Dec. 14, 2023

Viral Video Marketing and AI Content Creation with Tomer Dean, CEO of Lychee

This week on The Index, host Alex Kehaya welcomes Tomer Dean, CEO of Lychee, a premium video content marketing platform. Tomer, along with his partners, started Lychee to assist brands and content creators in producing viral hits for TikTok, Instagram, YouTube, and more to reach over 1 million views.  Learn how Lychee's unique hybrid AI architecture delivers high-end content for social media channels. 

Host - Alex Kehaya

Producer - Shawn Nova

 

 

Chapters

00:04 - AI and Future of Content Creation

09:08 - Scaling Communication Efforts With Leaky

14:07 - Entrepreneurial Ventures and Learning Experiences

19:03 - AI, Blockchain, and Content Distribution Intersection

Transcript
WEBVTT

00:00:04.427 --> 00:00:08.013
Welcome to the Index Podcast hosted by Alex Kahaya.

00:00:08.720 --> 00:00:26.547
Plug in as we explore new frontiers with Web3 and the decentralized future, hey everyone and welcome to the Index brought to you by the Graph, where we talk with the entrepreneurs building the future of the internet.

00:00:26.547 --> 00:00:37.851
I'm your host, alex Kahaya, and today I'm excited to introduce you to Tomer Dean, co-founder and CEO of Leachy, a premium video content marketing platform.

00:00:37.851 --> 00:00:42.771
Tomer is a serial entrepreneur who sold his first company at the age of 15.

00:00:42.771 --> 00:00:59.267
And I think Leachy is super cool as someone who produces a podcast and also recently started doing a lot of video on TikTok and Instagram and YouTube, because they leverage this hybrid AI architecture to deliver high-quality content to their customers, and most of you guys listen to this show.

00:00:59.267 --> 00:01:20.170
We primarily talked about Web3 in the context of crypto and the blockchain space, but I also feel like the future of the internet leverages a lot of AI and things like virtual reality and augmented reality, so we've decided to expand a little bit on that, and I think that these technologies are twiner and are going to empower people in a whole new way, especially creators, even creators like us making the Index.

00:01:20.902 --> 00:01:21.971
Tomer, thanks for being here.

00:01:21.971 --> 00:01:22.799
Welcome to the show.

00:01:22.799 --> 00:01:24.947
Thanks, alex, good to be here.

00:01:24.947 --> 00:01:30.343
So maybe, just to start out, if you could just tell us a little bit about yourself and your background.

00:01:30.343 --> 00:01:31.668
I would love to hear your story.

00:01:32.140 --> 00:01:36.231
Yeah, so I grew up right in the center of the tech world in Silicon Valley.

00:01:36.231 --> 00:01:41.132
Both my parents were engineers, so I guess the Apple didn't fall far from the tree.

00:01:41.132 --> 00:02:00.537
And they started coding from a really young age, after starting to build computers a bit earlier and it was just fascinated with the internet and this was the early days of e-commerce and I was in the eBay and some dropshipping back then, which was still not as sexy as today.

00:02:00.537 --> 00:02:16.203
In recent years, I've been working with content, a lot of content creators and start becoming a creator myself, and this you know leachy, what we're building today actually started as a side project when I was looking.

00:02:16.403 --> 00:02:22.074
How can I repurpose my own podcasts in this whole video AI world?

00:02:22.074 --> 00:02:46.294
The paradigm is shifting very quickly and it's interesting to see how many things AI can help solve, but also, on the other hand, where AI is still lacking and will likely always lack some elements, especially when we're talking about content which is very subjective, and I know a lot of people don't want to consume AI generated content and I can resonate with that.

00:02:46.294 --> 00:02:55.224
So how do we still use these amazing advancements but still keep that creativity, which is what people are drawn to?

00:02:55.224 --> 00:02:58.413
I think it's a really interesting problem to solve.

00:02:59.060 --> 00:03:03.931
For me as a creator, my outlet for creativity is this show that's doing these interviews.

00:03:03.931 --> 00:03:06.790
But I'm not a video editor, like I'm not a designer.

00:03:06.790 --> 00:03:08.519
I can design a landing page.

00:03:08.519 --> 00:03:09.563
My life depends on it.

00:03:09.563 --> 00:03:11.794
I mean, I could get a basic one up, but it would look terrible.

00:03:12.578 --> 00:03:24.867
What I like about your service because I've tried a lot of the AI driven ones just to like, hit a button in it, like, or give it the link to my YouTube video, and typically the output of like cutting the clips for me isn't that great.

00:03:25.068 --> 00:03:48.132
It's getting close but it's still not good enough and it's a lot of work to go through and like edit those videos or cut it even further, like watch the whole episode and cut it and you know what you guys do is super interesting to me because it's that 80, 20 year old, right, like 80% of the work is completed by the AI to short circuit the process, and then you have editors who come in and like just polish it off so that it's going to look really great.

00:03:48.132 --> 00:03:52.150
It has like a really professional look and also just make sure that that content makes sense.

00:03:52.150 --> 00:04:02.800
Because if you were to just give AI today at least the services I've checked out you know a clip to edit or a video to edit and said distribute this formula on all my social channels.

00:04:02.800 --> 00:04:04.865
It would come out not making quite sense.

00:04:04.865 --> 00:04:08.830
Like they might hit the mark sometimes, but like other times the story is just way off.

00:04:09.031 --> 00:04:10.114
It would be a spray and pray.

00:04:10.114 --> 00:04:13.162
They would most likely be second rate.

00:04:13.162 --> 00:04:14.846
But you know, here's the dilemma.

00:04:14.846 --> 00:04:17.733
Let's say the all these AI platforms get better.

00:04:17.733 --> 00:04:19.201
They obviously will get better.

00:04:19.201 --> 00:04:25.355
What happens when everyone can do 80% and when everyone's doing the same level?

00:04:25.355 --> 00:04:32.581
It just increases the competition, Because you as a creator and me on my own channels, we're basically competing with everyone.

00:04:32.581 --> 00:04:34.112
It's a zero sum game.

00:04:34.112 --> 00:04:40.872
There's only a finite amount of human attention for everyone, so we can't all be going viral.

00:04:40.872 --> 00:04:54.055
So I think AI will have a huge effect in increasing the standard of video, but to actually succeed in getting performance right, getting views that's what we all want that would be even harder.

00:04:54.055 --> 00:04:56.841
It'll all be in between 80 and 100%.

00:04:56.841 --> 00:04:58.163
How close can you get to 100?

00:04:59.589 --> 00:04:59.810
Yep.

00:04:59.810 --> 00:05:10.451
So tell me more about, like, how you guys got started as a company and how did you land on this split between the AI and the, because you mentioned to me before the show, you started out as just pure AI.

00:05:10.451 --> 00:05:12.980
You wanted to be a pure AI player, which I thought this is how we got connected.

00:05:12.980 --> 00:05:32.375
It's talking to my producer and I was like man, I wish I could just upload my content and then AI will just deploy that content with really high quality clips and sound bites and tweets and all that stuff for me and like I feel like that's got to be possible in the future, but we're not there yet, and so I'm curious how did you land where you are as someone who's actually tried to build an AI platform that does this?

00:05:33.658 --> 00:05:33.858
Yeah.

00:05:33.858 --> 00:05:47.098
So we started in 2021 building AI tools for content creators and we started it was actually a funny story with the whole clubhouse trend that was happening in 21 there and then also Twitter spaces.

00:05:47.098 --> 00:05:52.197
There was the same issue with podcasting how do I repurpose, how do I distribute my content?

00:05:52.197 --> 00:06:20.168
And we had a very novel idea of let's use these new AI developments chat GPT was getting released to automate this whole process of repurposing, clipping, posting and what we saw is, even if we use the best frameworks out there today, the best we can do is 60% of the quality that is needed to be at the top 1% dial level.

00:06:20.187 --> 00:06:32.805
And we had to make a decision Do we want to continue building a product that's going to continue sucking, in a sense, right in that 60% mark and slowly inching our way to 70?

00:06:32.805 --> 00:06:35.149
Or do we want to provide 100%?

00:06:35.149 --> 00:06:43.348
And we decided we're going to offer 100% polished results as good as a human editor.

00:06:43.348 --> 00:06:52.048
That was our framework and we just saw that it wasn't possible to do this with AI in 2023.

00:06:52.048 --> 00:07:07.048
That's why we started understanding that, all right, if we add a very minimal amount of human work on every piece of asset, can we increase that 80% to 100 or to 99?

00:07:07.048 --> 00:07:18.014
And we saw it's definitely possible with a very lean layer of human service that's very efficient, to properly deliver 100% outputs at scale.

00:07:18.014 --> 00:07:20.826
That's how it evolved to wear this today.

00:07:21.971 --> 00:07:25.185
Yeah, I think it's super interesting and one of the reasons why I wanted to have you on the show.

00:07:25.185 --> 00:08:00.367
It's not just because of the AI piece that's relevant to what we just talked about the future of the internet but also my background is primarily, the last six, seven years has been in crypto, and part of the reason why I started the show is because one I believe people are worth knowing and I believe that if we can tell the stories of the people who are building the future of the internet, specifically in the space that I'm in, then the general public will understand why we're here and why it matters, and versus the narrative that they're seeing today from Congress and things like that, which is pretty negative.

00:08:00.367 --> 00:08:04.908
Even people in my neighborhood, where I live I, have a pretty negative opinion about crypto, as they should.

00:08:04.908 --> 00:08:53.341
I mean, there's been a lot of problems, especially last couple of years, with FTX especially, so we've gotten a bad wrap, and I think One of the solutions is producing really high quality content at scale that reaches a large audience, and that's my goal with the show, and I want every single founder who's listening to this show, who works in my space, to start thinking about how can I use tools like what you guys have built to achieve that goal, to disseminate the message around why they're building this through like more channels than just Twitter, right Like, we are in a complete echo chamber on Twitter with like a very small audience of crypto enthusiasts and early adopters, versus like all these other channels that have much broader reach, especially for the younger generation, who consumes things on like TikTok, for example.

00:08:53.623 --> 00:09:07.869
Most of what I've seen on TikTok that relates to crypto and blockchain technology and AI, even but especially crypto, is just like about token price and being a crypto expert, and that's not why anybody I know is building in this space.

00:09:07.869 --> 00:09:17.285
My hope is that, like people can use these ideas and the tools that you're working on to scale up their efforts to communicate with the world, and that's why I'm doing it.

00:09:17.285 --> 00:09:19.046
So I can hire you guys.

00:09:19.046 --> 00:09:22.208
First of all, it's not just production of the content and handing the clips back to me.

00:09:22.208 --> 00:09:34.024
I can actually hire you guys to do full service, right, like, even distribute it through my social media channels, and the pricing you gave me was pretty affordable relative to like if I hired an agency that doesn't use the tools or have the tools that you guys have.

00:09:34.024 --> 00:09:35.551
You're talking about 10,.

00:09:35.551 --> 00:09:40.869
In my experience, like 10, 15k a month of, like you know, four people working for you.

00:09:41.740 --> 00:09:43.287
And so like what does the pricing look like?

00:09:43.287 --> 00:09:47.710
What are the different options and what do you get if you're trying to use Leaky to produce content?

00:09:48.580 --> 00:09:55.200
You know, we started with just the offering the video, editing the clips, and what we saw is that it's very difficult.

00:09:55.500 --> 00:10:15.051
There's a lot of nuances on these platforms, and knowing how to actually use these videos, upload them in the right way, in the right format, with the right hashtags, the right time of day in some cases is very important, and very few people know these things and stay on top of all the trends.

00:10:15.820 --> 00:10:20.600
So today, you know, to make sure that our customers are successful that's our ultimate goal.

00:10:20.600 --> 00:10:54.572
We offer the fully managed solution, which is very similar to what you mentioned with the traditional agency, but because around 80% of the work is automated and only 20% is done by humans, we're able to offer this at a very different price point than what was acknowledged today, and we can offer a fully managed podcast, repurposed to three platforms for $3,000 in some cases, compared to that around $15,000, $20,000 a month for a traditional agency.

00:10:54.572 --> 00:10:57.967
We don't do it by decreasing our margins.

00:10:57.967 --> 00:11:11.907
We're also earning a healthy margin, which is important as you're growing your company, which involves also building AI and also servicing many customers, and that offers a really interesting win-win.

00:11:11.907 --> 00:11:33.874
Somebody can access a team of three focused personnel, in a sense, for your show without hiring those three, four people or hiring an agency and be able to control this device, but also never sacrificing that quality we were talking about, which often is sacrificed with these other tools.

00:11:34.741 --> 00:11:35.663
Just so I get it straight.

00:11:35.663 --> 00:11:37.246
You guys can do the audio.

00:11:37.246 --> 00:11:39.010
You can put that on my RSS feed.

00:11:39.010 --> 00:11:43.369
It gets distributed and then you can do the video on two other channels of my choosing.

00:11:43.369 --> 00:11:49.686
You could do that full soup tonight for $3,000 a month, and how many clips do you get out of that?

00:11:49.686 --> 00:11:52.206
How much content does that actually end up being?

00:11:52.206 --> 00:11:54.884
Let's say, I picked YouTube and TikTok.

00:11:54.884 --> 00:11:57.610
How much stuff am I going to get distributed that way per month?

00:11:57.610 --> 00:11:59.365
Let's use my podcast as an example.

00:11:59.365 --> 00:12:01.344
We're doing now.

00:12:01.344 --> 00:12:09.546
We just bumped up to four episodes a month, so one per week which has audio and video with it.

00:12:09.546 --> 00:12:11.025
So we use that as the example.

00:12:11.701 --> 00:12:11.880
Yeah.

00:12:11.880 --> 00:12:17.511
So what we would do here, we would create 25 clips a month from those four episodes.

00:12:17.511 --> 00:12:23.753
That's roughly five an episode plus another one we would do on the most successful clip of that batch.

00:12:23.753 --> 00:12:28.826
We would do another round, those we published to two or three platforms.

00:12:28.826 --> 00:12:36.948
So that would be 50 to 75 clips being published on those channels and maybe a few more on LinkedIn as well.

00:12:37.659 --> 00:12:38.804
And you guys do the copy too.

00:12:38.804 --> 00:12:42.349
For those clips You'll write like the TikTok message or whatever it is.

00:12:42.349 --> 00:12:45.265
You'll do all that and you can just distribute it kind of on autopilot.

00:12:45.265 --> 00:12:46.744
Exactly exactly.

00:12:46.744 --> 00:12:48.405
That's amazing for everyone else.

00:12:48.405 --> 00:12:49.946
I have no stake in this company.

00:12:49.946 --> 00:12:53.711
I'm not an investor or a partner or anything like that.

00:12:53.711 --> 00:12:57.087
I just am genuinely interested in planning on using this.

00:12:57.087 --> 00:13:05.457
So you'll see over the next month or so some stuff coming out on our channels that leverage this so you can see for yourself, including this episode.

00:13:05.457 --> 00:13:11.099
You'll be able to see for yourself what that process looks like and I'm excited to test that out.

00:13:11.099 --> 00:13:18.845
But, yeah, super scalable and not that expensive relative to for this kind of level of what I think you're going to see comes out of it.

00:13:18.845 --> 00:13:23.235
Let's go back in time a little bit, because you've been an entrepreneur for basically your whole life.

00:13:23.235 --> 00:13:26.283
It sounds like what was the first company you built you sold when you were 15.

00:13:26.283 --> 00:13:26.985
What was that?

00:13:27.416 --> 00:13:34.880
All right, the first company is actually a company I flipped and it's a funny story how I got to flipping a company at 14.

00:13:34.880 --> 00:13:37.128
But I was really into the forums.

00:13:37.128 --> 00:13:38.993
This was in the mid 90s.

00:13:38.993 --> 00:13:42.019
There was all these new websites, had forums.

00:13:42.019 --> 00:13:47.115
You remember those old school forums with the threads, yeah, yeah.

00:13:47.115 --> 00:13:53.210
And through the forums I found this one website called SitePoint.

00:13:53.210 --> 00:13:55.875
Today's is called 99 Designs.

00:13:55.875 --> 00:14:01.107
Today it's a service that you can order logos, but back then sorry this was called Flippa.

00:14:01.107 --> 00:14:03.340
It was a brokerage website.

00:14:03.340 --> 00:14:06.567
So back then it was a forum where people would be selling their businesses.

00:14:06.567 --> 00:14:11.331
And I was really engaged with business on this forum and I saw this.

00:14:11.461 --> 00:14:16.808
One person was selling a website that looked like an amazing business to invest in.

00:14:16.808 --> 00:14:18.434
All right, here are the specs.

00:14:18.434 --> 00:14:26.054
It was a content website that had tutorials on how to do things on Excel All right, microsoft Excel.

00:14:26.054 --> 00:14:28.844
And this was in, you know, in the late 90s.

00:14:28.844 --> 00:14:38.394
And this website had 20,000 unique visitors a month and was earning $200 a month in ad revenue and they were selling the website for $2,000.

00:14:38.394 --> 00:14:40.259
That was the asking price.

00:14:41.462 --> 00:14:46.235
So you know, under one year, multiple, pretty good, but I don't have $2,000.

00:14:46.235 --> 00:14:47.096
That's the problem.

00:14:47.096 --> 00:14:53.794
So I told my parents about the opportunity and my dad agreed to loan me the $2,000.

00:14:53.794 --> 00:14:59.548
As you know, he thought it would fail and I would get a great business experience.

00:14:59.548 --> 00:15:06.048
I bought this website the whole negotiation, you know, transferring of the assets, even the contract.

00:15:06.048 --> 00:15:09.000
I took control of the website.

00:15:09.000 --> 00:15:20.822
I started increasing the content in terms of quality and quantity, so I would be writing some of the tutorials myself, I would hire some other freelancers and this was, you know, at the age of 14.

00:15:20.822 --> 00:15:33.010
And I got to the point where the website tripled in revenue and started doing $600, $700 a month and eventually sold it for $9,000.

00:15:33.010 --> 00:15:36.855
So around four, five X higher than what I purchased it for.

00:15:38.039 --> 00:15:39.384
I love that learning experience.

00:15:39.384 --> 00:15:41.875
It's super cool that your dad and your parents agreed to support you on that.

00:15:41.875 --> 00:15:45.514
You don't know this and probably some people listening don't know this, but I was an educator.

00:15:45.514 --> 00:15:48.229
I started out teaching high school and middle school kids.

00:15:48.229 --> 00:16:05.115
One of the first things I did that was entrepreneurial was build an entrepreneurship program at a school in California that was designed specifically to enable the kind of thing that you're talking about, which is like allow students to actually do real problem based learning, building real businesses and actually talking to customers.

00:16:05.115 --> 00:16:09.211
It really focused on that early stage like market validation process.

00:16:09.231 --> 00:16:09.794
So you have an idea.

00:16:09.794 --> 00:16:13.143
How do you like actually try to make that idea a reality?

00:16:13.143 --> 00:16:17.264
And typically for something that isn't a small business, right, like a startup.

00:16:17.264 --> 00:16:19.446
Small business is not a derogatory term here.

00:16:19.446 --> 00:16:30.100
It just means that there's like a unknown business model that you can use, versus a startup is like there is no business model in many cases, where there's some like big difference in what you're doing than what's what like already exists.

00:16:30.100 --> 00:16:31.725
Anyway, very cool.

00:16:31.725 --> 00:16:32.971
So I have a soft spot for that.

00:16:32.971 --> 00:16:34.697
That story, it's really really cool.

00:16:34.697 --> 00:16:41.301
Did you have more companies between then and now with your current company or like what else have you done between that?

00:16:41.301 --> 00:16:46.644
I know you were in the military for a little while, but did you have any other entrepreneurial ventures before starting this business?

00:16:47.096 --> 00:16:51.096
It's always fun to talk about the winners, but let's talk about one of the ones that didn't work out.

00:16:51.096 --> 00:17:01.628
This is a company called Blush I co-founded six years ago and it was active for four years and we raised close to $2 million in venture funding.

00:17:01.628 --> 00:17:24.288
And what we did is we helped e-commerce retailers find and use different UGC from Instagram of influencers tagging their products or dressed in similar products, and we helped them utilize that on the website to increase conversion rates, and it seemed like a very good idea at the time.

00:17:24.288 --> 00:17:38.888
Looking back, I think we were a bit too late to this kind of UGC free commerce and that took me a while to understand how important is to be early and be at the right timing when you're opening a new business.

00:17:39.791 --> 00:17:42.138
We had, I would say, half product market fit.

00:17:42.138 --> 00:17:53.729
We reached 500K in the annual recurring revenue, which is often at the point where you feel you have PMF, but it wasn't nice to have.

00:17:53.729 --> 00:18:01.510
We weren't able to turn into a must have and the business shut down at a point because money ran out.

00:18:01.510 --> 00:18:13.166
We were offered more money to continue, but we just didn't believe there was a clear path to making this successful enough to justify the VC returns and that whole thing.

00:18:13.166 --> 00:18:17.518
So we decided to close up shop and look for our next opportunity.

00:18:18.961 --> 00:18:31.625
I have been there before as well, several times actually and that happens Like sometimes you get product market fit but you don't have business model market fit right, like the business model just doesn't pan out.

00:18:31.625 --> 00:18:38.577
One of the first companies I started was a company called NextMover and we started out of Startup Weekend in Santa Barbara.

00:18:38.577 --> 00:18:45.201
We actually won the event and like at the final pitch when we won, the judges invested in the business.

00:18:45.201 --> 00:18:46.967
It was crazy and it hadn't happened before.

00:18:46.967 --> 00:18:49.307
But we raised like I don't know 100K.

00:18:49.307 --> 00:18:54.726
It was not a lot of money but it was enough for me back at the time with like two other people to try to see if it would work.

00:18:54.726 --> 00:19:02.224
And we spent like nine months just with that budget trying to figure it out and we did a bunch of manual testing and stuff and we generated revenue and first customers.

00:19:02.224 --> 00:19:07.103
We just realized that the cacti LTV ratios were just never going to work out.

00:19:07.103 --> 00:19:09.099
At least we didn't find the solution to that.

00:19:09.099 --> 00:19:13.759
It doesn't mean someone else won't figure it out, but we never really determined a profitable way to do that business.

00:19:13.759 --> 00:19:24.355
Maybe as a small business it would work, but it wasn't going to be venture-fundable and I even had investment for a seed round like an actual seed round, lined up and turned it down and said I'll call you on the next one.

00:19:24.355 --> 00:19:28.297
I'm not going to spend the next five years trying to build something that's not going to scale.

00:19:28.297 --> 00:19:31.265
And that's like an important lesson as an entrepreneur to learn.

00:19:31.404 --> 00:19:45.723
A friend of mine, an investor in Olaplex this guy, vinnie Lingham, who many of you in our space will know, told me that one time you only have so many cycles, that you only get so many chances, because you just don't have more time on this planet to build things.

00:19:45.723 --> 00:20:12.242
And so the adage of failing fast is actually not just about the business stuff, right, like trying to win and generate revenue and all that stuff and find the right product, but it's also about your personal time and how many times you're going to get to do this, to actually try to build a business as an entrepreneur, which I think it requires a lot of sacrifice I know it does, but it's also a huge opportunity and privilege to be in the position that we're in to try to build companies.

00:20:12.242 --> 00:20:18.535
If you're listening to this and it's the bear market and you're struggling to find product market fit or you just don't think it's there, do something else.

00:20:18.535 --> 00:20:19.719
You're not giving up.

00:20:20.019 --> 00:20:20.981
It's not that you're quitting.

00:20:20.981 --> 00:20:30.250
It's that you're evaluating based on the data you have like the opportunity, and making sure you're spending your precious time and skills on something that's going to be impactful in the world.

00:20:30.250 --> 00:20:33.200
So I think it's a really poignant story from you.

00:20:33.200 --> 00:20:39.684
Appreciate that as we get towards the top of the show, we still have like five or seven minutes left here, but I can't always ask the same thing.

00:20:39.684 --> 00:20:43.565
But what are some things that I should have asked you, that I should have asked that I didn't?

00:20:44.434 --> 00:20:51.760
I think the whole metaverse, crypto, ai there's this whole feeling that it's moving, it's going to combine in some way.

00:20:51.760 --> 00:20:58.442
These worlds are going to be connected, maybe through the metaverse, where all these things will be intact.

00:20:58.442 --> 00:21:03.442
I think that's something, or a discussion I think, it's really interesting to have within these topics.

00:21:03.442 --> 00:21:10.586
How will the future pan out in five years, where we're going to be talking about how we're going to be buying things?

00:21:10.586 --> 00:21:12.617
How will these things come together?

00:21:12.617 --> 00:21:15.306
I think that'll be very hard for you to anticipate.

00:21:15.306 --> 00:21:21.107
I wanted to talk about that, but that's something that came up while reading the show description.

00:21:21.555 --> 00:21:25.265
I'm starting to see some connection points.

00:21:25.265 --> 00:21:36.027
I'm sure there's going to be some DeFi primitive that uses AI or something in some unique way, but that's not the thing that I think is the really interesting thing that's happening.

00:21:36.027 --> 00:21:40.809
The thing that I think is really interesting, that's already happening and I've spoken about this on a previous episode.

00:21:40.809 --> 00:21:51.249
But companies like Render Network, companies like Uproc, who I mentioned previously their startup they have built networks of people, people-powered networks.

00:21:51.249 --> 00:21:54.599
In Render's case, it's about they have GPUs.

00:21:54.599 --> 00:21:59.940
These people are contributing GPUs to a network and then companies or individuals can use those GPUs.

00:21:59.940 --> 00:22:11.561
They can use them for rendering, which is what Render initially started out doing, or it can be used for AI, like models powering these models, because there's a big GPU shortage right now.

00:22:11.561 --> 00:22:20.005
You can have all these people contributing the GPUs from their laptop to solve that problem and provide inexpensive access to GPUs.

00:22:20.005 --> 00:22:27.028
That's all power where the accounting for paying people for that GPU processing power is done on the blockchain.

00:22:27.028 --> 00:22:28.500
It's done on Solana.

00:22:28.500 --> 00:22:40.169
That is something that is really only possible at scale on Solana because the speed and efficiency of that network, and really only possible when you have the incentive alignment to get people to contribute this compute power.

00:22:40.169 --> 00:22:43.538
A Cache Network is another really interesting company that's doing that.

00:22:43.538 --> 00:22:53.625
That's more on the compute resource problem If you have access compute resources or you don't have enough, you can use these networks, these networks, to get access to that resource.

00:22:54.268 --> 00:23:02.702
In Uproc's case, they're actually going after the data problem, because data you have to feed these models, these AI models, these machine learning models massive amounts of data.

00:23:02.702 --> 00:23:06.939
In many cases, getting that data is very hard.

00:23:06.939 --> 00:23:12.116
When OpenAI did that, it was a tremendous amount of resource put into getting that data into the model.

00:23:12.116 --> 00:23:32.878
What Uproc's doing is they're creating a proxy network that is on individual people's phones and that proxy network is used to scrape data from the internet so that you can get real time Data feeds on current data that's on the web into an API that then paying customer can access and use.

00:23:33.484 --> 00:23:40.479
I think that's another really interesting intersection of like blockchain technology, token economic structures and AI.

00:23:40.479 --> 00:23:51.296
So for me I don't have a vision of like how is AI going to be used in that way, like connecting all those tools, but I think it's more how are these tools going to feed AI or contribute resources to AI?

00:23:51.296 --> 00:23:52.625
That's what I see.

00:23:52.625 --> 00:24:00.839
I mean the same thing with even for the podcast editing, like if you need compute resources to do something to that video, you can get access to that via these networks.

00:24:00.839 --> 00:24:09.765
I think demand for renders there and the GPUs, I think, like some of it, is lagging and lagging in demand, but I think it's just going to take a little while to get product market fit for those things.

00:24:09.765 --> 00:24:10.928
So I'm just saying it's early days.

00:24:10.928 --> 00:24:13.895
We'll see how much usage these things actually end up getting.

00:24:14.605 --> 00:24:16.550
Yeah it's just a liquidity problem.

00:24:16.550 --> 00:24:18.496
Someone has to connect the gaps.

00:24:18.496 --> 00:24:31.105
I'm not sure it's going to be how will AI be applied, but how will we live and how will we interface, and I think that will define which technologies will be used to solve that problem Exactly.

00:24:31.105 --> 00:24:39.740
You mentioned with this, with the blockchain, the problem with compute is a hard one, incentivizing all these P2P random people out there sharing their resources.

00:24:39.740 --> 00:24:41.730
That's an interesting observation.

00:24:42.405 --> 00:24:48.480
One question I'm starting to ask people on every show now because I haven't done is a good enough job of like really drilling into this.

00:24:48.480 --> 00:24:49.845
But what is your why?

00:24:49.845 --> 00:24:52.010
Like my why for the show is people are worth knowing.

00:24:52.010 --> 00:24:53.173
Right, that's what I believe.

00:24:53.173 --> 00:24:54.356
I believe people are worth knowing.

00:24:54.356 --> 00:25:07.637
This kind of goes back to like Simon Sennick's work, and I've spent a lot of time last like six months really thinking about like the why behind what drives me in the things that I do, and I'm curious if you have an answer to that what is the why driving the work that you're doing right now?

00:25:08.125 --> 00:25:10.250
I think good stories need to be told.

00:25:10.250 --> 00:25:12.096
I'll be more specific.

00:25:12.096 --> 00:25:27.813
The amount of content that's out there already recorded, already published and not being promoted in the right way, so it's not getting into the right people's feeds I think that's a huge, huge waste.

00:25:27.813 --> 00:25:34.517
In our world, content should be consumed and, of course, the good content should stand out.

00:25:34.517 --> 00:25:43.704
It just always makes me so furious when I see an amazing company, let's say a technology company, doing this amazing technology and their marketing just sucks.

00:25:44.527 --> 00:25:46.770
And it's not because of lack of trying or resources.

00:25:46.770 --> 00:25:53.625
They are trying, they're producing, but nobody's watching this stuff because it's not being presented or repackaged in the right way.

00:25:53.625 --> 00:26:03.005
I think that's a shame, that people can't access these ideas just because whoever created them didn't know the right way to package that.

00:26:03.005 --> 00:26:13.109
So if we can help bridge that gap and help spread this content, we're not obviously creating any content, we're just helping the distribution layer.

00:26:13.109 --> 00:26:19.839
I think that's going to be a very important resource in our world, perhaps even a commodity.

00:26:19.839 --> 00:26:25.217
Content is becoming a commodity being able to have an efficient way to distribute it.

00:26:25.217 --> 00:26:30.396
That's my why I think distribution of good ideas needs to be easier.

00:26:31.125 --> 00:26:32.690
I love it, obviously, I agree.

00:26:32.690 --> 00:26:36.992
I mean it kind of aligns with people worth knowing, people worth knowing, and the story has got to get out there somehow.

00:26:36.992 --> 00:26:38.768
And I think about this all the time.

00:26:38.768 --> 00:26:48.415
Maybe it's just the algorithm, but I find that on social media channels except Twitter Twitter I get all the same people, but not all the same content over and over again.

00:26:48.415 --> 00:26:55.856
There seems to be fresh content there for me, but on Instagram, which is really the only other one that I actually use, it's just the same stuff.

00:26:55.856 --> 00:27:02.230
Like I see almost all the same stuff all the time and I'm like literally the same clips and I'm like how is that possible?

00:27:02.230 --> 00:27:03.255
The same ones?

00:27:03.255 --> 00:27:07.496
Yeah, like literally I'll see the same stuff and I'm like it's like recycling the same clips.

00:27:07.516 --> 00:27:15.050
To me, that's got to be partly the algorithm, but I also just wonder if it's just not enough people producing the kind of content that I want to see, and maybe that's part of the problem.

00:27:15.050 --> 00:27:19.845
In general, I grew through, like our space crypto and the blockchain space has been guilty of this.

00:27:19.845 --> 00:27:26.858
Like, marketing is very hard and it's not a core competency of many of these, like very heavy engineering and product focused teams.

00:27:26.858 --> 00:27:28.347
They need help with that.

00:27:28.347 --> 00:27:30.114
So I think you're definitely doing good work.

00:27:30.114 --> 00:27:33.085
I look forward to seeing how it pans out with our stuff.

00:27:33.085 --> 00:27:43.311
So, if you're listening to the show, listen to this episode, go to our Instagram, go to our Twitter, go to our TikTok and you should see an upgrade in some of the content we've been producing.

00:27:43.311 --> 00:27:46.557
That is going to be done by Tomer and his team.

00:27:46.557 --> 00:27:47.318
So thanks so much.

00:27:47.318 --> 00:27:48.441
Where can people find you?

00:27:48.800 --> 00:27:49.501
We're on Leechie.

00:27:49.501 --> 00:27:51.005
We're on Twitter, we're on LinkedIn.

00:27:51.005 --> 00:27:53.816
You can check us out online at leechieso.

00:27:53.816 --> 00:27:57.484
That's right, we're starting a new cool domain.

00:27:58.066 --> 00:28:06.853
So it's lyceeso for people like me who are spelling challenged lyceeso.

00:28:06.853 --> 00:28:07.734
Thanks for coming on the show.

00:28:07.734 --> 00:28:08.276
I appreciate it.

00:28:08.276 --> 00:28:09.117
Appreciate it, Alex.

00:28:12.627 --> 00:28:16.125
You just heard the index podcast with your host, Alex Kahaya.

00:28:16.125 --> 00:28:24.444
If you enjoyed this episode, please give the show a five star rating and subscribe on Apple, Spotify, Google or your favorite streaming platform.

00:28:24.444 --> 00:28:27.404
New episodes available every other Wednesday.

00:28:27.404 --> 00:28:29.126
Thanks for tuning in.