Transcript 0:00 Welcome back to the Middle Tech Podcast. You got Logan Jones and Evan Knowles recording out of the Sales River studio. Hope everyone had an awesome Fourth of July, got to enjoy some time off. 0:08 Hopefully, you went to a barbecue of some sort, spent some time with your family, maybe watching some fireworks. Uh, we got to go to a pretty cool party here in Lexington, uh- Little rooftop ludicrous action. 0:19 [laughs] I like to say Lexington's leveling up- [laughs]... with that party. It was on top of the Manchester Hotel, which is an awesome new, uh, development here in Lexington. Middle Ground Capital threw a great party. 0:27 Yeah, shout out to John Stewart and Lex Exotics, uh, and my friend Wes Brown, and your friend TJ Burnett for getting us in there. Um, so we've got some awesome stories to go over in this episode. 0:37 We're gonna be talking about Rivian's $10 million investment here in Kentucky. 0:42 Uh, we're gonna be talking about Meta launching Threads, which happened, uh, today when we're recording this, which is Thursday, July 6th, and we're also gonna be talking about Twitter introducing rate limits, which has a lot to do with that previous story. 0:54 So before we dive into those, we just wanna remind everyone to subscribe to the Middle Tech YouTube channel. Uh, follow us on all of our socials, @middletechpod, to keep up with everything going on with Middle Tech. 1:04 And before we dive in, we just wanna get a quick word from our sponsors. 1:07 Before highlighting our sponsors, we'd just like to state that the views and content shared on this platform do not necessarily reflect those of our show sponsors. 1:15 Middle Tech is presented by KY Innovation, the Kentucky Cabinet for Economic Development's Office of Entrepreneurship. 1:21 KY Innovation exists to support and develop Kentucky's startup ecosystem, and we are proud to be supported by an organization whose mission aligns so closely with ours. 1:30 If you're a founder building in Kentucky, you need to check out the resources that KY Innovation has to offer. You can find more information by clicking the link in our show notes or going to kyinnovation.com. 1:41 Middle Tech is sponsored by Bolt Marketing. Take your website to the next level with a website that's built to work. 1:47 At Bolt Marketing, they're revolutionizing websites for small businesses that are affordable, customizable, and hassle-free. 1:53 Whether you have a construction company, a boutique clothing store, or you own a hot yoga studio, they have options for you. 1:59 Click the link in our show notes to explore their marketing options that can transform your marketing and grow your business. 2:04 And as a personal note, Bolt Marketing built our website, and they were awesome to work with throughout the entire process. We highly recommend working with them. All right. 2:12 So let's talk about this Rivian, uh, Kentucky investment. So Rivian announced that they're gonna invest $10 million and create 218 jobs in Bullitt County. 2:21 So specifically, it's gonna be in Shepherdsville, where they already have a large facility that's about 326 square feet, uh, for, uh, remanufactured parts, so basically they're replacement parts. 2:34 Um, so they're gonna expand on this facility and make it, um, an operational, uh, facility when it comes to manufacturing. So they're gonna remanufacture parts, um, and this is a news article actually from April. 2:46 So, uh, it's kind of older news, but, uh, Governor Beshear just put on his socials to share this story and got some good video content with Rivian. 2:55 Um, so this is a pretty cool story because it's another EV company here in Kentucky, and Kentucky's been working very hard to bring, you know, battery manufacturing and EV manufacturing, uh, to our state. 3:05 They've done an amazing job. 3:07 Uh, we're probably currently the leader, uh, in the United States when it comes to both battery and electric car manufacturing, when you take into account that, uh, Toyota is transitioning their factory in Georgetown to manufacture EVs. 3:20 Uh, we're by far the EV capital of the United States. So super exciting. The state of Kentucky and, and Governor Beshear's cabinet's doing an amazing job. Uh, so, um, one more to track. 3:30 So Rivian's a pretty re- uh, innovative company. 3:32 Yeah, and, uh, just because we've talked to one of the, uh, um, senior computer vision engineers on the podcast, I don't believe Angelo's still with Rivian, but wanted to call out that episode. 3:43 Uh, Angelo Stikardus, who was a University of Kentucky graduate, uh, he worked at Rivian for a while, and I saw him post on his story- He's still there. He is still there? Yeah. Okay. That was my mistake. 3:52 Um, so he's still there. I saw him post on his story that, uh, he was excited to see Governor Beshear calling it out. 3:57 Um, so very exciting to see Rivian putting some money into Kentucky, and Kentucky continuing to position themselves as an electric vehicle, uh, manufacturing capital. 4:05 So, um, moving on from that story, let's go ahead and talk about, uh, what everyone's been talking about, uh, today at least, which is Meta launching Threads. 4:14 Uh, so Threads, which has kind of been dubbed as a big competitor for Twitter, uh, officially went live today, which again, uh, is Thursday, July 6th. 4:23 Um, so it came out as kind of like a pre-order on the App Store, and then when they officially launched it, you could download it today. 4:30 Uh, so let's first start by just establishing what Meta has said about Threads, what they want it to be. 4:35 Uh, they have dubbed it as, uh, the vision being that Threads will be a new app more focused on text and dialogue modeled after what Instagram has done for photo and video. 4:46 Uh, so this launch came earlier than expected, um, likely due to Twitter, uh, enacting rate limits. We're gonna talk about, uh, that in a l- in a little bit, uh, why Twitter enacted those rate limits. 4:57 But it essentially made it to where, uh, users of the platform could not consume as much content, um, and that just provided a perfect, uh, perfect point for Meta to launch this new platform. 5:10 So I went ahead and created an account on Threads. Incredibly easy user experience. Have you gone and created an account yet? Uh, it's-- So it's connected to Instagram. It's connected to your Instagram account. 5:20 It took me maybe, uh, 10 seconds to actually get set up and follow the same people that I was following on Instagram, which is a huge deal how easy they've made it, uh, to create an account. 5:30 Um, and that is gonna lead to a pretty high conversion rate, uh, in, in my opinion. 5:35 So they, they would only need to convert 17.5% of Instagram's two billion monthly active users to Threads for it to be as large as Twitter, uh, with Twitter having around 350 million monthly active users. 5:48 Uh, and in the first seven hours of launching Threads, it got to 10 million users. So- They're at 30 million today. They're at 30 million already? Yep. 5:55 Okay, so that was after seven, seven hours of launching it, they got to 10 million. 5:59 By the, uh, by the time of this recording, they had 30 million users on it, so I don't think there's any doubt that, uh, that Meta is gonna end up, um- You know, matching the number of users on Threads is what Twitter has right, right now. 6:11 Um, a couple comments on this. I wanna hear your thoughts on this as well. First off, we've seen this whole thing [chuckles] play out several times now with Meta. We saw it back in 2016 when Instagram introduced Stories. 6:21 They basically copied that feature, uh, from Snapchat. We saw it again in 2020 with TikTok. They copied that feature... Or I'm sorry, with, uh, with Reels, they copied that feature from TikTok. 6:32 Now we're seeing it again, uh, with, uh, you know, Zuck seeing that Twitter is at a vulnerable moment with the turbulence that Elon Musk has caused. 6:39 Now they're introducing essentially a whole new platform to try and take on Twitter. Uh, so I wanna pause for a second. I wanna get some of your thoughts on it. 6:45 I have a few more thoughts, but, uh, off the bat here, what are you thinking of, of Threads so far? Yeah. 6:51 Um, well, I think they definitely have a great chance of, you know, becoming, uh, the platform where there's more users. Um, I think, you know, I think, um, it's gonna be interesting to see if the developer... 7:04 You know, Twitter right now is, is heavy, um, developer community. You know, the tech community's on there. 7:10 Political community is by far mostly on Twitter, so we'll see if that moves to Facebook and those kind of core demographics. 7:18 Um, another thing that, that concerns me is, uh, just the amount of open sourcing Twitter has done. You know, I'm not sure we're gonna see the same from Facebook. 7:28 I value that a lot 'cause, you know, trust and safety and, you know, free speech is something that, you know, I really have enjoyed Twitter, uh, implementing. Um, so we'll see what happens there. 7:37 But, uh, Elon actually is suing, uh, Facebook, uh, over this, and so he's claiming that Facebook hired ex-Twitter- Meta... engineers and is now cloning Twitter with those engineers. 7:50 Um, and so they're also claiming, uh, that they scraped their data. 7:53 Um, so there's several theories out there, and I think I've seen Elon actually engage with some of these theories, where essentially they're saying that Elon introduced this rate limit because Facebook or Meta was, uh, basically scraping the social graph of Twitter, and so basically that made it easier for suggesting who to follow on Threads. 8:14 Um, so that was interesting. Now, another th- another thing I saw on Twitter that was really interesting was, um, Facebook would not be allowed, Meta would not be allowed to acquire Twitter. 8:25 The government would not allow that. Mm. But all of a sudden they can launch a clone and it's okay. Um, so that, that's a big, I think, issue, so we'll see what happens with this lawsuit. 8:35 Um, I don't know if there's a case or not. You know, I think that the antitrust stuff is, um, case by case. It's hard to tell these days. Um, but I thought those were some interesting comments. 8:45 Um, I would probably prefer, um... I'll probably use Threads. We'll see, 'cause I, I've seen a lot of cool integration with Instagram, so you can put your Thread, you know, on your Instagram story. 8:56 But, um, yeah, again, I think I value the kind of openness and trust and, and free speech of, of Twitter and the transparency. Yeah. I hope Twitter continues to differentiate itself with that. My fear is that, 9:10 uh, it's gonna start getting split down political lines between Twitter and Threads. Um, you'll start seeing... 9:16 I mean, you've already seen kind of a big movement of right-wing people moving onto Twitter and hosting a lot of their content on Twitter. 9:22 I hope that it doesn't become split as, like, Threads is where, you know, left-leaning people go and then Twitter is where right-leaning people go. 9:29 I think that'd be a shame if we created, like, thought bubbles within platforms. That would, that would not be good. Um, I will say, I don't think Twitter is just gonna die because of this. 9:37 I've, I've seen some conversation around like this is the Twitter killer. 9:40 It's certainly putting Twitter on high alert, which I think is a good thing, but if you think about all the other companies that, uh, Meta copied the features of, Snapchat is still around, TikTok obviously, uh, a huge company. 9:52 Um, I'm not saying that's gonna be certain for Twitter, but just because they have, you know, essentially copied a lot of their features doesn't mean they're gonna instantly die. 10:00 Um, but at the end of the day, I'm kind of along with Evan here. I really don't want Meta to just be able to dominate a large percentage of social media usage. I think that that's just a lot of control. 10:09 I think that's dangerous. And also from a data perspective, they already have so much data on us from our activity on Instagram, our activity on Facebook. We've seen that go awry with the Cambridge Analytica scandal. 10:21 I'm not saying that all of that stuff is gonna happen again, but if you give a company access to that much data, uh, and that much insight into our, our social media activities, I just think it's a dangerous thing. 10:31 Um, and I also am very much for open sourcing the algorithms and, uh, the free speech measures that Elon Musk is taking on Twitter. 10:37 So, um, I think now is probably a good point to move on to, uh, kinda why this was a good time for Meta to launch Threads, which was Elon introducing, uh, these rate limit, uh, this rate limit, um, thing to, to Twitter. 10:51 And so a lot of people are speculating about r- why this happened. I think that it's pretty clear what's going on, and I think it's the reason that Elon, um, sided. 11:00 So Elon is siding that, uh, these AI companies are essentially scraping Twitter, uh, to collect content to train their large language models with, uh, which then causes Twitter's servers costs, server costs to skyrocket. 11:15 Uh, so they have now, uh, introduced rate limits, which essentially just limits how much content a user can consume. Um, and so what that actually looks like, uh, is for, uh, for an unverified user- 11:28 Let me make sure I get the, uh, the numbers right here. Uh, for an unverified user... Well, first off, users cannot view tweets without logging into the platform. So you have to have an account to be able to view tweets. 11:38 Unverified accounts can only view 600 posts right now. New unverified accounts can only view 300 posts, and then verified accounts can view 6,000 posts. 11:48 Uh, he's already said that the limit will soon increase to 10,000 for verified, 1,000 for unverified, and 500 for new unverified. 11:55 But I wanna pass this off to you, Evan, for the second point of why this is happening with, uh, in regards to, to scraping and bot activity on the platform, uh, because I think that's a really important piece of this as well. 12:05 Yeah. 12:06 I mean, for the AI piece, you know, I think that what's gonna happen with the internet is that, uh, the battle, uh, on kind of the m- the macro level within the tech space right now is who can train the best foundational model, uh, which is basically what OpenAI and Google and Anthropic with, uh, their Claude model, um, have created. 12:27 That's called a foundational model, which is kind of what all these-- a lot of the AI companies will build on top of. 12:32 Um, and the battle is over these foundational models which are scraping, you know, the open internet for this data to train. 12:39 And I think one of the big battles going forward is, um, companies protecting their data from being within these foundational models, um, so they can either have a competitive advantage for their own foundational model, foundation model, um, or because they claim that they own that content. 12:54 With social media companies, when you put something on their platform, you know, they, they own that content. They get to put ads on it. They get to do what they want. 13:02 Um, so that's one thing that I think I wanted to add there. Now, one of the other bits of information here, um, that's out there related to this rate limiting is controlling bots. 13:11 So we know that bot activity has been a huge issue. 13:14 Uh, people are able to put, you know, bot farms and, you know, bot armies out there and change dialogues and engage with tweets, which changes, uh, where they show up in the algorithm, and they can basically flame a fire, um, and expand, you know, how many people are viewing content with these bots. 13:31 And so the theory is that essentially they created these rate limits to figure out who's a bot and who's not, and then if it's a bot account, they flag it, and they can remove those accounts, um, or just figure out a way to stop their activity. 13:44 So that's what these limits are for. Um, these bots are, you know, viewing a lot of tweets, and they're engaging with a lot of tweets in a very short period of time, and these rate limits prevent that. 13:54 It's very important that this happens, and other platforms like Threads will have to do this, and if they don't, they can't be trusted. 14:02 I think that's a, a great point, and I wanna go back to the AI thing just real quick, just because it, it kinda spurs a thought for me. 14:08 Does that give companies like OpenAI even more of a head start than they already had? 14:13 Because it, it sounds like it's now gonna get much more expensive and much more difficult to train these large language models if all these companies that host so much of the internet's data, like Reddit. 14:22 We talked about this story a couple weeks ago, uh, Reddit now charging for their API, making it much more difficult to, uh, to get a lot of this data that is used to train these models. 14:31 I'm wondering if that's gonna further isolate some of these early movers that have been able to get these large language models, you know, Meta being one of them. 14:38 They have tons of, tons and tons of access to all of this. Well, I think that's a good intro into the next story here with AI Edge. Well, let's go ahead and get into the AI Edge. 14:46 Uh, so the AI Edge, uh, [chuckles] the AI Edge segment, uh, stories, tips, and topics for gaining an advantage with artificial intelligence. This story has to do with OpenAI and ChatGPT. 14:56 If you had ChatGPT Plus, you might have noticed that you no longer have access to browsing, uh, if you're using ChatGPT-4, which, uh, which is tragic. 15:05 It was an awesome, [chuckles] awesome feature that I used all the time. Um, which, you know, hopefully it'll come back. But tell us what's going on. Tell us about the lawsuit here. Yeah. 15:14 So, uh, OpenAI basically put out the beta for, for browsing. They made it very clear this was a beta. And what ended up happening was the browser was able to go scrape and basically navigate the internet. 15:23 And several users brought to the attention of OpenAI on Reddit that, "Hey, I asked your ChatGPT model to go find a mar- an article that was behind a paywall and show me that article," and it did it. 15:36 And so it got around the paywall. There's a couple theories on how it did it, but the big point here is that it was able to navigate the internet and get behind paywalls without there being a transaction there. 15:46 Um, so paying a company like The Atlantic, for instance, that has a paywall on their articles, it was able to get an Atlantic article and share it with the user on ChatGPT without, you know, them paying for that article. 15:56 Um, and so that is clearly a copyright issue, um, and that's not legal. And so they're stopping, you know, this browser and figuring out how to solve that, um, because there's gonna be lawsuits. 16:07 There's already several lawsuits, like I said. Twitter is suing, many companies are suing OpenAI, um, over this data scraping and over basically the ability for them to get access to personal information as well. 16:19 Um, so there's several class action lawsuits out of the state of California, uh, against OpenAI because they're claiming that OpenAI is collecting personal identifiable information and putting it into their model. 16:30 Um, and so part of that scraping also brought privacy concerns. Um, so OpenAI is just growing super fast. 16:37 Uh, these, uh, large language models and these kind of AI agents that act autonomously on the internet doing things like browsing and, uh, other tasks, um, you know, they're really hard to control how, based on how fast these companies are moving. 16:50 So things are breaking. You know, the whole adage within Silicon Valley is move fast and break things, um, is happening right now with OpenAI and all these large language model companies. 17:00 And so there's gonna be, you know, a lot of lawsuits. And so we'll see what happens here, um, over the next several months. Yeah. 17:07 I feel like this whole AI movement is just totally gonna change the entire infrastructure of the internet and how we consume content. We've been saying that for a while. This makes it a little bit more apparent, I think. 17:16 A thought experiment, and there's probably not an answer to this, but I think it's an interesting thought, is couldn't somebody essentially just pay for all of the publications for, you know, let's say Wall Street Journal, Kentucky Kernel, whatever, whatever else you want to add in there... 17:29 connect a, a large language model to it and just basically say, "Hey, you subscribe to this large language model, you can get these articles rewritten." 17:38 It's the same information, but it's just rewritten content, and then just you have access to all of the, all of the publications. The way I would think it would play out is somebody's not gonna wanna pay for all those. 17:48 What they're gonna do is microtransactions, so they're gonna have to figure out how to hit an API and call on an article, and whenever that call comes back with that article, they're gonna be-- there's gonna be a microtransaction there. 17:58 Like an API call? Yeah. And so as a user, you're paying for a subscription plus some kind of microtransaction when you're asking it for an article. Yeah. I was more so thinking if there's, like, a layer in between. 18:08 Like, somebody starts a company that's the large language model that just gives summaries of, you know, let's say 10 different paywalled publications. 18:17 You just pay for this large language model access to it, and it's recreating the same content, but it's, it-- theoretically, you would think it would get past copyright laws because it's just recreating it. 18:26 I would probably say that the content creators are probably just gonna create plugins. Yeah. Just a thought. 18:32 Um, have you been doing anything cool with AI recently that you would like to, uh, give up as a tip of the week? Um, no. You're gonna pass, huh? You know, I think I'm tinkering with a concept of, um... 18:43 You know, one of our previous episodes, we talked about using it to, uh, basically proofread and, and pick out biases in your writing and help you improve as a writer and, and figure out, okay, um, you know, um, before I put this out, if it's a blog or I'm writing something for my, you know, company content, before I put this out, let me make sure that there's no, you know, biases that can get me in trouble, um, or, you know, sway the reader's opinion. 19:08 'Cause as a writer, uh, especially in the, in the business realm, you don't necessarily wanna do that. You wanna just put out, you know, neutral information. 19:14 Um, so one thing I've also been tinkering with is along those lines of, um, basically, can you use these models to, uh, help you with decision-making? 19:23 So whenever you make a decision, um, you have a thread that has memory in it, and you put your decisions in there. 19:29 And as you make more decisions, it learns your decision-making patterns, and then it helps you, you know, continue to make decisions. 19:34 So that's one thing I've been tinkering with in the models and figuring out if I can fine-tune, um, one of those to do a better job of that. 19:42 So that's just a, a thought that people, uh, listening can use ChatGPT for is, you know, putting your decisions in there and just having another perspective on those decisions to, again, find biases or, you know, have the memory of your previous decisions so you can continue to improve. 19:56 Now, I don't know if a normal consumer would do that, but it's a very interesting, uh, business use case. Little Easter egg there. I love it. Um, all right. 20:03 Well, wrapping up the, uh, Edge segment, just wanna give a quick shout-out to the Spotlight series that we have coming out on Wednesday. 20:11 Uh, so we're gonna be releasing our Spotlight series with Randall Stevens of Avail, uh, actually one of the companies that Evan worked at on his, uh, entrepreneurial journey. 20:19 Uh, so what year were you with Avail, back in, like, 2020, 2019? Yep. Uh, 2019. Yeah, which was around probably the last time that we actually talked to Randall. 20:27 So we sat down with, uh, Randall again to just get an update on his company, uh, how things have changed, where things are going. So be sure to tune into that one. 20:35 And if you did not have a chance to listen to last week's Spotlight series, would highly recommend that. That was with JP Blevins of Live Oak Bank and nCino, two hyper-successful companies. 20:46 Uh, JP recently came back to Lexington, uh, and became friends of, uh, of both Evan and I, uh, and his story is just incredible. 20:53 Um, it's somebody that was an early employee at a company that managed to take that all the way to being a publicly traded company. 21:00 Uh, and it's just a really cool story for maybe you don't think you're gonna be a founder, um, but you wanna be an early employee. This is a great success story of somebody that took that exact path. 21:08 So be sure to check that one out as well. Other than that, thank you guys for tuning in. We hope you've learned something today. We will see you next week. 21:16 [upbeat music]