're writing in. So there is, there's like, and I think there's like, there's been like this new version announced for like quite a while.

It's kind of hard to tell when they upgrade the products, like there's not like, I'm not sure which GPT version I am currently, but it's like improving every, you know, every month or so.

Ryan Lazuka: Mark, can you tell us a little bit just more about your background? You said you're overseas, where did you grow up and how did you get integrated into AI?

Mark Hendriks: The AI is only like a year old, feels like, feels longer, right? So I'm Dutch, I grew up in the Netherlands. Like, worked with like, sort of with computers early on, and my parents weren't like technical at all, I had like a very kind uncle that always helped me get unstuck when I brought the computer, he came over to fix it, so I like to like, try to build things with computers from an early age. And I'm like primarily a designer. So I throughout the years, I learned myself how to code in order to like, make my own games or apps, just enough to kind of make things like a whole product.

So that's what I've been doing, like for a very long time, like work with different startups, start a lot of companies, some went well, lots of things you never heard from, of course, as these things work, as this works out. And then like four years ago, my wife, with my wife and son, we moved to Spain, because we wanted to live a little bit more outdoors, which is quite hard in the Netherlands, because there's not a whole lot of nature out there. Plus the climate is better here. So that's what we did. So we got like a house on a nice piece of land, more outdoors, and getting to enjoy the sun as well.

Ryan Lazuka: So you said you got to, you dabbled in making some games, what were some of the games that you made?

Mark Hendriks: I'm mostly trying to recreate games, like, like sort of breakout or pang, like the very basic stuff, like a platform game. And they're all like this, like, what they weren't called, like no code tools back then, but you had like click and play, and multi-mating fusions, like some, maybe some people recognize this, but you can kind of like click together your games and like in sort of in a fast way without having to learn all the code. And as I get a little further, then I, you know, understand programming a little better, but it was like a nice stepping stone. So mostly like recreating games to teach how that out of everything works.

Ryan Lazuka: Then one thing you did, Mark, I remember you had a really cool weather app on the iPhone store. Do you still have that? Yeah.

Mark Hendriks: The old one now, it's no longer there. You mean well better? I think so, yeah. Yeah, yeah. That was the first step. So that's how I really got into programming. So I did this together with my wife, she's an illustrator, and we just sort of like found side branch project idea to create like a weather app together, but I didn't know how to go that well only from like websites. So I had to learn Swift, and that's how really that's like seven years ago now. That's how I got like more like a program instead of like not knowing what I was doing.

Ryan Lazuka: Your design works really awesome. It is like where do you see, I guess we'll sort of jump around here, but where do you see like design with AI going? You know, you can go on mid-journey and in Dolly 3 pretty incredible designs right now for things like I just actually one of the, you know, like you mentioned chat GPTs before the interview just came out last week, and I was playing around with one of them yesterday. It was like a logo maker, and I just said like it was a robot cupcake, robot cupcake company, and it made it this logo, and it was a really nice looking logo. Like I was blown away by how good of a job it did. So what do you think? How is design going to play with AI?

Mark Hendriks: I think that like it depends on who you are. Like if you're like another designer, and you're looking to get like a logo done or an app icon, it can get like pretty close to something that looks like great or especially like good enough. If you're not like a sort of like a side project, you don't have like the budget to spend hire like a really good designer. So it gets like it's it can produce some really cool stuff. And then like sort of more as a designer, I think for me it's mostly a way to kind of kickstart ideas like let's say you want to design a logo or like a new app interface, like instead of like searching around for inspiration, which you kind of do anyway, these prompts can give you some like unexpected results, even if they're like a little funny. Sometimes like they're not like the end product, but they can really bring some new ideas to life.

Ryan Lazuka: Definitely. One of the things it seems the one thing it seems to struggle with that designers do a much better job at is like giving it feedback. Because a lot of times you can go in and give a prompt for something for an image.

So you want to create a thumbnail or something like that, it outputs it and there's one little thing you need to tweak. And if you don't know how to use Photoshop or something like that, or even you know, Canva or Figma, it's so hard to write a prompt to say just tweak this one part or change this one part, like Firefly has that in Adobe, but still like it's still you still need a human there to help you in those situations, at least from my experience.

Mark Hendriks: Yeah, and that's why I think it's helpful. Like if you like a program or design, like I think it's like a like another thread of it, maybe in five years, but right now, not so much because it's like you, like to your point, like you can, you know what to ask for or if the AI writes some code and you can program, you kind of tell if you're going in a wrong or right direction. And the same with design.

So I think it's like a huge benefit. If you're not a programmer, you try to like create a program from scratch or like a designer, like you don't know what to look for, right? You don't know what what is correct code, are these colors like matching? But if you do, then it's like a really good help.

Ryan Lazuka: No, one thing I think you do need help with a designer with is like the experience that you guys have with like contrast complementary colors, the contrasting colors, like what everything, all the little tweaks behind things really make a difference in the final outcome of whether it be a website, a logo or an image you have to edit. So eventually, I will be able to do all that little tweaking for you. But right now, at least, I think designers still the way to go. Yep, yeah, we're safe for now.

For now, though, but what do you think? Do you think that's going to change like, you know, five, ten years from now where you'll be more of a front manager than a designer?

Mark Hendriks: It's, I have no idea really. Like, it's so, it's kind of like the internet in the early days, like there's a lot of exciting stuff and and a lot of hype as well. It's probably going to like on the short term, like there's all these amazing things, but in the long term, it's probably going to take longer than you would think to get to that, like, like where it's like really dependable. And because now it can be like hit and miss, like it's going to be very hard to get like to the last, let's say five or 10% to make it where you can like really replace someone on like a good level.

That's that's kind of what I think, but it's it's hard. Like, if you want to talk about like last week with the whole open open AI release and and a whole new GPT thing, like I've been impressed, like this seems to be like a big step forward. I have this like little side project called daily four, which is like a like a to do item app, which only allows you to have like four items at any given time to kind of help you focus. And that's been sort of my playground to test AI stuff with and I made like this productivity coach based on that principle and some other stuff. And it got like really helpful. Like before then it was like sort of like a cool demo, but I can see how this would actually work. So seems to be like, like this last week seems to be like a pretty big step forward.

Ryan Lazuka: So you made it, did you did you make a productivity coach with the chat GPT's?

Mark Hendriks: Yeah, I got a GPT floating around. Yeah, since yesterday. Do you mind sharing the name of it is a public? Yes, it's public like it's called daily four. And I'm not sure like right now you need still need like the GPT plus subscription to access this, but apparently there's coming some like store or GPT thing you can visit, but it's publicly available and some people have been been using it.

Ryan Lazuka: Yeah, I think I was just I just did a video on this last night is right now you can't go on chat GPT and like search for GPT chat GPT's.

Yeah, you Google that name, it will come up like Google. What is it again for the daily four, daily four and then just put like space to GPT after it and it'll probably come up in the Google search.

Mark Hendriks: Yeah, but like a lot of people are creating GPTs now like it's it's it's I'm seeing all kinds of interesting things like being able to learn Spanish, like sort of this like this tutor kind of set up and like for programming and like cooking like it's it's it's fun to see what people are people are creating.

Ryan Lazuka: Did you see that the Humane AI pin that came out last week as well that

Mark Hendriks: I did not see the presentation saw some stuff on on Twitter. Yeah, but it's using open AI.

Ryan Lazuka: The cool thing about it was you mentioned translations it would do if you wear this pin, I could speak with someone that speaks Spanish that doesn't know English and vice versa, and it translates on the fly so it would translate their Spanish to English on the fly, and then I would respond in English to them, and then it would translate to Spanish back to them so like in real time as a school.

Mark Hendriks: Yeah, I think some things like like an iPhone app, but if it's like sort of more like sort of almost natural like you don't have this sort of your phone for that could be pretty cool.

Ryan Lazuka: Yeah, I guess it does already exist in technology, but yeah, I see what you're saying. Back to the the heart of everything the light year. Yeah, yeah. Can you tell us a little bit how you got into the design work for that? I forget if you mentioned or not already, but how did you you and the co-founder get together and start this project?

Mark Hendriks: So my co-founder Eric, we've known each other for seven years. So the way I got to know him was like he had like a different startup and he needed the designer. So we got to work together and we worked really well and at some point we said like we should start a company together without like sort of having the idea yet. So we talked about that and then like, they like six months later pitched me on this like early prototype for the integration. software and I said, let's build this thing together. So we go found the company and we've been building this whole thing sort of in secret and having some people on like a small list of like a small set of users like using it and improving it with it. So that's been like over a year now. So that's how I got to design the whole thing because we started together.

Ryan Lazuka: And did you guys know each other from the Netherlands or?

Mark Hendriks: No, just like the internet. We did met in the US some years ago, but it's all through the internet.

Ryan Lazuka: So it is a US based company then? For the most part? Yeah, it's US based. So what is the specific problems that you've run into in trying to develop the project and then how have you overcome those problems? Is there any big stumbling blocks that you've come across?

Mark Hendriks: Some products you create, it's okay if they don't work that well, like this early stage thing, but if the integrations don't work and you don't get your slack notifications, then you're not going to be happy. So it's, I guess, a bit more upfront work to make sure works reliably and well. And the other part would be getting like enough support for like integrations, you know, trying to, you know, you don't have to match the API, but you need like a few to make it like versatile and useful for people. So the way we got around that is like using a thing called custom apps. So you don't need us to add certain services or if like your own API, for instance, something like within your team or company, you can also add it to the platform. So we try to make it sort of like you're not blocked by what we are doing. Like if you want to write your own service there, that's like an option. So that's one way we kind of got around that roadblock of where do you start with integrations.

Ryan Lazuka: And what are you guys working on right now? What's the current next step?

Mark Hendriks: We just finished the Assistant API that we have support. So I think it's rolling out today. So you can, instead of using the GPT, there's also like an API. So if you program, then you can create your own assistance through our platform. So that's been like the, I would say the bigger release for this week and then being generally available. So you know, normal wait list, people can just sign up. That's been pretty big week for us.

Ryan Lazuka: Very cool. So we caught you at a decent time. Yeah, very. And then we kind of ask a lot of people this, but we like to get everybody's opinion, people who are really immersed in this AI space. Where do you see AI going generally? What's the direction of AI more generally? Some people think it's like AI relationships and characters. Some people think it's a doomsday.

Mark Hendriks: What do you think about this? I've been mostly approaching it from like a work perspective because of like design and the whole light year product. So I mostly see it as like a enabler or like if you think of like, if you're like a company, which can be like, you think like a bad scenario, like, like, if you're like new to programming, you're like a junior developer, I think it's going to be a little harder than it used to be because AI can do sort of more mundane tasks. I would say that's may not be the most positive thing. Or maybe you can learn a lot faster AI. It's kind of hard to tell, but it's definitely some stuff where it's getting more capable, which can be hard for some people.

So most like a work, I would say most from like a work perspective, that's kind of how I'm using it. I also have like a six year old who is, is, does like, like he has like this wild story idea and does like the visual prompt of like sort of visualizing this, you know, like, like the, the Earth cloud can talk and it's like bacon flying around stuff. It is fun to see that you may like to draw, of course, and kind of support that, but like being able to create all these funny, funny things is, is, I mean, it's pretty cool for him.

Ryan Lazuka: Yeah, one of my favorite parts about AI is just exploring all the different tools. Yeah. And I always go to like the fun sections of the tools because I think it's pretty interesting. Some things you just never think of that. That's probably my favorite part of that. Yeah, I mean, it's, it's,

Mark Hendriks: I say it's, I'm, I'm generally like excited about this, like this, there's so much cool stuff people are doing, like seeing now with these old GPTs, but like sort of all these services that people create, you did not think of. So I think I need to like look in like sort of more like a personal level, like what kind of do like maybe we're like, make food like I saw one where you can list like I've like these six things in the fridge. What can you like cook today? That is fun. Like, it might be interesting or not. I haven't tried it yet.

Ryan Lazuka: I feel like the bartender one where it's like create your own types of drinks and right kind of type in what you like and it'll give you some brand new recipe about, I mean, that stuff's really interesting. I think there was a AI girlfriend one so you can be that's it. That's the name GPT AI girlfriend.

So I tried it for a video I was doing. It said it was like, well, I can't officially be your girlfriend, but I'll basically do everything a girlfriend would do. I mean, other than certain things.

Mark Hendriks: So the like they had like chat one of the things chat GP chat GPT and opening are doing a lot more sensory things. So it was pretty dumbed down a little bit. But yeah, it's fascinating how many, how many tools are out there for sure.

In terms of light year. Can you tell us a little bit about like what the what you see are the best use cases are mark of like someone that wants to use the tool. Are there any specific examples that can really maybe help someone's productivity, whether it be in their personal life or for business. Well, because it's like it is like bill for for developers have been a lot of like workflows for developers or like connecting, you know productivity after their code like get up to to linear and slack and being like really specific on how that works. So we've been seeing like people going like more beyond standard integrations and they really have like specific requests and how it should work. That's kind of where code for them shines. We also like the interesting thing where sometimes more basic integrations don't work as expected from other platforms, which I did not really foresee but that has also been sometimes a reason why people switch to light year. So it's it's it's mostly like trying to be very hard to expression express in like an official way where you like that like these building blocks and trying to tie them together and like these weird adc cases that are very hard to capture. That might be very good reason to to use code and use my dear that like has been one of the main things for us.

Ryan Lazuka: So it's almost like the tool the tool itself is the efficiency.

Mark Hendriks: It is. Yeah, it feels more natural like this whole facial part might be not so natural because the whole like if you building software you're writing stuff you're writing code all day long and then you get to the part of integrations and you kind of have to like learn this new tool and all facial and it doesn't always work out.

Ryan Lazuka: It's it's for developers but it's not like you can if you have a little bit of experience the code you guys have very clean code like you should be able to catch up but fairly easily I would imagine.

Mark Hendriks: I'm not like I don't consider myself to be an engineer. Like I can I can code I could not create light year on my own. So that's kind of like how I've been seeing it like I'm a designer I've been doing a few apps and then it's for me very easy to learn light year I know I'm the co found some pretty biased and deep into the product but it has been like when there's like like with the assist assistant API like playing with that it. It's very easy to pick up and we we spend like a good amount of time on documentation.

That's what like if you think like being frustrated with other products like having like outdated documentation or incomplete or non working examples really trying to make a good effort on on making a whole lot better. Because I think that's like maybe the biggest thing where you get judged on by all the different appraisal. Like if they can't get through the getting started tutorials like unclear or it takes a very long time you kind of sort of like a very bad impression.

Ryan Lazuka: Yeah, and sometimes those projects are so big that it's hard to like sit through everything that you need to get to like their documentation can be so thick where you guys are a newer startup. You don't have many creations but that's can be a benefit

Mark Hendriks: in a way you know yeah it is yeah and then there's like go by it so if you. If you're using if you're like a programmer using go by it you will see that some of the integration work on like it can be written by co pilot. So it kind of goes back and forth like it's one big sort of go by it's a thing where like

Ryan Lazuka: it's pretty crazy even if you're a you're not a developer you don't have it hardly any understanding. Of code but you're smart enough to like look at read some documentation. You could and you don't even know co pilot you could put the code into chat you will help you as well so if you if you have familiar familiarity with chat you can do use it that for code as well. But yeah it's a it's a cool it's a really cool product and like if you thanks yeah it's if you have any experience anyone had experience with the Zappier make or something like that.

And you get frustrated. Light year is a great alternative and again if you look at the code. I know I was given I gave that example the beginning of this podcast is like I was doing something to summarize podcasts and get a summary from it from it and write it have chat chat EPT write a summary for me. For YouTube for the YouTube description description and it's just very very hard in those tools.

Even though it's supposed to be easy and if you just like maybe give a later try or for something like that or something similar. Their documentation is very well done so even if you don't even if you don't know code. It should be pretty clear on

Mark Hendriks: how to use it that that was my take at least yeah and we have like quite a few examples for every integration I think the like 20 get up example so. If you look at those examples you have a pretty good idea of how to work or you get inspired on how to create your own. Integrate integrations that is like as a designer has always helped me like learning I try to learn to go to like a book. Now going through all the functions and I'm like I just want to make a weather app like how do I make a weather app like just show me an example and that's kind of how I look how I learn best and not for everybody but like. That's one way to approach this make it sort of like you kind of want to have like this thing run run in a few minutes like if you try something new and not be to 10 pages of like try documentation.

Ryan Lazuka: Someone recently I saw said the best way to learn is just try it and then if you have a question about it look it up on Google or the documentation. You know play as you go so so mark. What's your long term vision for this project where you see it five years five years ten years down the road ten years at a long time.

Mark Hendriks: When it comes to like the integration part it's it's easier and better done code and I think with AI. I think it's going to be even easier also like if you're like not that experienced there's so much like. I can do for you like writing a code like I mentioned I that you can just prompt the integration you want to create and it will like buy the code for you like that's also fine with me like it's and you can modify it like it's that it will be like the somewhat longer term vision where you have to.

If you want to write write code you can if you don't want to write a lot of codes have someone else do it for you. So I feel I feel like that's sort of going for us being able to create a product faster and follow people to use the product also fast like it seems to go like a very like natural way. Yeah so that's that's kind of how we see it and like ten years I don't know how we'll be very successful.

Ryan Lazuka: Do you see a world where coding is no longer needed.

Mark Hendriks: Unlikely less needed or likely will be different use cases like maybe the things we're doing right now will be like automated away but there might be new challenges or new things to develop or more complex things so I think some of the maybe like sort of easier or basic things will go away. To some extent but probably it's kind of just going to be a whole lot more code. In the future.

Ryan Lazuka: Is it possible you think light year will come to a point where it will almost be like a chat GPT where you can like write a prompt and say I want to tie in. Yeah. These three products to do it for me. Yes that that is something.

Mark Hendriks: But what we're experimenting experimenting now is like giving a documentation from some other servers who want to integrate with like read this documentation and write the integration for us using you know the structure we have in place or using these examples.

That's kind of the first thing we're trying to prove. And if that works I would say the next step is that you have some sort of integrated. Promptability to write like integrations for you. That would be that this something we're experimenting with.

Ryan Lazuka: Yeah that'd be that'd be a problem.

Mark Hendriks: That would be yeah I would be amazing. I mean it's it's kind of like I think it's. Especially after last week I feel like a definitely step closer like it's going to be hard just to give like the GPT like a URL like a code create this whole thing but you know certain steps and certain parts and we'll be doing that more and more.

Ryan Lazuka: I mean just a few months ago you couldn't use like banging integration or search integration with chat GPT and you can do that now to ask for real time data so.

Mark Hendriks: Yeah and you can add like very specific models for programming or you may do something on your own like we're not like a you know I primary AI company so I wouldn't be you know wouldn't be the best use of time to try to create your own model but I can see how that's going to be more. Possible like these things are getting smaller and smaller and. You may be able to create your own model for your own

Ryan Lazuka: product for the aspect of late year you guys are clearly aren't it's not your main focus we understand that but what. What aspects of AI are being used in the product right now in terms of for the end customer like they want to tie in the chat GPT they can do that it sounds like. What is there anything else they can tie into that will help them with the aspect of things

Mark Hendriks: because it's like an integration thing you can use you could use something like get all the latest issues from get up. And summarizes using AI in like a sort of a daily digest and send it on slack like in an email so you like basically leveraging all these services. And so you can use as much AI as you think is helpful to do the whole business and I will be adding more integrations right now that's been a bit more on. Like the developer side of things but we have been adding like all the Google tools and it will be like more content stuff.

Maybe even something for podcasting like we'll be adding like other areas as well where. I mean I think you mentioned like I'm sure if you already doing this right but like generating images for like the blog post or Twitter like it's like already automated at this point or do you still like do this.

Ryan Lazuka: No I really I was doing Twitter posts for a while on my own like one every day and that was the one thing I struggled with the most is finding images and I cannot find a tool to do that for me that did it well. So is it sounds like there's something out there that can do that now.

Mark Hendriks: I don't know maybe GPT like I think it's getting close on the trip is like do you want to have like an automated phrase like completely like independent and it will just do it Twitter and you kind of trust the image is that kind of the end goal or.

Ryan Lazuka: That would be the ultimate end goal yeah I don't think it's there yet there's some tools out there I think it's post wise. If they do that for you but they're just it seems to the tweets and the end tweet doesn't sound good you can tell it's generated. The main thing it starts with it like say if I said find me five images about that humane pin that we talked about the AI pin and just give me stock photos for that like it won't do that. Or I will so things like that would be a huge help in the future and then like I know the thumbnails are a lot better there was a thumb there's a thumbnail chat GPT and it did a pretty I did test it one last night and did a pretty pretty pretty great job I was kind of blown away by the results so everything's getting better very quickly

Mark Hendriks: and it makes a hard to predict I think it still struggles with text like if you want to create. Yes like that was in the sense gets like off words

Ryan Lazuka: like misspell it or put a letter back words event. And then you love the image but you want to get rid of the taxes.

Mark Hendriks: An open Photoshop or something like back

Ryan Lazuka: to where you are yeah like with everything. Lightyear all these new tools everything's progressing very quickly so and if you want to promote anything now is the time to do a mark like

Mark Hendriks: we're opening up sort of publicly today so you can go to the chat. You can go to run lightyear dot com you can create a few accounts and have like your first integration on the way pretty fast. The other big thing I would say is like we have a development environment so that's something that runs on your machine. It's free there are no limits you can write as many integrates as you want and by the time you're ready you can run on another platform so like give the try doesn't cost you anything with like a lot of cool examples out there and we'll be adding more integrations very soon.

Ryan Lazuka: So if you're in run lightyear dot com check out their projects also subscribe to Ryan eyes weekday newsletter fry hyphen ai dot com you could weekday news three top stories of the day along with some cool schools and community interactions. So just thank you so much mark for coming on today we appreciate having you and wish you the best of luck.