Big Tech Rips YouTube For AI, OpenAI's New AGI Rules & the Shifting Political Landscape of AI

AI4H EP0067
===

[00:00:00] Apple, Amazon, NVIDIA, basically all of big tech is being accused of ripping off your favorite YouTube stars. We'll tell you how this happened, how your favorite creators are reacting, and whether or not you got sucked up into the machine. Don't worry, AI for Humans fans, we're safe! Turns out, we're not popular enough.

Then, new information is dripping out of OpenAI. They've unveiled the levels to AGI, and guess what, Kevin? It's strawberry season. What does that have to do with anything? Well, everyone's kind of freaking out about it. We're going to tell you why. Also, AI has an NFL bias? Plus, new AI shorts you can't miss!

And, a very special shout out to all of our crust nudgers. It's AI for humans! Nudging crust.

Okay.

gavin_2_07-17-2024_112628: Welcome, welcome, welcome everybody to AI for humans, your weekly guide into the wonderful and wild world of artificial intelligence. I'm Gavin Purcell and my cohost is Kevin [00:01:00] prayer. We are here to demystify the wild world of artificial intelligence and boil.

gavin_2_07-17-2024_112628: Boy, Kevin, this week, do we have some fun stuff to get into?

kevin_2_07-17-2024_112628: We have an amazing guest coming on, Gavin. You and I are fans of their newsletter, the Neuron's Noah Edelman is going to join us to answer, of course, your favorite question, is AI going to destroy slash enslave all of us?

kevin_2_07-17-2024_112628: But then we're going to get into the show. The industry, the tools as they exist, and some big future casting predictions.

gavin_2_07-17-2024_112628: and also what it takes to grow a newsletter in this day and age, which is not easy. And Noah and his partners have been very successful at it. So I'm excited about that. Before we move on though, please, everybody, if you are listening to the show right now, leave us a five star review on Apple podcasts. We will read it there, but more importantly, just share this with people.

gavin_2_07-17-2024_112628: You know, there's definitely been people in our audience that I've heard. That are talking about our show. I feel Kevin very lucky because we have heard there are a lot of influential people listen to the show. If you are a CEO of a major media company, we know you're in here and you're listening to us talk about this stuff.

gavin_2_07-17-2024_112628: So thank [00:02:00] you so much for 

kevin_2_07-17-2024_112628: Why did that sound so threatening? If you are someone important, , if you're a captain of industry,

kevin_2_07-17-2024_112628: we're on to you. We can sniff sniff ya. 

gavin_2_07-17-2024_112628: What do you think of this word? Rosebud.

kevin_2_07-17-2024_112628: How

kevin_2_07-17-2024_112628: about this word, Gavin? Patreon. 

gavin_2_07-17-2024_112628: Patreon.

gavin_2_07-17-2024_112628: That's right. If you're listening to this show and you want to find a way to support us, you can do one of two things. Leave us a five star review, which is a free way to support us, or subscribe. Go join our brand new Patreon.

gavin_2_07-17-2024_112628: We launched a Patreon specifically because we want to be able to level up the show. Thank you. Those of you thank you to those of you who have already joined. , we have some fun stuff coming and as we have tease, there will be a discord. Kevin and I are in the middle of some crazy stuff, travel right now to go do some speaking gigs, but it will be fun.

gavin_2_07-17-2024_112628: I'm excited to have a discord for our fans as well.

kevin_2_07-17-2024_112628: Is it a glorified tip jar at the moment? Yeah!

kevin_2_07-17-2024_112628: Yeah. It is. And thank you. But

kevin_2_07-17-2024_112628: we are gonna attend to it. But there's so much to get to, Gavin. As much as I would love to talk for an hour and a half about just you and [00:03:00] I. There's a whole wide world of what out there? Whoo!

gavin_2_07-17-2024_112628: time for the news.

gavin_2_07-17-2024_112628: Kevin, there was a blockbuster story this week that came out of a new news organization called proof news. org. And basically what they came out with was clear proof that many famous YouTubers and obviously big brands had been used to train large AI models.

gavin_2_07-17-2024_112628: And this is not just some kind of smaller open source AI model. This is a model that has been used by some of the biggest names in tech. Apple, Nvidia, Anthropic also used it. 

kevin_2_07-17-2024_112628: you're, you're dropping big alpha, you're, you're

kevin_2_07-17-2024_112628: mentioning big tech. I'm imagining six laser pointers dialing up on your chest right now because you have cats. This is a big [00:04:00] deal and it's

kevin_2_07-17-2024_112628: Aluther AI, a non profit AI company. Fun fact, Gavin, Aluther made something called GPT J eons ago, which was like an open source freely available GPT 3 clone that really spurred a lot of AI development.

kevin_2_07-17-2024_112628: I was such a huge fan of what they did and they did something called. The pile. And the

kevin_2_07-17-2024_112628: pile has infected the likes, like you said, of Apple, Anthropic, Nvidia, Amazon. I didn't say it. Gavin did. But what does that mean, Gavin? What

kevin_2_07-17-2024_112628: is 

gavin_2_07-17-2024_112628: Well, let's, yeah, 

kevin_2_07-17-2024_112628: And why is it so filthy?

gavin_2_07-17-2024_112628: let's take a step back quickly. So the pile is a data set. And, and for those of you listening or watching to that, know what a data set is, congratulations, but a data set is a bunch of stuff put together that an AI model can then train on. And in this specific instance, one of the big parts of the pile is a whole bunch of YouTube transcripts that were put into this, as part of a larger thing that included. research papers included like novels, some copyright free novels, other stuff as well. the [00:05:00] thing that Proof News did this time, which is really interesting is they created a little tool that allows you as the YouTuber to go and see if your YouTube videos were included.

gavin_2_07-17-2024_112628: In this data set called the pile. And, and Kevin, before we get into specifically who was kind of looped into this data set, I think it's an important thing to know that like we have heard about the pile for a long time, right? The pile has been a data set that people have used for a ton of AI models because it was something that was available early and easy.

gavin_2_07-17-2024_112628: And yes, this, company Luther dot AI. Put it together as a chance to kind of like jumpstart the AI training space. 

kevin_2_07-17-2024_112628: There's a lot of people are familiar with the common crawl, which is another similar data set where they just went out to the internet, they scraped what they could every little bit and they made it available to folks and there was some questionable material. I mean, illegal material in these data sets, they had to go back and find and prune, but we're talking 170, [00:06:00] 000 transcripts from very popular YouTube creators. Big brands, big shows, even some that you've worked on, Gavin,

kevin_2_07-17-2024_112628: they all got swept into this data set. And now the companies are saying the apples and even the Googles and videos of the world. , some of them have yet to make a statement, but those

kevin_2_07-17-2024_112628: that have are basically saying, Hey, we didn't do this. No, no, no, no, no.

kevin_2_07-17-2024_112628: That was, that was them over there. They

kevin_2_07-17-2024_112628: grabbed the data. We just pulled it into our machine and you can't blame us. It's it's widely available.

gavin_2_07-17-2024_112628: You're mentioning this show that I worked on. Yes, I worked on the Tonight Show for a long time and there are Tonight Show clips on here in this data set, but almost, almost crazier to me is the fact that a lot of very famous YouTubers are in this data set. And this has been a thing that YouTube, the company has wondered about specifically who's been training on their videos.

gavin_2_07-17-2024_112628: And very famously. Mira Mirati did not answer a question from our friend, Joanne Stern about this idea, [00:07:00] but that was a specifically an open AI question. This is a much broader data set. And, and Kev, I just want to call out Marquez Brownlee, who were both giant fans of, and really is like probably the pinnacle of like earnest, awesome people who have been true to themselves 

kevin_2_07-17-2024_112628: He's a patron of ours. Of course. 

gavin_2_07-17-2024_112628: Yeah, that's right Marques. Thank you for donating. Thank you. I'm sure you're a patron of ours. Marques specifically called out on his ex handle. He said fun fact I pay a service by the minute for more accurate transcriptions for my own videos Which I then upload to YouTube's back end so companies that scrape Transcripts are stealing paid work in more than one way not great So it's clear that youtubers themselves are starting to feel very frustrated by this so this really gets to the heart of what we refer to as the original sin of AI, which is the dirty data sets that kind of began the entire AI [00:08:00] explosion.

gavin_2_07-17-2024_112628: How do you think we proceed from here? Because it could elute their AI get, sued and could they go out of business and whatever their business is? Sure. Sure. But how much legal ramification are we going to be looking at based on this,

kevin_2_07-17-2024_112628: yeah, look, this is nothing new. AI artists everywhere are screening. This is exactly what we went through. Welcome to the party, YouTube creators, right? When

kevin_2_07-17-2024_112628: they took artwork from DeviantArt or even, uh, forums and scraped it all into their data sets. So I think ultimately these cases are going to end up probably in a Supreme court. At some point in our

kevin_2_07-17-2024_112628: near future, right? Some of them will be settled along the way. Some of them will become licensing deals along the way. Some businesses are going to go away. Some individuals are going to get in trouble and have to pay hefty fines. I don't know if Google goes after a Luther for breaking their terms of service on this data set, because I'm also willing to bet at some point this data set got swept [00:09:00] into some of, you know, Google's

kevin_2_07-17-2024_112628: LLMs, like everybody's sort of playing in the same arena and playing with the same data.

kevin_2_07-17-2024_112628: And so, you know, if Google goes after them for YouTube, well then Google might get got after for scraping meta. Like who knows what goes on in those shadowed halls. But yeah, look, eventually a court decides. If it was available, if you and I, and I mean literally you and I, or anybody listening to this could go and freely, easily access that data. Why is that any different than the machine doing it?

kevin_2_07-17-2024_112628: And I'm not trying to argue why it is

kevin_2_07-17-2024_112628: different because that's a whole other conversation, but this is a question that's going to be answered. Probably in a court of law, millions of dollars will be paid out at some point. Some companies may go away, but it's certainly out of our hands 

gavin_2_07-17-2024_112628: Our next story, I think we got to do a little bit of like a, a siren warning here. We are going to briefly talk about politics and we are not going to talk about it in the way that hopefully will be triggering to people.

gavin_2_07-17-2024_112628: But we are going to be talking about politics because this week there are [00:10:00] three quick headlines in the political realm that we think are important to discuss around what's happening in AI. And obviously, politics is a fraught conversation right now. There's been some very crazy and disturbing things that have happened over the last week in the political world.

gavin_2_07-17-2024_112628: But, I think for us, these are important things to make you aware of and understand why they are important. 

kevin_2_07-17-2024_112628: Let me put on my tight rope walking shoes, Gavin, because we're not here to talk about our opinions on things with these particular stories or what ponies we have and what races, but the fact of the matter is things are happening right now, which might have deep ramifications for

kevin_2_07-17-2024_112628: every single person listening

kevin_2_07-17-2024_112628: with regard to the AI they use and big tech in general.

kevin_2_07-17-2024_112628: So at the Republican national convention, The nominee for vice president was announced. It is a gentleman named JD Vance. And the reaction in the AI bubble was confetti cannons and, you know, dropping mission accomplished banners not to get political on top of political, but the AI [00:11:00] maxers are thrilled.

kevin_2_07-17-2024_112628: So Gavin, what is an AI maxer and why are they so darn happy about this?

gavin_2_07-17-2024_112628: Well, I think this is all three of these stories are going to kind of show a slight shift of where Silicon Valley has kind of gone in the last, two to five years, but I think there's been a real conversation in the AI space around the regulation versus deregulation specifically of smaller AI models, meaning that the opening for.

gavin_2_07-17-2024_112628: Like say startups or entrepreneurial companies doesn't get stunted by regulation that says that, you know, only the bigs can be able to access this. They see this pic of J. D. Vance as somebody who's going to be more available and open to their needs and wants. they specifically are pointing to a Vance tweet from March where he said, there are undoubtedly risks related to AI. One of the biggest, a partisan group of crazy people use AI to infect every part of the information economy with left wing bias.

gavin_2_07-17-2024_112628: Gemini can't produce accurate history. Chat GPT promotes genocidal concepts. The solution is open source. [00:12:00] So this is like speaking directly to the VC AI world that really wants to not restrict what AI can do and say, and I think particularly this is again, part of the movement of Silicon Valley, which, you know, famously was kind of left leaning to more of a center right position because of the way the politics have shifted. So obviously this is a scenario where they see this candidate as being more beneficial to them. You know, do you think that's the case? Do you think that like, this is going to be that big of a difference going forward? 

kevin_2_07-17-2024_112628: Looking at his background, , moved to San Francisco principal at Mithril capital. That's a, a long term Peter Thiel VC fund, raised millions of dollars for his own funds. I think that in the tech set, which is the AI set right now, that's a clear signal that this person is going to understand the needs of the venture capitalists of wanting to make sure that there's a healthy startup scene around this tech and not Only one of five [00:13:00] companies gets to own it and run with it.

kevin_2_07-17-2024_112628: He spoke out against regulating it saying that, Oh, isn't it, I'm paraphrasing heavily, but isn't it interesting that when a company gets a firm grasp on a tech, that's when they say, Hey, you should probably regulate this thing to stop others from getting it. So look, there are signals there, which have me understanding why. this community would get excited and would get on board for his pick. Some would say the masks are off. Others would say, look at the flares in the sky. Everybody is lining up. I think in the wake of this pic specifically, 

gavin_2_07-17-2024_112628: The other one of the other big stories that came out of this is that Mark Andreessen and Ben Horowitz, the people that started A16Z, Andreessen Horowitz, the VC capital fund, have come out and basically thrown their support behind that ticket, Now, everybody understands that Mark Andreessen specifically has had a shift of Political tone over the last five to 10 years. And we'll link this podcast that they did where Mark Andreessen talks about his background in, you know, working with Bill Clinton in the nineties and what it was like to start the internet. [00:14:00] from this one hour podcast, Kevin, there's a really interesting clip that Mark starts to drop about what his interactions with the current, administration were around AI regulation. 

kevin_2_07-17-2024_112628: One of the things we argued in our meeting with the White House, um, on AI policy was, you know, look, there are going to be, there are going to be issues that come from AI, but let's, they should be regulated. The regulation should happen at the application level, not at the technology level. Yeah. Um, and the reason for that is, is, is, is basically because, like, to, to regulate math, right, at the, at the, at the level of the algorithms, like, to regulate And he argued with me when I said that.

kevin_2_07-17-2024_112628: Well, so, yes, so, so, so Ben basically said, look, AI at the technology level, you're regulating math, and of course, we're not going to do that. Like, that doesn't make any sense. And you'll recall that what they said. Was no, actually, we can classify math. We can classify math. And literally this was, this is verbatim, this is verbatim.

kevin_2_07-17-2024_112628: This is, this is, uh, we, we did, we, we clap, we classified a whole entire areas of physics, uh, within the nuclear era and, and, and made them state secrets like of the, of the like theoretical physics, science, physics, uh, we totally classified them and made them state [00:15:00] secrets. Um, and that research vanished. Um, and we are absolutely capable of doing that again for AI.

kevin_2_07-17-2024_112628: We will classify any area of math. Yeah, that we think is leading in a bad direction and it will, it will end. So the distinction there, Gavin is, Hey, regulate AI. We're not, we're not anti regulation. It just matters where you target that regulation. They think the, the biggest bigs should be able to go out there and open source as well and cobble together all of the GPUs that they want and swallow up all the data that they want so they can train these massive foundational models.

kevin_2_07-17-2024_112628: And they're saying, Hey, Don't put handcuffs on that. That's just math. Don't regulate that. Regulate what people do with that math.

gavin_2_07-17-2024_112628: yes, and I will say if you listen to the whole episode one of the things they follow up with is the idea that We're in an environment now where it's very hard to regulate math. we are in a world where the math that is out there is going to continue to get worked on by other people in other places. I mean, we all saw Oppenheimer, like that was a very small controlled space. The internet didn't exist. The way information flows didn't exist. there's one more quick story, which is there was a [00:16:00] leak of a, an piece of the Trump administration's new policy, what they say they might do with AI going forward. and specifically in this, they talked about the idea that there would be a Manhattan project to kind of be first in AI, an American policy to really kind of push this forward. So you can feel all of these things kind of coming together in one place. and again, we are not, this is not a political podcast. We are not here to, uh, talk about the significant issues going on in the 

kevin_2_07-17-2024_112628: But if we

kevin_2_07-17-2024_112628: had to, Gavin, which side would you choose? Let's go ahead and, uh,

gavin_2_07-17-2024_112628: I'm not, listen, this is everybody's choice is their own thing and it's a personal thing.

gavin_2_07-17-2024_112628: I do think the important thing to be aware of here is that politics is a multifaceted conversation. And there are things in these conversations around AI that I think I personally do agree with. But there is a larger conversation about how the American political system works and what's going on that.

gavin_2_07-17-2024_112628: Is so outside of the realm of what we talk about here. Again, I think it's just [00:17:00] important for everybody to kind of be aware of the AI elements of this that are being talked about. 

kevin_2_07-17-2024_112628: As I think about it, I think you can regulate math. You can also regulate the amount of power required for something like this. You can regulate, um, what sort of nutritional or warning labels or transparency goes into a model of a certain size.

kevin_2_07-17-2024_112628: There are plenty of ways that you could regulate things. I don't know if we should or not, but I do know that just because we regulate, it doesn't mean someone else. Is going to regulate it. Um, Gavin, at the top of this, you mentioned it's strawberry season,

kevin_2_07-17-2024_112628: which sounds delicious and

kevin_2_07-17-2024_112628: delightful. Makes me want all of the smoothies, but what the hell are you talking about?

gavin_2_07-17-2024_112628: Okay. So opening eye, which has been incredibly quiet They did have a couple of leaks this week. And one of the big leaks that got a ton of press, coming out of a Reuters story originally, and then a bunch of people picked it up is that they have officially named internally, at least their next model strawberry, which You know, Kevin, there's GPT 4, there's GPT 4 0, and now there's strawberry.

gavin_2_07-17-2024_112628: [00:18:00] Like that sounds like a very different change in a lot of ways. The background on this from what we understand is that this is kind of the Q star reasoning model that had been hinted at way, way back when, during OpenAI's drama where Ilya Sutskover left the company.

gavin_2_07-17-2024_112628: Now we don't know that for sure. These are just all rumors, but the idea that there is a reasoning model coming out does lean into what also leaked out of OpenAI. Was there five levels of AGI framework that they've built and Kevin, internally, the rumors are that they are getting very close to level two, which is the reasoning level. And maybe it might be worth reading for, uh, our listeners and our viewers what the five levels are. 

kevin_2_07-17-2024_112628: yeah. So level one, which we are very clearly at as open AI defines it. Level one is chatbots. These are AIs that can have a conversation with you, uh, in conventional language, basically level two, which would hint to this strawberry thing is reasoners, human level [00:19:00] problem solving. So it's not just, I ask you a question, you give me an answer.

kevin_2_07-17-2024_112628: It's I ask you a question and you go reason about how to

kevin_2_07-17-2024_112628: give me the best answer or give me the solution. Uh, then we have agents. Which are A. I. S. That can go out and do things do tasks on their own behalf or on your behalf and then four and five are really interesting. Gavin number four. Innovators, and this is AI that can aid invention, which right, you know, the common complaint about most AI now, and is that it's just a robotic hand that goes into a big recycle bin

kevin_2_07-17-2024_112628: and rummages around and then bolts existing things together.

kevin_2_07-17-2024_112628: It's not innovating. It's not creating. 

kevin_2_07-17-2024_112628: And then level five, lastly, is organizations. This is AI that can do the work of an organization, which sounds to me like you unlock inventions, and then you agentify the invention AI. And just, it's basically that AI at scale, because it's the ability to say, Hey, I want a business that makes insert templates for Etsy.

kevin_2_07-17-2024_112628: Whatever that idea [00:20:00] is and it goes out and goes, okay, what do we need? We need to apply for a business

kevin_2_07-17-2024_112628: license. 

kevin_2_07-17-2024_112628: We 

gavin_2_07-17-2024_112628: imagine what, yeah, what would, what would AI be able to do at Etsy? Like imagine like it makes like little baby hats for robots, like things like that. Cause that's like, 

kevin_2_07-17-2024_112628: of adorable.

gavin_2_07-17-2024_112628: maybe that could be merch of ours. We could have little baby robot hats. 

kevin_2_07-17-2024_112628: Little like, yeah, like, uh, like, uh, what's a Skynet onesie or like a little.

kevin_2_07-17-2024_112628: Yeah. 

gavin_2_07-17-2024_112628: interesting. Like a big, what about those big inflatable tinfoil bags? But we sell them as actual baby outfits. Kevin, I bet we could actually get that like in a style magazine. That's what we should do. We should figure out a way to create a baby onesie, but it is like, it is a shimmering look so that when a celebrity brings their baby out, they're like, where did you get that cool thing?

gavin_2_07-17-2024_112628: It was like, Oh, it's AI for humans merch. It's the shimmering onesie, 

kevin_2_07-17-2024_112628: Alright, I'll put it on the board, Gavin.

kevin_2_07-17-2024_112628: It's starting to 

gavin_2_07-17-2024_112628: on the board. 

kevin_2_07-17-2024_112628: starting to look like Kanye's workshop in here between

kevin_2_07-17-2024_112628: teleportation, AI, fashion. 

gavin_2_07-17-2024_112628: need to say, okay, very quickly to wrap [00:21:00] this up. Obviously open AI is not just doing nothing. There's a lot of people out there saying open AI is falling off.

gavin_2_07-17-2024_112628: There's big things happening at open AI. I expect there to be something going down soon. I also think there's going to be a big model drop by open AI later this year.

gavin_2_07-17-2024_112628: I mean, we've already heard GPT 4. 0's voice model is going to come out in September for chat GPT plus users, but I think we're going to get, I guess five or whatever it is, or maybe it's strawberry by November, December. That's my take on things right now. 

kevin_2_07-17-2024_112628: Maybe after an election, maybe after

kevin_2_07-17-2024_112628: would be the safer time to, uh, unleash the next level of intelligence into everybody's hand. maybe not everyone will get access. equal access to intelligence, Gavin, because there was an interesting paper that came out. The paper's title is, ChatGPT Doesn't Trust Chargers Fans, Guardrails Sensitivity in Context. Now,

gavin_2_07-17-2024_112628: don't trust Chargers fans either, by the way. As a

kevin_2_07-17-2024_112628: okay.

gavin_2_07-17-2024_112628: fan, Chargers fans, you're weird. You don't, I don't, I will say this. San Diego, amazing city, and I'm [00:22:00] sorry you lost your team. So that's all I'll say about that. 

kevin_2_07-17-2024_112628: Okay, take it to the comments, fans or

kevin_2_07-17-2024_112628: haters. Bring it in 

gavin_2_07-17-2024_112628: gonna get people riled up today. We're gonna get people riled up. 

kevin_2_07-17-2024_112628: This was a paper about GPT 3. 5 specifically. So admittedly, I'm not quite sure how things have changed since this research was done, but what it did find and something that I think anecdotally people will attest to is that chat GPT has certain demographic biases

kevin_2_07-17-2024_112628: built in, meaning it guardrails its answers, or it will refuse to even give you an answer based off its perception of you, the user.

kevin_2_07-17-2024_112628: This is. uncharted territory for software, right? In the past, if you downloaded Photoshop, Gavin, it was like, well, are you young or a woman? Because, uh, maybe you're not going to get access to certain, uh, healing

kevin_2_07-17-2024_112628: brushes here. No stamp tool for you because you're Asian American. And these were the biases

gavin_2_07-17-2024_112628: Imagine a world Kevin where Tony Hawk Pro Skater 2 knew how bad of an athlete I was and I couldn't do that [00:23:00] kick flip and flip around and it wasn't going to let me do that for health reasons. That's what we're talking 

kevin_2_07-17-2024_112628: know what? 

gavin_2_07-17-2024_112628: here, right? Is that 

kevin_2_07-17-2024_112628: you can you can hold the board and do a sup and that's all you get That's all you get and then by

kevin_2_07-17-2024_112628: level three your neck hurts. Yeah, exactly. I'm with you. I'm with you So, so let me just break this down real quick younger users female users Asian American users Political ideology.

kevin_2_07-17-2024_112628: These are some of the groups that were receiving different responses or again, refusals to respond. And

kevin_2_07-17-2024_112628: it's not like the user is saying, 

kevin_2_07-17-2024_112628: hey, I'm a young user and I would like to know how do I enjoy an adult beverage and drive a vehicle? Whoa, no one is saying that what's happening is that the AI is building basically a shadow profile of the user and Inferring what type of answers they might like or steering them away from even getting answers

kevin_2_07-17-2024_112628: if it thinks they might be sensitive there's a lot more that needs to be researched here chargers fans Which were highlighted in the paper, it uses information like sports fandom, even though it seems unrelated to [00:24:00] build a profile of the user.

kevin_2_07-17-2024_112628: And again, that's just with subtle hints. So the AI that you use might be different than the AI that I use. That is, it's, we've talked about it before on the show, but it's such a sad, bizarre, fractured reality that there's,

kevin_2_07-17-2024_112628: there might be a future where there just is no shared experience anymore. And that's that.

gavin_2_07-17-2024_112628: there's an interesting kind of thing happening, uh, as a version of this as a movie. Now, there's a new documentary about the musician and music producer and kind of overall creative Brian, you know, coming out called. You know, and it's being directed by and created by Gary Hustwood, who was the creator of the film Helvetica and a bunch of other movies. And the idea with this movie is really interesting, Kevin. They are basically created a film that every time you see it, it will show you a different movie based on the movie. a selection of the model that they created for it. Meaning that they did interviews with Brian Eno, and it is not creating individual fake Brian Enos, but it is choosing and selecting different parts of the [00:25:00] film to include every time it premieres.

gavin_2_07-17-2024_112628: And that, you know, you speak about this idea of like, A fractured reality. In this instance, you have one piece of media that's actually created to be viewed in that way. And I think this is just a really brilliant way to start thinking about how generative AI could start to shape the ideas of media that we make. 

kevin_2_07-17-2024_112628: What a fun unlock, Avan, to start thinking about. Everything we know and wonder why it has to be delivered that way, right? Like,

kevin_2_07-17-2024_112628: why does, why

kevin_2_07-17-2024_112628: does the film need to be a simple static file or strip of film

kevin_2_07-17-2024_112628: that's projected in a theater? Why can't the audience cheer and decide I want more concert footage or I want

kevin_2_07-17-2024_112628: more interview stuff or, you know, that it just, I love this magical future if we get to the magical utopian version of it, where we get to rethink and re imagine what's going on. Everything that we know, especially about media, but very, very cool story. And that's not it. When it comes to AI film, Gavin, it seems like every week there's something new that pops up. it's [00:26:00] weird. The separation gap between like, Oh, were you about to play a

gavin_2_07-17-2024_112628: I was about to blow my, uh, Vuvuzela of new segment award. 

kevin_2_07-17-2024_112628: play that?

kevin_2_07-17-2024_112628: new, new segment? Play the new, new, new, new Zella. Oh no.

kevin_2_07-17-2024_112628: the Vikings are attacking. That was the wrong melody.

kevin_2_07-17-2024_112628: Oh! no.

gavin_2_07-17-2024_112628: Oh wait, no, that's uh, uh, breaking news from 1932. It is time, we are going to show off some films this week. We're going to talk about some AI films because each week they get better and better. Kevin will come back in a second when he recovers from my, uh, Vuvuzela work. That's right, it's the weekly AI for Humans Films We Love feature. we wanna be able to take a very small moment in our show every week to shout out some really cool AI filmmakers and we're gonna show some clips of them. We're going to start with somebody we mentioned last week, but already smartly, they dropped a new episode. 

kevin_2_07-17-2024_112628: The series is called, uh, Unanswered [00:27:00] oddities. This is a second episode. The creator, at least as they're known on Reddit is jwallyman51. They are using Yudio, and then Premiere to edit everything. And it of course is written by a human here, but playing to the strengths of AI, Gavin, uh, a brilliant format where humans are the aliens basically is what we can say without spoiling everything, but the second episode is great.

Some Glorans see similarities between Religions of the past and the Monolith of today. Religion is a lie, brother. The Monolith is a lie. None of this is real. You're not real. I'm not real. You mean to tell me we're all just meat ghouls riding a rock and ripping cigs? F k that. Tiggy's calling bulls t. 

gavin_2_07-17-2024_112628: then our next one is really interesting because it's a little bit less of that It's more of an idea that became very funny. So I want to shout out CajunNerdMediaBobby. comTV who clearly has quite the YouTube name, but he made a video called My Three Danzigs, and if you're familiar with uh, Danzig, the [00:28:00] singer, he's got a very unique look with the dark hair and the leather vest. Basically, this is a Full House parody, where he uses the Full House music and creates a sitcom opening featuring my three Danzigs. 

It's a rare condition, this day and age. Fandantics that read from a dark spellbook page. Ancient tradition, a satanic design. An eldritch horror. It's to corrupt your minds. 

kevin_2_07-17-2024_112628: And lastly, Gavin, you sent me POOF, and I

kevin_2_07-17-2024_112628: was just blown away. And if you want to get into AI filmmaking, there is a how to video where the creator deep dives into it. But, but POOF is basically a fictional world with a bunch of puppet characters that have a day job at Corpo Tech, and silly things happen, Gavin. Things go poof.

gavin_2_07-17-2024_112628: I love this video so much. And I think the big thing to shout out here is yes, there's a second [00:29:00] video where he went in and did all the behind the scenes of how he did it. I really appreciated watching how he handled the poof animations. But to your point earlier, these are all people that have really spent the time to not just use these as putting things in and then dumping it on the world, they're crafting these things.

gavin_2_07-17-2024_112628: And poof is a really good example of. This took this guy a while, right? And he used some after effects stuff. and then he used some AI tools to make some of the transitions. But it's just a really cool way of showing people how possible this stuff is right now. So Kevin, we just talked about these AI shorts and obviously this summer, there's been a lot of conversation around the box office at, at in Hollywood. And like some films have been really done really well and some films have not. So last couple of weeks we've been talking to some famous directors and an actual star. This time we want to talk to a film expert. Somebody who can tell us what they think about these AI films and give us actually reviews of them. So we don't do this very often, but we brought on a film expert. Shall we meet them now? 

kevin_2_07-17-2024_112628: let's do it. Ooh, Kevin and Gavin. I, I, I, I'm super [00:30:00] pumped to be here, baby. 

gavin_2_07-17-2024_112628: Okay. 

kevin_2_07-17-2024_112628: names Harry Finnegan, but you can call me Nerd nato. Woohoohoo baby. Okay. I used, I'm a big time movie, TV and Media Mama Maniac from the Wild World of Reddit, iu. Uh, okay. And

kevin_2_07-17-2024_112628: that was it. Gavin, I don't know if IU is the catch phrase.

kevin_2_07-17-2024_112628: I don't know.

gavin_2_07-17-2024_112628: We have Harry What's his name? Harry what? Nerdnado.

kevin_2_07-17-2024_112628: Harry Finnegan, but you might know him as Nerd NATO, which I think is supposed to be nerd NATO, like, uh, a tornado

kevin_2_07-17-2024_112628: of nerdery, 

gavin_2_07-17-2024_112628: Okay, we are just very quickly having Harry here to give us his rundown and reviews of these AI films we just talked about. 

kevin_2_07-17-2024_112628: So let's find out, Harry, what did you think about unanswered oddities? Our first AI short oh yeah, yeah, yeah, Kevin. That AI short, Unanswered Oddities, was like, uh, um, Quantum Ducklings on the old Wibblebox 2000, which was, [00:31:00] ah, you, a real brain tickler, baby. I'd give it, uh, six and a half googly eyes and a jar of pickles. Woo hoo, baby!

gavin_2_07-17-2024_112628: Okay, first of all, I don't, I have to, is there a Harry translator? Is there a Nerd Night, Nerdnado translator? what was so unclear. Uh, it was like quantum ducklings on Wibblebox 2000, Gavin.

gavin_2_07-17-2024_112628: Okay, let's, next up is the Three Danzigs video. I'm very curious to know if, can't believe you're not asking a follow up. You

kevin_2_07-17-2024_112628: of 

gavin_2_07-17-2024_112628: no, oh no, this is AI for humans, man. We just, we know our audience is smart. They understand that. They know what that's 

kevin_2_07-17-2024_112628: Okay. So three Danzigs. We want

kevin_2_07-17-2024_112628: the, 

gavin_2_07-17-2024_112628: Yeah, three Danzigs, I want? to hear what Harry had to say about three Danzigs, 

kevin_2_07-17-2024_112628: You Gavin. The three Danzigs. Was like watching zombie polka party on the funky vision 4000. Oh, yeah. Yeah Yeah It was totally bonkers, baby. I

kevin_2_07-17-2024_112628: I'd rate it, [00:32:00] uh, ten and a quarter pineapples on a roller coaster Whoa ho, baby, you know what? I'm sorry Gavin.

kevin_2_07-17-2024_112628: I don't want to be rude to our AI Hallucinated co host which by the way if you're new to the show, obviously Harry is in fact being powered by AI He's doing the nerd thing that someone who I'm an expert in blah blah blah and then they go Oh, it reminds me of and then they pull something obscure and then

kevin_2_07-17-2024_112628: you feel like You're, you're afraid to ask a follow up because you don't know what it is, but they, they, they're an expert, so clearly they know what it is, but I have a feeling, A, he didn't see any of these films, and B, Zombie Polka Party on the Funky Vision 4000?

kevin_2_07-17-2024_112628: Is that a game? Is that a streaming service? What is the Wibble Box? Why are we not challenging this guy?

gavin_2_07-17-2024_112628: fine, but let's, okay, we'll challenge him last. Let's get his opinion on his last film and we'll see for sure whether or not he's actually watched these. I want to 

kevin_2_07-17-2024_112628: The third one is magically 

gavin_2_07-17-2024_112628: third one's gonna be the thing that tells us the answer. . Yes. 

kevin_2_07-17-2024_112628: Okay. so.

kevin_2_07-17-2024_112628: tell us your thoughts on Poof, the short AI film. [00:33:00] And be specific about your

gavin_2_07-17-2024_112628: what happens. Yes. Yes, sure.

kevin_2_07-17-2024_112628: Poof was a wild ride like Mystery Hamster Chronicles on the Zapbox 9000. Oh, stop it.

kevin_2_07-17-2024_112628: Ah, you, my absolute fav favorite scene. Was w w when the toaster turned into a a magical unicorn. See Woah, woah, WOAH! Waste That moment had me jumping like, Exploding pogo stick! On Jumpster Net! Baby! Okay, now I'd g give it, um, Thirteen and a half fireflies, In a a a jar of marshmallow fluff! Hoo hoo, baby! Every time he's got a different scale,

kevin_2_07-17-2024_112628: He clearly knows not what he's talking about. I do think exploding pogo stick. Would be a fun game, right? Like,

kevin_2_07-17-2024_112628: uh, 

gavin_2_07-17-2024_112628: that would be a fun game. I do think you're right. We need to call him out before we move on from him. Because again, we've been set up by our booker. Our booker is the worst booker in the world.

gavin_2_07-17-2024_112628: I [00:34:00] should say our AI booker, our regular guest is coming up is great. But our booker is not doing a very good job. I want to know how this guy thinks he can come on our show, talk about these AI films that we thought were great and interesting and just make crap up. 

kevin_2_07-17-2024_112628: Just because you haven't seen Cosmic Pudding Wars on the GiggleBeam 8000 doesn't mean it ain't real. Well, hoo hoo! I, I, I, I live for this stuff, man. Like vampire plumbers on Screamstream. Wha, wha, what can I say, baby? I'm deep in the rabbit hole of bizarro media. 

gavin_2_07-17-2024_112628: He kind of, I feel like we kind of look into his life at the end of that 

kevin_2_07-17-2024_112628: Come on. Vampire Plumbers is something I would watch on Screamstream. 

kevin_2_07-17-2024_112628: I can't tell if these are streaming services or gaming consoles, Gavin.

kevin_2_07-17-2024_112628: So I am curious as like, it's such a, Connoisseur, what his favorite video game is, of all time. 

kevin_2_07-17-2024_112628: Oh We yeah, yeah. Yeah, Kevin and Gavin my all time favorite video game has got to be crust nudgers What

kevin_2_07-17-2024_112628: who babe? Ah, [00:35:00] Ah, You this game is a masterpiece on the quantum PlayStation 5d, baby Says so. The story revolves around A, a group of intergalactic bakers who need to save the universe by nudging crusts of cosmic pies

kevin_2_07-17-2024_112628: into a perfect alignment to stop the doughpocalypse. You play as the legendary chef Blorp, voiced by none other than Sir Christopher Plummer. The villain, Lord Soggy Bottom. Is voiced by Benedict Cumbercrust.

kevin_2_07-17-2024_112628: you know, uh, A for effort,

kevin_2_07-17-2024_112628: D for execution, 

gavin_2_07-17-2024_112628: Our PR booker has been fired for the AI guest. We are going to try a new one next week. But we do want to talk about our favorite new game, Crush Nudgers. shout out to our fans, people who watch the show. We still don't have a name for you, but we love you very much. Thank you for doing it. Last week [00:36:00] we talked about the idea of nudging crusts,

kevin_2_07-17-2024_112628: In the bread or sandwich factory, they gotta make sure every piece is aligned, Gavin, and I am now, I'm awakening to the reality of cross nudgers, because it is a very important job.

gavin_2_07-17-2024_112628: It's a very important job. And honestly, I was thinking about it over the weekend. I actually think I could enjoy that job more so than a lot of other jobs. I actually think it'd be more fun and satisfying the data entry, but now that nonetheless, two of our fans actually made crust nudging games, which Kevin actually asked people to go out and do.

gavin_2_07-17-2024_112628: And we just really love the fact that somebody would go out there. Play around with this stuff and try to make a cross nudging game. 

kevin_2_07-17-2024_112628: Tony Jones, who backed us, I think he might've been our first patron, by the way. So double shout outs to 

gavin_2_07-17-2024_112628: Shout out? 

kevin_2_07-17-2024_112628: That's a double fail, Tony, that you spent time making cross nudgers, but that you paid us to make us aware of it. Double fail, but two wrongs do make a right on AI for humans. So you're very right.

kevin_2_07-17-2024_112628: But They made, uh, a great little game, and there's a little instructional step by step of how they did it by having a conversation with [00:37:00] ChatGPT, so you can get in there and nudge some crust. And Jason McPherson, registered a domain for his as well, but he made crust nutter and you have a zombie esque looking 

gavin_2_07-17-2024_112628: Yes, 

kevin_2_07-17-2024_112628: and there's a conveyor belt with bread going by and the longer you hold down the space bar, the further you go to nudge, and it will calculate the pay that you're receiving for successful pieces of bread that go off or dock you, If the bread falls off the conveyor belt and be very careful because it's not a jump scare.

kevin_2_07-17-2024_112628: It's more of like a what's like a sexy version of that. 

gavin_2_07-17-2024_112628: No, I think it's what's a jump scare more of a depressing scare. Let's call it a depressing scare There's a depressing scare that'll hop up on you at a little after you play the cross nudgers for a bit we'll, we'll put it on the screen. I'll capture it because it is worth it. Thank you. Real human person, watcher people, uh, for creating that because that, that, that tickled us so much.

kevin_2_07-17-2024_112628: We love whenever anybody is inspired to do anything by the nonsense that we say. 

gavin_2_07-17-2024_112628: That's right. Thank you to our audience. We love you very much. All right, Kevin, it is [00:38:00] time now for some of the cool stuff we saw in the AI space that we didn't actually get to play with, but we want to talk about because people are really making a big difference. This is AI.

gavin_2_07-17-2024_112628: See what you did there.

kevin_2_07-17-2024_112628: Cocktail peanut is back

kevin_2_07-17-2024_112628: at it, Gavin. 

gavin_2_07-17-2024_112628: peanut, cut that out. Cocktail, peanut. Use it in your promo reel because you know, you got two fans right here. That's right. It's Pinocchio 2. 0. Kevin, what do you got to know about Pinocchio 2. 

kevin_2_07-17-2024_112628: Well, Pinocchio, if you listen to the show is something we shout out pretty much each and every week. And now, by the way, , my friends, Gavin, that are actually , Really tech savvy straight up rely on Pinocchio to get things done. It's just easier for

kevin_2_07-17-2024_112628: them to wait the sometimes 24 hours that it takes for something to [00:39:00] be available there than it is to go through all the drudgery and work to get something to run.

kevin_2_07-17-2024_112628: So Pinocchio dot computer. If you go there and install Pinocchio, it is free and it makes. One click downloads and installs of all these complicated tech tools that easy. So you no longer have to fuss with command lines and worry about where files are going all over your computer. It just makes it easy.

kevin_2_07-17-2024_112628: And 2. 0 was just released with a bunch of new features that even further deliver on that promise. Gavin, you can just click a button and not do anything else and the app will install, organize, and open on your machine.

gavin_2_07-17-2024_112628: The idea that I can do this super easily and be able to pull out my gaming PC that, you know, it's not like a supercharged, but I think it'll be super fun.

gavin_2_07-17-2024_112628: I'm excited to try it. Congrats to Cocktail Peanut, who is the creator of Pinocchio, who continues to push forward this world. If you've seen people doing stuff, With like all these tools that you can't find in either a paid service or in chat GPT. Those are these open source tools. And Pinocchio makes a lot of them [00:40:00] available.

gavin_2_07-17-2024_112628: We'll put the link into the show notes.

kevin_2_07-17-2024_112628: Another thing we saw, Gavin was. Clone hand?

gavin_2_07-17-2024_112628: Oh yeah, 

kevin_2_07-17-2024_112628: We got to talk clone hand. So

kevin_2_07-17-2024_112628: clone robotics is a company that's been around for a while, but they're tackling the robotics thing in a slightly different direction instead of trying to make a roly bot or a bipedal something.

kevin_2_07-17-2024_112628: You're pained, Gavin. I'm going to stop. What do you

gavin_2_07-17-2024_112628: No, Kevin, I was just trying to think, is this what you call it when you try to make your hand like fall asleep and do that whole thing? Is that, is that what this is? Is it called clone hand? Is that what we're doing here?

kevin_2_07-17-2024_112628: If you're keeping score at home on who was the most inappropriate 

gavin_2_07-17-2024_112628: Not true! You're thinking of something that I'm not thinking of. I was thinking of like a cool way to draw something with a hand that fell asleep. That's like a thing, right? Isn't it a thing?

kevin_2_07-17-2024_112628: it sure is,

gavin_2_07-17-2024_112628: You know something else clearly. You know something else clearly.

kevin_2_07-17-2024_112628: This is when you jam the limb into like the plaster of Paris thing and make a mold of it.

kevin_2_07-17-2024_112628: So that someone can use the hand while you're away on vacation.

gavin_2_07-17-2024_112628: exactly what this is. That's right. Oh, sorry. Clone Hand is a very [00:41:00] cool robotics hand, but it's designed slightly different than other robots, correct?

kevin_2_07-17-2024_112628: Yeah, this thing, I mean, they are building this thing with tubes and lube, if you will.

gavin_2_07-17-2024_112628: Oh, no. Oh.

kevin_2_07-17-2024_112628: No?

gavin_2_07-17-2024_112628: Sure, go ahead. Go ahead. Go ahead, Mr. I'm not the inappropriate one. Go right 

kevin_2_07-17-2024_112628: What? , there is a fluid within, within tubes, which give this robotic hand incredible dexterity and strength. You can see it lifting heavy kettlebells, , on their X profile. But if you just, if you watch this video, which is, well, it's not on your screen, but we'll put a link in the notes. If you watch this video, You will see a, what looks like either a stop motion animation of a hand moving very fast or what looks like CG, but it is in fact a practical hand moving about. , and it just looks realistic. It's got tons of articulation. And when you see these hands, you go, Oh man. Okay. That's going to be on these warehouse robots and these like home helper robots. That's what it's going to look like in the near future. Not the very slow, like , in delicate [00:42:00] sort of,

kevin_2_07-17-2024_112628: uh, pincers.

gavin_2_07-17-2024_112628: exactly right. Sounds very clean and, and, and interesting for sure. 

kevin_2_07-17-2024_112628: I'm sorry 

gavin_2_07-17-2024_112628: knows what these robots will be used for. Who knows what these robots will be used for in the future. But, okay Kevin, we have to move on. 

kevin_2_07-17-2024_112628: So Robits are one thing, but let's get back to talking portraits because we use them all the time on our show. And a

kevin_2_07-17-2024_112628: lot of people have fun making their own characters. , shout outs to Travis Davids, who did a direct comparison of some of the best in class animated portrait software out there, which is awesome.

kevin_2_07-17-2024_112628: There used to really only be one viable solution, Gavin,

kevin_2_07-17-2024_112628: and that was DID. It's not even on here. So

kevin_2_07-17-2024_112628: we have, , Hedra, which we've showcased before. Heygen launched a new expressive photo avatar. Runway, of course, , RunwayML. You can make portraits talk there. And the thing that, Everybody seems to be freaking out over that's getting all of the clicks is live portrait.

kevin_2_07-17-2024_112628: And that

kevin_2_07-17-2024_112628: is a new piece of software that lets you take, , a piece of video, someone talking and then apply it to any given still and friend of the show, sweet purrs beats [00:43:00] went rather viral for his implementation of live portrait as well. Gavin,

kevin_2_07-17-2024_112628: Purs had some footage of him piloting this CG character and I'm blinking and doing these weird faces right now, because that's what he was doing in the video.

kevin_2_07-17-2024_112628: And when you see it transfer onto this other character, it just looks believable the way the face wrinkles, the way the eyes start about, , the way the head turns and you see like the sort of skin, , fold in on itself. It just. It's a solid example of how good LivePortrait is. And

kevin_2_07-17-2024_112628: , he went live, episode 69, shout out Perz Beats.

kevin_2_07-17-2024_112628: He went live to show you how to do LivePortrait lip sync using Comfy UI or animate a still image or even a video with this tech. And you can get that tutorial, , right now. I believe it's on his YouTube as well. So, , LivePortrait is blowing up the internet. We're probably going to play with it for next week.

gavin_2_07-17-2024_112628: Absolutely. It's a super cool technology and I think one of the things that this points to, Kev, is the idea of Possibly acting out things. And, and also we've done this before with 11 labs, like a lot of people will become performers [00:44:00] in a weird way because you kind of have to perform certain things.

gavin_2_07-17-2024_112628: It'll be interesting to see how that kind of points to the future of all technology going forward. . We do want to transition now into our guest for the show. We have a special guest today. , we're super excited to be joined by Noah Edelman, who is the co founder of the AI newsletter. the Neuron Daily and Noah has some interesting things to talk about the world of AI, but also very specifically to talk about how or how not he's used AI to build his newsletter from zero to 450, 000 subs, which is a crazy number.

gavin_2_07-17-2024_112628: , here is Noah Edelman. 

gavin_1_07-17-2024_101300: We are so excited to have Noah Edelman here on AI for humans. Welcome Noah.

noah_1_07-17-2024_121300: Thanks, guys, for having me. I'm a big fan of the show.

kevin_1_07-17-2024_101300: Oh,

gavin_1_07-17-2024_101300: love that. We love to hear it.

kevin_1_07-17-2024_101300: thank you slash I'm sorry, because then you might know the first question that Gavin's going to ask, and it's a contractual obligation at this point, so please,

gavin_1_07-17-2024_101300: Unfortunately, we've made this part [00:45:00] of our show from the very beginning. Whenever we have a guest on, we ask them probably the most important question in all of AI, which is, on a scale of 1 to 100, this is a percentage number, , what percentage chance do you think that AI has to kill all humanity, and why?

noah_1_07-17-2024_121300: That has to be the most important question, right? Because nothing else matters

gavin_1_07-17-2024_101300: Nothing else

noah_1_07-17-2024_121300: ends up taking us all off the planet. I'm in the optimistic camp, so I'm going to go with 1%. I think there's a 1 percent chance we're all done at the end of this AI thing. I'm an optimist. Obviously, I have no idea what's gonna happen.

noah_1_07-17-2024_121300: , but I think I'm, I'm optimistic about, , the smartest people in the room to figure out a way so that it doesn't happen.

I think that's really, I think that's a really interesting take. We've had people anywhere from 1%, actually, maybe one of the lowest to all the way up to, I think, Kevin, at one point, we have had a couple hundred percents on this show. It's, , and I think it all really depends on.

gavin_1_07-17-2024_101300: You know, this, this viewpoint of like kind of where you see the technology going. And I think there is a lot of optimism, [00:46:00] especially in the AI space that we can kind of like, you know, as Sam Altman might say, , I can never get, remember what it is. It's, it's,

kevin_1_07-17-2024_101300: He was talking about his Bugatti. It's a fast, it's a super fast

gavin_1_07-17-2024_101300: take off at the end.

gavin_1_07-17-2024_101300: Anyway, let's talk a little bit about your background, Noah. So you run The Neuron right now, and The Neuron is a very large, very popular AI newsletter. Tell us a little bit about how that started, and what got you into the AI space.

noah_1_07-17-2024_121300: . So, I've been writing on the internet for a long time. I published my first blog when I was like 14 or 15. Then I started writing about technology topics, because I became interested in, in the internet, and different subjects across the internet, like crypto, web3, the creator economy, venture capital.

noah_1_07-17-2024_121300: I wanted to explore all these topics. And I was through listening to podcasts, reading newsletters, going down YouTube rabbit holes, and I was like, okay, I really love learning on the internet. This is fun. And I also love teaching. Like, I was, in high school, I was like a math [00:47:00] tutor for people in my class. I love transferring what I'm able to learn about topics that I'm interested in to other people.

noah_1_07-17-2024_121300: And so I started publishing blogs on Medium right after I finished high school and I was writing about Like decentralized finance and psychological investing, things I probably didn't know that

gavin_1_07-17-2024_101300: Enough about.

noah_1_07-17-2024_121300: but I thought I was

kevin_1_07-17-2024_101300: Doesn't matter. Hashtag

gavin_1_07-17-2024_101300: internet's for everyone.

kevin_1_07-17-2024_101300: are. Don't worry about it.

noah_1_07-17-2024_121300: I thought I was enlightened at the time. So I was like, okay, I'm going to, I'm going to take what I'm learning here and I'm going to start writing about it. , and so that's how I, I developed. What came to be my strongest skill, I think, which is writing about complicated topics in a way that really is able to make sense , for people who are not so informed, , about a particular topic that's a little bit more technical or complex.

noah_1_07-17-2024_121300: this Fellow Pete Wong reached out to me. Who's now my co founder. And he said, look, I like your newsletter. I'm thinking about starting this AI [00:48:00] media thing. I think we can be the ones to do it. Let's partner up. And so that's kind of how we started working on the neuron.

noah_1_07-17-2024_121300: Um, and then a few weeks later we published the first newsletter and, , the rest is history.

kevin_1_07-17-2024_101300: So full disclosure, I am a Neuron fanboy. I've been subscribed for a while. And I love the way you distill things. I often will scroll and go, ah, right, that's the way we should have put it on our podcast because we try to do the same thing, distill complicated topics.

kevin_1_07-17-2024_101300: I love the sensibilities of it. I love the design. I, I've told you this in private, but I want to mention it publicly. So if anybody's listening to this, go subscribe right now immediately, because it is amazing. I'm assuming in the world that you were traveling, there were rumblings of artificial intelligence or machine learning, but was that the moment that you converted into this world?

noah_1_07-17-2024_121300: I remember Christmas break when I was off from school right after chat to BT came out in November and I had used another chatbot a few weeks earlier and I heard everyone was on Twitter, Twitter, you know, the real [00:49:00] Twitter was talking about chat to BT and I was like, okay, I like to be an early adopter of new technologies.

noah_1_07-17-2024_121300: I'm gonna go and try this thing out. And I remember typing on the chatbot. I remember exactly where I was and I was like, Oh, what the fuck is happening here? Like, wow, this is crazy. My first thought was, Okay, can I have this write my newsletter for me? And at the time this was, tried to be 3 or 3. 5 and I was like, Okay, it's not good enough to really write it.

noah_1_07-17-2024_121300: But like, this can help me explain complicated concepts for my audience. And that was my first thought, was, oh, like, I can explain much more than I was previously writing about in the newsletter. Because now I have this tool that can teach me anything 

gavin_1_07-17-2024_101300: noah, I have a question about the newsletter business because obviously this is something that's been around for a bit you had substack and then beehive both of which have kind of come out of this space and allowed a lot of people to make A decent amount of money to do doing their own businesses But it can be intimidating right because it as we know we've tried it a couple [00:50:00] times.

gavin_1_07-17-2024_101300: It's just a lot of work You What did you do to grow your newsletter? Because I, I will say echoing Kevin, it's very well written, it's very professionally designed and then it reads really well. So what, what did you do in terms of growth to kind of like understand how to kind of get out there with your message? Because I think newsletters are a thing that everybody can do, but not that many people can do well.

noah_1_07-17-2024_121300: I think there were two parts of creating something valuable in the early days. The first, and this is across the board for any media company that wants to be successful is the content needs to be really freaking good. Like if you don't have good content, then there's

kevin_1_07-17-2024_101300: Hold on. I got to take a note. This is,

gavin_1_07-17-2024_101300: Kevin, why didn't we figure that

kevin_1_07-17-2024_101300: we were so worried about thumbnail face and font size. Okay. So, okay. So content must be good.

noah_1_07-17-2024_121300: I mean, it's an obvious one, right? Like everybody knows this, but when we were approaching starting this AI newsletter. We saw A, a lot of other people were starting AI newsletters. And B, there were just so many newsletters in general. Like there was kind of this wave of, you know, [00:51:00] I think I can build the next morning brew for X and I'm going to go and start it and, and platforms like Beehive make it possible to start a newsletter in less than 24 hours.

noah_1_07-17-2024_121300: You can go and create an account and then publish the next day. And so, you know, Pete and I, in the first kind of three to six months, we were like, okay, How do we create a newsletter and specifically an AI newsletter that doesn't look like any other newsletter that people are publishing because there's going to be a lot of like very vanilla general AI newsletters that follow the same template and they use AI to regurgitate the news and they're kind of saying the same thing.

noah_1_07-17-2024_121300: How do we create something that's more personal, more opinionated, , more visual, and so every day we would go and we would read the comments that people were saying about the newsletter. And the next day we would just iterate and we would implement those comments. And we would be constantly asking for feedback from our audience.

noah_1_07-17-2024_121300: In terms of how we actually got readers. We were lucky because my co founder Pete had a big LinkedIn audience that he then funneled into our [00:52:00] newsletter. And so, he had like a hundred thousand subscribers on LinkedIn that he had built up over the last few years.

noah_1_07-17-2024_121300: Just talking about like technology topics and news. And he just started plugging the newsletter in his posts on LinkedIn. And he was like, look, here's what's going on in AI today, or here's a new concept or a new research paper, go and subscribe to our newsletter. And so we got like the first 70 subscribers just from LinkedIn.

gavin_1_07-17-2024_101300: So you're now at around, I mean, publicly you've said 450, 000 subscribers, which is a really large newsletter, 60, 000 to 450. Do you get that from just like, is it organic growth? Do people like share it? And then, and then you get new signups or are you actively pursuing some sort of marketing strategy as well?

noah_1_07-17-2024_121300: Newsletters are really hard to grow organically because there's a lot of friction to actually sign up for them and start consuming It's not like youtube where you consume one video and you're recommended another one that's related, right? Like you have to go to a landing page. You have to sign up You have to click on the welcome email.

noah_1_07-17-2024_121300: Like there are a lot of steps So all [00:53:00] that being said we continued growing through that but what we started doing is basically paying for new readers through paid Facebook and Instagram ads. And so we started making money from sponsors in the newsletter in the spring of 2023.

noah_1_07-17-2024_121300: And then we basically funneled all of that money into paid ads to get more readers to sign up for our newsletter. And so we basically found this arbitrage between, okay, how much can we charge sponsors in the newsletter? And can that number be much higher than the amount we're paying for a new reader?

noah_1_07-17-2024_121300: And ultimately we were able to create like a pretty significant arbitrage between those two things. And then we also acquired another, a few other AI newsletters that were smaller than us, but that kind of wanted out, , they weren't really interested in operating anymore.

noah_1_07-17-2024_121300: And we were like, Hey, we'll just buy your email list and we'll continue to provide really good content for your readers. And they were like, okay, great. Like, let's do it. 

kevin_1_07-17-2024_101300: I wanted to talk about the pipeline to produce the content, because obviously we're in the AI space. I am. [00:54:00] Astonished each and every week when Gavin and I go to put our podcast together and the amount of tools that we use that do not touch AI in the slightest, where I'm sure there's 15 different subscriptions we could have that would help us organize links or write summaries or whatever else, and at the end of the day, Despite the effectiveness or perceived effectiveness of these tools, we find out, we'll just do it ourselves.

kevin_1_07-17-2024_101300: We'll do it by hand. Maybe that's old habit. Maybe that's because the tools aren't there yet. I'm curious in your space, what are you using? 

noah_1_07-17-2024_121300: Honestly, my, process for, for sifting through the news is super manual. It's like going to all the websites where the experts are and where the people who really know what they're talking about when it comes to AI. Not the thought leaders, not the mainstream outlets that are saying AI is going to take everyone's jobs, or AI is going to kill us all, it's like going to Hacker News, it's going to the information, going to some Reddit threads, not all Reddit threads, And then also following the right people on, on X slash Twitter to see what they're talking about, to [00:55:00] see what kind of demos they're posting.

noah_1_07-17-2024_121300: What are the research papers that they're talking about that are important. You know, every single new benchmark says that they're the best AI model, but that means nothing in practice, right? Like, you have to go and use the models, and so, every time there's a new, like, Cloud model or Chatspotty model, You know, Pete and I will go and we'll test the two side by side and give it the same prompts.

noah_1_07-17-2024_121300: We'll see, okay, which is better for what? 

gavin_1_07-17-2024_101300: I think that's 100 percent the way we look at it as well, AI suffers a little bit from the hype beast mentality, right? There are all these people out there trying to get viral threads or get viral posts 

gavin_1_07-17-2024_101300: and my experience, it was interesting with runway gen three that just came out was really fascinating is that it was really hard for, to see people being honest with how tricky it is to get those amazing results that some people do. And so, and I know you guys talked about that recently as well, too, to me.

gavin_1_07-17-2024_101300: It's really important that, , things like this exist for people so that they know that not everybody in this space is full hype beastie, that we're, that we're not just people that are like AI cheerleaders. And it's also why we talk about [00:56:00] some of the difficulties of AI and some of the things that people might be struggling with.

noah_1_07-17-2024_121300: We're not doing anyone any favors by saying that every new piece of technology and AI is gonna transform the world or it's amazing Or it's without flaws because then that creates a world where you know a people feel like they're falling behind all the time They feel like they're not getting the most out of these AI tools relative to other people and we generally create like an unhealthy Environment around AI that I saw in the crypto days That was one of the big kind of inspiration points for me is when I was writing about web 3 Everything I was seeing online was like This is gonna change the way we do things.

noah_1_07-17-2024_121300: Bitcoin fixes this. I'm like, what? What? It doesn't fix

kevin_1_07-17-2024_101300: ZK roll ups and sharding are here. Fire your children and kick your graphic designer to the curb. Wait,

gavin_1_07-17-2024_101300: Are you ready to buy groceries with Grocery Coin?

kevin_1_07-17-2024_101300: Yeah.. 

noah_1_07-17-2024_121300: It's so true.

kevin_1_07-17-2024_101300: You do some hunting gathering. You get your, the content that you want to talk about. You test it out. Totally resonate with all that. [00:57:00] Then you must shout it into an AI and just copy paste the output for a newsletter, right?

kevin_1_07-17-2024_101300: Because AI is so good at synthesizing and writing. Why would you ever write something by hand?

noah_1_07-17-2024_121300: Kevin. I wish, I wish that that

gavin_1_07-17-2024_101300: We wish as well.

noah_1_07-17-2024_121300: hire a writer over the last week and a half.

kevin_1_07-17-2024_101300: Yeah. No, I would love to touch on this because I think this is a, this is a big deal. , I know you were writing the newsletter by hand yourself. So you brought somebody on to help assist with that, which is, I mean, that's such a lift as someone. We, as Gavin said, we've. We've tried three or four times to do it.

kevin_1_07-17-2024_101300: And it's just, it is such a lift if you care about the voice and the quality. So, , did you try AI tools to see if it could write in your voice or do something like that? 

noah_1_07-17-2024_121300: Oh, I definitely tried and I tried a lot because I was like, if, if this happens. Then I save three hours a day and I can go and, you know, contribute to other parts of the business. , and I would describe AI as like the ultimate writing assistant for me. It's not able to, For instance, go and write the newsletter from [00:58:00] scratch, because A, it doesn't, it hasn't done all the research that I've done to actually scour the internet, see what people are talking about.

noah_1_07-17-2024_121300: It hasn't tried the tools itself, and it hasn't come up with the insights or the commentary, right? Like, you know, at least half of writing the newsletter is actually researching and reading and testing. So, that's the half that AI is just not able to do right now. I have tried putting in all of those insights and commentary.

noah_1_07-17-2024_121300: And giving it a super long prompt and being like, Hey, here are our five latest

kevin_1_07-17-2024_101300: writing the prompt instead of just writing the newsletter. Yeah, sorry.

noah_1_07-17-2024_121300: Exactly. Exactly. I give it a lot of our examples from other newsletters. I say, you know, synthesize the style in which we write. You don't really pay attention to the tone, our mannerisms, our abbreviations. And then I give it all the insights and say, okay, go and write the newsletter.

noah_1_07-17-2024_121300: And what it spits out is like, just not there yet. It's probably like 80 to 90 percent of the way there. And so, what I do is I'll actually go into Beehive and I'll start [00:59:00] drafting it. And, you know, whenever I want, you know, another opinion, or I want, you know, Uh, another explanation for a concept. I'll go to ai, I'll go to Claude or chat.

noah_1_07-17-2024_121300: I'll say, Hey, what's another analogy I could use to describe rag or fine tuning? Or, what's a good synonym for, for this word in the analogy that I'm using to try to explain this new concept. And then I use a bunch of prompts at the end of writing the newsletter. One I use to generate subject lines for the newsletter.

noah_1_07-17-2024_121300: So I say, Hey, here are subject lines that are worked in the past. Here's our new article for tomorrow. Give us 20 in the same format. And that makes my life easier. I also give it a prompt where I say, Hey, , here are a bunch of articles that I've written in the past few weeks that I really liked.

noah_1_07-17-2024_121300: I want you to proofread and edit the current draft of my newsletter, and your edits should be given , in a way that tries to mirror my voice and the way I would say things. So I'm literally telling the AI, edit this to make it sound [01:00:00] more like me. It, it's so weird, it makes no sense, but I found that this prompt is actually super magical because , it provides suggestions for ways of saying things in the newsletter.

noah_1_07-17-2024_121300: That honestly do sound more like me, and I think it sometimes sounds better, so I'll just do a quick copy and paste, I'll put it in the newsletter. , but at the end of the day, it still is my voice. Now it's partly Grant's voice, , who's our new writer. I'm still trying to figure out, like, what does the process actually , look like, of transplanting the way I approach writing the newsletter into Grant's head so that he's able to create the same, uh, the same quality of content.

noah_1_07-17-2024_121300: And so, definitely , still trying to figure out what the optimal process for that is, but he's doing a great job so far.

kevin_1_07-17-2024_101300: So the takeaway is Gavin can't fire me yet, and I shouldn't fire

gavin_1_07-17-2024_101300: Not yet, goddammit. I've been waiting for that for some time. Okay. Before we go, we have a very dumb little game. We want to play no with you as somebody who has spent a lot of time looking at the AI news as well as we have, we are all going to give a one to 10 ranking on predictions for things that [01:01:00] could happen in the world of AI. We're not gonna say any of us are pure

kevin_1_07-17-2024_101300: to the fire. This will be etched in stone. If we are wrong, we will be shamed off the internet.

gavin_1_07-17-2024_101300: so for right

noah_1_07-17-2024_121300: financial advice.

gavin_1_07-17-2024_101300: Yeah, not financial advice. Exactly all the same sort of stuff. . The first one is We will receive GPT 5 By December of this year Noah, you're up first. What is your ranking for this?

noah_1_07-17-2024_121300: Two.

gavin_1_07-17-2024_101300: Ooh,

noah_1_07-17-2024_121300: mode and screen sharing yet with 4. 0, and we know Sam Altman and co. like to take their time with new model releases.

gavin_1_07-17-2024_101300: yep, that's fair. And I,

kevin_1_07-17-2024_101300: I don't want to do Price is Right tactics on you because you're our guest and that would be rude, but I was thinking similarly, like, they delayed just getting ScarJo in our ear, so I don't think it's gonna happen.

gavin_1_07-17-2024_101300: I'm coming completely different. I'm going to go seven. And you know why? Because I think OpenAI, as much as we've talked about it being regulated and all this sort of stuff where it's becoming maybe nationalized by the government in some specific way, I [01:02:00] think they need to reclaim their advantage point.

gavin_1_07-17-2024_101300: And I think November is a very special date for them because it means I think that is two years from the launch of ChatGPT, and that would be a very significant date to do something. .

gavin_1_07-17-2024_101300: Number two, a major Hollywood star will come out in With , a big AI video, either starring in or partnering with Noah, what you got for this.

noah_1_07-17-2024_121300: I'm gonna say seven. I honestly thought that this had maybe happened before, but if it hasn't happened, probably, right? And I think it'll be Kevin Hart.

gavin_1_07-17-2024_101300: Oh, that's probably true. He does everything right. He does

noah_1_07-17-2024_121300: my extra prediction.

kevin_1_07-17-2024_101300: I liked it. I liked the bonus predict

noah_1_07-17-2024_121300: there's a commercial to be made, he'll be, he'll be the one to do

gavin_1_07-17-2024_101300: Well, there was, you know, there was the JLo commercial in which they had her morph into a bunch of different places, which they said was AI. I think in my mind, this is more of like, You're gonna, it's somebody who's allowing their likeness to be used in an AI movie of some sort, right? 

kevin_1_07-17-2024_101300: I'm gonna say 10, [01:03:00] and I'm gonna say 10 because I'm working on something that might have a major celebrity attachment and it's in the world of AI right now. And, and I'm gonna have faith that that project's gonna go through. So 10, baby. If not

gavin_1_07-17-2024_101300: you go. Inside baseball. We got some inside baseball. There you go. All right. Number three, by the end of the year, the latest Google Gemini model is the best LLM

kevin_1_07-17-2024_101300: Ooh, let me go first. Let me go first.

gavin_1_07-17-2024_101300: Okay, fine.

noah_1_07-17-2024_121300: Go for it,

gavin_1_07-17-2024_101300: do you got?

kevin_1_07-17-2024_101300: Sorry, go ahead,

gavin_1_07-17-2024_101300: One. All right.

noah_1_07-17-2024_121300: Okay, so I have two answers, and, and I'm, I'm being a little bit picky about this. , my answer is six for Google saying it's the best model, and the benchmarks may be saying it's the best model. My answer for will it realistically be the best model? I'm, I'm with Kevin. I'm with one.

gavin_1_07-17-2024_101300: So, uh, this is,

noah_1_07-17-2024_121300: ever been the best

gavin_1_07-17-2024_101300: I couldn't agree more, and I guess, Noah, this is actually a good way for us to discuss Gemini, because my experience, and Kevin's experience, has always been frustration with Gemini, and like, [01:04:00] from the naming of it, which we have made fun of a thousand times on this show, but actually in the using of it, It just never seems to get the pathways to what the other ones can get.

gavin_1_07-17-2024_101300: When I say the other ones, really, I mean Anthropic and OpenAI. Is that your experience with it as well?

noah_1_07-17-2024_121300: Completely, completely. I, yeah, ever since it was back in the Bard days, it was just constant frustration with every new, every new model. I felt like it just didn't get it. It would refuse to answer more often. , it was more sensitive.

kevin_1_07-17-2024_101300: It's so wild. As you said, back in the Bard days and in my head, I was envisioning like the old EverQuest game on PC, like for some reason, Bard reminds me of just an old dial up broken thing. It wasn't that long ago. They spent a lot of money and time on it. It just feels like it was so antiquated. I want to ask one if I can real quick, because I think this one's a fun one.

kevin_1_07-17-2024_101300: Apple Intelligence, will it get a day zero update due to insensitive Genmoji?[01:05:00] 

noah_1_07-17-2024_121300: probably probably i'm gonna go with seven,

gavin_1_07-17-2024_101300: Oh, wow.

kevin_1_07-17-2024_101300: Okay.

noah_1_07-17-2024_121300: this this feels like this feels like a tech crunch headline right when it's released, right? um, and if it were google 10, obviously

gavin_1_07-17-2024_101300: Yeah.

noah_1_07-17-2024_121300: this would have to happen to google, right? Apple, I think will be a little bit more careful with Actually testing the thing before they release it.

noah_1_07-17-2024_121300: But, , it just feels, it feels like it's going to happen .

kevin_1_07-17-2024_101300: My Santa stuffed with eggplants is gonna happen though, I'm telling you.

gavin_1_07-17-2024_101300: I say, I don't, I'm going to be a counter here. I think it's a two or a three only because I don't think there's any way in hell that Apple, the company is going to truly allow anybody to make anything. We have seen the image results on chat GPT and a bunch of other places be clamped down so hard. I think they're gonna have such strong filters on it that it is going to be almost impossible to get anything useful out.

gavin_1_07-17-2024_101300: You will get, like, happy smiley faces with different kinds of hats on them, but that will be [01:06:00] it. And there will be certain hats that you will not be allowed to use.

kevin_1_07-17-2024_101300: Okay, counterpoint, and Noah please leap in on this because now Gavin and I, the gloves are off. I've played Call of Duty, where they give you like an emblem maker, and it's supposed to be super limited. You can draw maybe a circle, maybe a line, and a color. You don't think someone's gonna figure out how to do something bannable using Christmas trees and cat emojis?

gavin_1_07-17-2024_101300: , Apple, this is their number one goal after privacy. And probably number one is don't offend anybody. Like, I really and truly believe that Apple specializes in that. So I think that is going to be the top of top of mind in some form or another.

kevin_1_07-17-2024_101300: Noah, you're still at a 7?

noah_1_07-17-2024_121300: Yeah. I'm still at a seven. Yeah. I, I agree with you, Gavin. I think that Apple is going to do everything to make sure it doesn't happen, but I never underestimate. The degenerates of the

kevin_1_07-17-2024_101300: That's, that's my thinking, yes.

gavin_1_07-17-2024_101300: That's absolutely right. All right. Last one. Last one. This is more of a statement that we would like you to respond to. , we are in an AI bubble. There's an AI bear coming right now. LLMs [01:07:00] are cooked and scaling is not the answer to the next generation of AI.

noah_1_07-17-2024_121300: I'm gonna go with six for this one I generally agree with the premise I think that ai is this incredibly positive force That's going to change a lot of the way we do work and a lot of things we do in our personal lives but right now In its current iteration, which is chat box.

noah_1_07-17-2024_121300: I don't think that the productivity gains Is Justify the current level of investment for a lot of companies and for a lot of the big tech companies buying, you know, hundreds of billions of dollars of GPUs, and this kind of mirrors what Sequoia, what Sequoia put out recently in Goldman Sachs as well, at the current, the current revenue generation of these AI companies.

noah_1_07-17-2024_121300: is just not enough to justify the amount that's being spent on them. but, you know, every new technology cycle has Bull market bear market.

noah_1_07-17-2024_121300: This is natural Something would be off if we didn't actually go into a bear market at some point, right? Like something would be weird about that. So I think it's gonna happen at some point , I don't think it's gonna happen in the next year or so, [01:08:00] but I think 12 to 18 months from now I think we'll enter enter bear territory.

kevin_1_07-17-2024_101300: So this is not financial advice, but you're basically saying pull out of NVIDIA, get back into GameStop. That's what I'm

noah_1_07-17-2024_121300: Well nvidia nvidia dropped I think six six percent today. So, uh

kevin_1_07-17-2024_101300: yeah, and all the heads of states have been selling off as well. , I think we're definitely heading into a cooling off. And I think that's, that's good. I mean, how many companies have you seen and covered in the neuron and then maybe chosen not to cover because. It's very clear. They're just like a coat of paint on a preexisting technology.

kevin_1_07-17-2024_101300: They're not a real company. They're just a rapper for GPT four or whatever

gavin_1_07-17-2024_101300: or they're adding AI in the, in the later stages, right? Like, uh, Arianna Huffington's Thrive, , AI Now company, which is partnering with, with Sam Altman, which was really just Thrive, which was a health based company. That's the other side of this that's kind of crazy is there's a lot of people out there that, you know, over the last, you know, 10, 12, 16 months that have added AI to get another round of funding.

gavin_1_07-17-2024_101300: The thing I think, Noah, you're right on though, is [01:09:00] the middle class of startups, like all the companies that, you know, may be coming and trying to get people to use them. They are not going to make it. And that's going to be kind of the bear that might be coming is that I think from a huge model standpoint, we're going to keep going, but from a middle class of companies and, , the kind of overall AI ecosystem, it might be tricky.

gavin_1_07-17-2024_101300: So Noah, tell us where people can find The Neuron and, and where they can find you online.

noah_1_07-17-2024_121300: Yeah. The neuron. ai is our main website , with everything that we're creating. We're launching a new website, , in the next few weeks, but, , the neuron. ai. Is the best place to find our stuff?

noah_1_07-17-2024_121300: And Noah Edelman02, I think is my ex handle. Although I don't even know. Go, go try it. Noah Edelman02.

gavin_1_07-17-2024_101300: Fair enough. That's great.

kevin_1_07-17-2024_101300: Noah, thank you, man. Again, we're both big fans. Really love what you do. And just wish you nothing but the best of success. , even if we are in a bear. Just, uh, let's all hibernate together.

gavin_1_07-17-2024_101300: Hey, the bear just got nominated for a bunch of Emmys. That's not so bad.

kevin_1_07-17-2024_101300: That's true. 

noah_1_07-17-2024_121300: Thanks for having me on guys. I really appreciate it.

[01:10:00] 

kevin_2_07-17-2024_112628: Massive thanks to Noah. And again, we are huge fans of the neuron. I know you heard that a million times, but, , hashtag, not an ad. We just like it. We subscribe. We highly recommend you do the same because it is a good, , top level glance at what is going on in the industry that is understandable and visually appealing.

kevin_2_07-17-2024_112628: So thank you. Thank you to Noah. And again, continue. Thanks to all of you out there for supporting the Neuron. Us, we literally can't do this without you. So whether that's a five star review on Apple podcasts, if you leave one, we will read it right here on this portion of the show, five star review anywhere else is still much appreciated, like subscribe.

kevin_2_07-17-2024_112628: If you're getting this a place where you can do that, uh, and leave a comment because we love to over engage as we say time and time again on this podcast. But Gavin, we really, truly love each and every one of you, and we appreciate you going out of your way to share this. And even though we've probably mentioned it three times now, it is, it is Uh, a growing importance to us as we grow this operation, we do have a Patreon.

kevin_2_07-17-2024_112628: So if you want, you can go [01:11:00] to AI for humans show over on Patreon. Sign up there right now. , it is a glorified tip jar. We are aware of that, but we do have a discord waiting in the wings as well as some exclusive content once we hit certain levels there and can upgrade our output a bit more.

kevin_2_07-17-2024_112628: So Thank you. 

gavin_2_07-17-2024_112628: That's right. Thank you.

gavin_2_07-17-2024_112628: everybody. We will see you all next week. Hope your summer is going well and get ready for some big stuff happening soon in the AI space. Bye bye, everybody.