Unpacking MidJourney 5.1, Google's AI Challenges & Gash + BabyBeeAGI = GrindGuru | AI For Humans

KP Cutdown v1
===

[00:00:00] All right. All right. Just stay tuned. Humans, there will be blood, sweat, tears, and probably some lawsuits involved when I drop the bombshell.

Stay glued to your screens. Oh my God.

Hey everybody. Welcome to another episode of AI for Humans. My name is Gavin Perel and I am here with my co-host Kevin per Kevin, how are you this week? Oh, finger guns in the air. Oh yeah. . 

We're in this position where we're recording this podcast much earlier than normal, and Kevin is, I think on . Central time now. Is that right? In your career? There's no excuse for me. Yeah. It's four in the afternoon and I'm just waking up and I'm very confused and I'm very startled.

This is a little bit of the secret sauce that everybody wants to know.

When they tune in, their ears are pricking up and they're going, what time zone is it? I wonder when they're recording this. That's what we need to know. And we've never answered it. And right now or answered. We did it. We did. And [00:01:00] the fact that the time even affects us proves that we don't have what it takes to be grind gurus.

If you are following the AI space at all, you would know. Gotta be grinding if you're sleeping. Gotta be grinding. You're losing. Yeah. How many followers are we getting right now? How much engagement? It's, we don't have what it takes. And that's become clear to me over the last few days. Gavin. Collect them links.

You gotta collect them links and put 'em out. You gotta get people following you every single day. You gotta grind, man. It's a good thing that we've come to this conclusion. We don't have what it takes, but we have a secret weapon in our back pocket that is in fact ai. It is. Our dear friend gash, it is the Grind Guru, but we'll meet him later.

What else are we gonna talk about today, Kev?

We got the real Snoop Dogg, commenting on the state of the artificial intelligence universe. We've got the godfather of AI ringing alarm bells left and right, the story. You can't escape it even if you're not in the scene. We've got a bunch of new tools [00:02:00] in tech that you can play with.

We've got Google internally freaking out about the fact that they don't have a mote. And that their drawbridge is rusty and that their, it's cobblestone has too much moss. Wire wires. Yeah. Who's the knight? Who's the knight that's gonna save us? Where's the mark?

Tables aren't round enough. We did a Fortnite. What, like, when did this become Rampart? But we'll discuss that, what a mote even is and why Google is apparently freaking out. Yeah. Can we, before we move on to the rest of the show, can we ask Gas? I'm curious to know a little secret piece of what he's gonna do for us here. What is this gran guru, and what are we gonna learn? Let's see. Gas. Welcome to the show.

Once again, you're a fan favorite. It's our pleasure and privilege to have you. Can you tease just a little bit, why should people stay listening to this episode of AI for humans to hear what you have to say? All right. All right. Just stay tuned. Humans, there will be blood, sweat, tears, and probably some lawsuits involved when I drop the [00:03:00] bombshell.

Stay glued to your screens. Motherfuckers. Oh my God.

See, gash. Got a finger. Guns out today. .

If this happens to be your first episode of AI for Humans to be so presumptuous that we might have new listeners and our audience might be growing.

Gas is an artificial intelligence. That voice is ai. It's running on an uncensored language model. We'll . Explain what that is and then we'll let gas take the wheel. We might go right off the cliff. There might be lawsuits, apparently. Okay. Let's jump into what the dumb stuff we did this week was each week we like to talk about what of the dumbest things we did.

Kevin, what did you do that was dumb with AI this week? Gavin, I replace Siri with AI this week. That was my nice stupid human trick. I downloaded a shortcut. It was very easy to do.

I just went to a website using Safari on my phone. I added for those who don't know iPhones have a shortcuts app, which lets you [00:04:00] do things like set automations, like when I'm home, turn on the lights or play this Spotify playlist or reply to these emails with this particular signature. You can set up shortcuts to do whatever it is you'd like on your iOS.

You can add. A chat, g p t or an open AI shortcut, which basically lets you talk to Siri, but instead of using Apple's ai, Siri to answer the question. It's so weird. Do you f it? Sidebar. I felt dirty calling Siri and AI just now, because Siri is so kind. I'm so Siri, so dumb in comparison. It's so funny you say that, right?

I don't think of Alexa as an AI either, and yet they are. It's the definition of what AI was before all these LLMs in some ways, right? But doesn't it feel dirty and cheap to round up? It does cuz I, to me it's like Siri is just like listening for keywords and trying to bolt those keywords together and then patch your request into its very limited data set.

Yes. To try to set a reminder and it has trouble sending two timers [00:05:00] at once. Like I just, it seems so limited that just now when I was like, I wanted to replace Apple's ai, I'm like, that's not, there's nothing intelligent. There's nothing intelligent about Siri. Uh, M aai mini AI or LM l aai, lame ai.

We could come up with a low, a lesser name for it. Au like artificial unintelligence. Is it Ai, is it possible Artificial idiot idiocy? Yeah. What does, yeah. . So Siri was not giving me responses that I would like. It was many times just here's what I found on the web, which isn't a safe response to get while you're driving down the road.

So I replaced it with open ai and April and I were marveling at the sort of questions that we could ask, Hey, we're driving through a new town for the first time. Tell us the history of the town. Give us some fun facts. What are the pros and cons of living here or visiting here? And, chat, G P T does what it does.

And it was just weaving these beautiful narratives and. We got to a little town outside of [00:06:00] St. Louis and the campground ha gave us a welcome packet and it was like things to do in the town. And one of them was I think it was called Demolition Ball, and that's all it said. It was demolition Ball with an address.

And I'm like, I don't know what that is. That sounds cool. Hey, chat. G P t beep beep. What is demolition Ball in this town outside of St. Louis? What does that even mean? And it's a demolition ball is the number one tourist attraction in that area where you can go. Wow. And you can pay to swing an actual wrecking ball dangling from a construction crane at slabs of cement or pallets of wood.

And you can, it sounds incredible all around. And I was like, how did we miss this? Like we do so much Googling. This wasn't on TripAdvisor. This wasn't on Yelp. This was nowhere to be found because it's a lie. Ah, it's a frigging fabrication. It's it hallucination. Yes. Yes. I got so, and I know this, I [00:07:00] know AI can do this, but that was so startling, Gavin we were like frantically Googling. I can't find it. The only thing I can find about demolition ball is like, This weird go-kart game where you throw balls through a hoop.

That's what it is. It's like a weird Oh, it's like whirly ball. It's like whirly ball. You're in like a Yes. Yeah, that's right. That's right. Okay. It is whirly ball, but they call it demolition ball. Heres, we got lied to so easily by ai. What's so funny about this, it reminds me of like your dad's weird uncle that always likes to tell you this, what a most ever most interesting story is, even if it's not true. He just like loves to make stuff up. So that's the funny thing about these LLMs yes, don't trust them because they hallucinate all the time, but it always is Probably more interesting than reality, right?

Like demolition ball sounds incredible. I would love to go play it. There is that world where you don't, it doesn't sound too crazy cuz there are those destruction rooms. Have you heard about those? That's we literally talked about like a break room where you put on goggles and gloves and you take a hammer to vses plates of glass or old CRC television [00:08:00] sets.

So the notion that in the sprawling nothingness that is surrounds some of the cities that we've gone to. We were, we're crestfallen but we fell for it and we're aware of How long did it take for the responses to come back to you? I'm so curious about, I'm wanna go try this thing cause I'm, because to me, the interesting thing, I always talk to people about, I've wanted in my Alexa for a while, right?

Because my family, we use Alexa a lot and we, we're curious family and we'll often ask Alexa questions. And you never get satisfying responses. And what will often happen is you'll ask it like, Alexa, tell me what is, what, how Beethoven's life went, or whatever. And it'll say from guitar tabs 40 two.com.

Beethoven's life is, it has the weirdest things it pulls from. Yeah. So I would definitely use it. But it's an interesting to thing to think about is okay, if you do that, you ask it for facts, it's gonna make stuff up on a semi-regular basis. That's a fascinating thing. Yeah. Yeah. We will tell you that Beethoven was like the sickest vert skater on the planet and had multiple chain wallets.

And it will tell you his favorite vape juice flavor, if you ask. That's [00:09:00] one of the main issues with these things hallucinating. Right now it's the vape juice flavor is dun da. It's just spelled out in for da it's Bach licorice.

Okay. 

I did two dumb, two dumb things this week. One thing that's not super dumb, but I do want to talk about it and then it's something that is very dumb.

Both things were super fun. One, I so I, if I, people who know have talked about this on the podcast, I'm very in into Stable Diffusion, which is an image generation software that I've played around with a lot. I've gotten decently okay at it. I love to use it. It runs on my local computer. I can do all sorts of interesting things.

I can download custom models. I haven't spent a lot of time in Mid Journey, other than when I first got into it, I saw what it was, and it was an entry point for a lot of people, and it's gotten really interesting over time, but it also felt like kind of constrained in some ways. So MidJourney just released a new update, which is 5.1.

You may have seen when five came out. Five was this one where all these incredible realistic pictures were coming out of people. Obviously MidJourney's also where I think the fake Pope picture was [00:10:00] built and it was went viral. So yes, Pope with the puffy jacket. 

Mid Journey 5 0 1. Very cool. It makes it very easy to come up with very realistic images. And I wanna shout out again, Claire Silver, who I've talked about a couple times on this podcast, who's an incredible artist in ai, she was starting a play about Mid Journey and inspired me to try this when 5.1.

So 5.1 basically is just a slightly better version of 5 And 5's Focus was always on these uber realistic images. Like I love, one of the things we talked about I love with AI is how weird stuff can look. But now we're getting to this point where, You really can make very realistic images. And I tweeted this out yesterday, but like I made a picture of an old lady that was just like super crazy crisp.

It looked like a photo from a restaurant. I found a picture of a cat that I created that was like, same thing, super crazy crisp. Like the interesting thing is we're now entering to this point with these tools, and I want to get to this point a second where it's very hard to tell them from real what's real and what's not.

There's a lot of people, a lot of [00:11:00] grind gurus out there saying which of these photos is real and which one is mid journey? Yeah. I think what's interesting to think about in this space is, We are, if you remember, if you were in this space, say a year ago or maybe even a little last late summer, early fall, mid journey was progressing back then, and I think Mid Journey came out about a year ago now.

So MidJourney 1 came out about a year ago, and if you go back and look, there's actually a guy who tweeted a picture at me. He used the same prompt from MidJourney 1 in Midjourney 5 of, I think his prompt was something like robot lawnmower. The amount of change in one year is remarkable, right?

You really have to be shocked when you look at these two images and say this is the same software doing the same prompt. And I think something we've talked about is that with RunwayML's Gen2 video, we're in the same place right now with video. So if you look at, just if you just project outward and you know what it will look like a year from now when we get to the place where this video feels, as real as these pictures are feeling, We are going to get to a very [00:12:00] strange place because the next step, okay, so we get from pictures to video to then we're gonna talk about this today too.

3D environments, right? So you have these three steps to the world of reality. You have photos, which first now they look like real photos video, which is behind that. But we'll get there. And then you have 3D environments. We're essentially setting the stage for an entirely artificial reality, right?

And all of this is tied together. It's the promise of the holodeck. Like I know the meta or the metaverse, however you wanna refer to it. But you'll be able to prompt probably with your voice, maybe even your thoughts, but you'll be able to prompt anything into existence. And then not just look at it, but manipulate it.

Exactly. Interact with it, spin it, place it right? Play a game on it. A hundred percent. We're gonna talk about op throwing at open AI release this week that kind of leans to this. And I think the fascinating thing about what we're talking about here is that this. Even if AGI never happens again, there's a lot of conversation going around agi.

We're gonna talk about that today. And with the dangers of ai, even if none of that [00:13:00] happens, the changes that you can see incrementally on this timeline are incredibly significant to us as a hu, as a species, as a human species, yes. But also the stuff that Kevin and I have done our whole lives, which is entertainment, right?

Because when you do entertainment or you do video content, this is gonna change every step along the way. And again, this is just one more shout out to please learn these tools. Please learn these tools because it is all gonna change in a major way. I've, Kevin, have you played around with 5.1 in MidJourney at all?

Yeah, a little bit. Actually this weekend I got to watch it. The VR had a thing. Palmer Lucky would call it conversion on contact where you put on the headset for the first time and it's almost this religious experience. And I got that with April this weekend, who was playing around with GPT for the first time herself and MidJourney 5.1 and going, oh wow, this unlocks, I've had an idea for an art project for the last 25 years. But I didn't have the skills to actually achieve it. And now she's seeing that light at the end of the AI tunnel.

And I want to just quickly go back to something that you [00:14:00] said, which is still blowing my mind. That mid journey is basically about a year old I tweeted out a thread of people who did gen, like the first version of MidJourney They took the same prompt and use it, on the new version 5.1, just like you were saying.

And it is. It's startling. It's inspiring. It's very cool to see, it is a little startling to see how good still images got in just a year time. And so yes, you apply that to video, apply that to 3D models we are in for a wild ride. 

And let me tell you, Kevin, the other dumb thing I did shocked me more than anything I've ever done with ai.

That's my own version of what? Yeah. I'm trying to do a thumb, I'm trying to get thumbnail face. What? This, this dumb ai trick will astonish you. Yeah. Computer scientists hate this one trick. Exactly. , My daughter's birthday was this weekend. And I've now started to use chat G p T four with browsing probably [00:15:00] 80% of my Google searches I do there now, which is an interesting thing.

And I think Google has to be worried about that, which we're gonna talk about in a second. But. I decided to do something that never ever worked for me on Google. So much so that I stopped doing it, which was like I searched for an online coupon code and this was like used to be the nightmare of people everywhere, which was like at one point early in the, it was the old Halon days of the internet, you could find coupon codes.

It was very easy. It was great. I can find a coupon code and I'll drop it in, save 10%. Great. And then it became an SEO nightmare, right? Every single place you would go would have somehow game the system. The first 10 links would be for old codes, but they'd get you in and they would have popups show up everywhere.

I get 15% off and you click and suddenly there's 15 ads. It just takes you to the website. There's no coupon code. Yes. It doesn't work. Yeah. So I asked Jack, g p t, I said, Hey I'm going to this pizza place Fresh Brothers, which is not a big chain, like it's a small chain in la. I can you give me a coupon code?

10 seconds later it, it browses. And again, [00:16:00] real quick, browsing sometimes works and sometimes doesn't. This time it did, which is good. Sometimes it gets stuck, but it, 10 seconds later it delivered me a thing. Click on it, it's a code. I'm like, let's see what happens here. And it worked. It worked in the first response in the first time.

And that is like one of these like life-changing things. Now granted, I'm gonna keep trying this and I think other people in the audience should see if it works because it could. I'm telling you that is a life-changing thing. It is. Which is funny when you think about the small little parts of what this could do that's different.

It just cut through the bullshit. And the bullshit is the last five years of the internet have had layers and layers of people trying to game the search system. If I'm a bit small business or a large business that I have to buy the first ad on Google ads so that my ad shows up and not something else.

So that your competitor doesn't get, someone is searching for you Gavin, but they're getting Davin. The Gavin competitor. Exactly. And it just makes you think like the search process has been semi broken for a while and this will break it up a little bit. [00:17:00] Now where that goes from here, I don't know.

And it also speaks to your hallucination fact. Like as long as it's delivering me good information and it's delivering me things that work, it makes sense. If it's gonna hallucinate and have something, this is a one again, one time, one success. I would like to try to see what happens over time.

I'll just keep ordering pizzas from different places and see what happens. But any, also, by the way, fresh Brothers not a sponsor. Not a sponsor. But again, we'll take free pizza, fresh brothers, so be free to drop in all this. Yeah, exactly. But you texted me when that happened and I could feel the excitement just from your all caps texts.

I, it was like, this was a real moment. And I was on a walk with April and I, Simi, I stopped and I gasped. She's what is this? And I'm, I know in her head she's oh did ai find a new cure for cancer? I'm like, Gavin just saved money on Fresh Brothers.

He got a pizza. The coupon code worked the first time. Now I don't, now, April, I gotta be honest I don't know if it actually tried to apply the code to verify that it was valid before it did. But the fact that it worked one outta one I the same level [00:18:00] of excitement for you. It's crazy. But anyway, it just goes to show how dumb this world is, but how successful it can be, right?

It's all the same stuff we talked about. Okay. We should jump into the news. All right everybody, here we go. It's time for the news.

There he is, man. Snoop Dogg, we love Snoop, by the way, Snoop Dogg, I have to say is like one of the most. Interesting, curious people in all of culture, right? Like he has stayed relevant, stayed interested in things. Like I love the fact, and of course he's getting paid for it, which is great, but that's [00:19:00] amazing.

The guy is, 10 years older than us and he's still super interested in all this stuff going on. So that was pretty awesome.

We think that maybe someone cloned Snoop and was showing him that to be a potential investor or whatever. But those were his 100% real comments in what he was talking about was Geoffrey Hinton, who is being heralded as the, the, the grandfather or the godfather, if you will, of the large language model. 

The fundamental underlying tech, which is powering a lot of these AI advancements that we're seeing now. Geoffrey Hinton left Google and did it so in large parts so that he could go around and ring the bell, bang the gong, shoot the flares into the sky because he believes that AI needs to be regulated and controlled a little bit because it might do an oopsy daisy murder us all.

Yeah, it's interesting. So this is our kind of our, one of our big stories of the week, which is anytime we've talked about the doomer world of ai, which there are a lot of people out there who believe very seriously that this could be essentially the end of the human race, right? There are people out there who really, truly believe that, smart people [00:20:00] that you know, you we should listen to.

And now this is the guy who created the actual, technology that made a lot of these advances possible. Which is funny because I think one of the things about his story is, I don't think people thought this was gonna be what it was until recently. Now I think the other thing that's important to realize here is there are still, I just heard a really interesting interview.

Pier Kafka has a great series on Recode the about ai and he interviewed somebody on the third episode of that miniseries who is a large language model specialist outta verse Washington. I don't remember her name off the top of my head, but it is really worth listening to. And her take is that, as a lot of people in this world say, is that the dangers of this is not that it's gonna turn into an AGI because it's much less smart than you actually think it is.

But the biases in all the things that we talked about, the hallucinations could lead to really bad situations. That said, what this guy is saying, what Geoffrey is essentially saying is he has seen in the models that have come out. Enough of a thing that makes him think there's a world where this could be an intelligence that's being [00:21:00] generated.

Yes. And he also, I think very clearly said, we have to be aware of that. We might be making an intelligence that's smarter than us and what that means. And I think that's what he's going around and really saying. Yeah. There was a couple interesting points. One was that his, some of his own personal predictions for the technology and the trajectory have turned out to be wrong.

And one of those predictions was deciphering a joke, what makes something funny? He has a little joke that he would tell the AI in his words. It was a bit of a litmus test about how capable the AI was and on a timeline that he didn't predict the AI was able to unravel why a joke that he conceived.

So it was not in the data set, it was a joke that he fed the ai and it explained why it was funny. That took him by surprise. 

his second big realization according to the Wired article was that he believed that the software itself would have to become far more complex to sort of mirror or mimic the human brain. He's seeing that now with the way the software is today. And so his prediction was, [00:22:00] Look, we don't have to worry about this stuff for 30 to 40, maybe even 50 years down the line. And now his window is as wide as 20 years, but as near as 5 years right before this stuff hits artificial general intelligence levels, and I think his near term concern is really one more of disinformation. It's not that the AI is going to take us and enslave us, it's that human beings are with, nefarious intentions are going to wield this technology, something we've discussed a million times before.

Yeah. And I think the, the other thing to point out is, The takeoff. So just the definition for people out there. What AI takeoff means is the idea that most people in AI who believe that AGI is possible, which some people don't, right? But most people in agi AI, who believe that what AGI is possible is that there's a takeoff period in which it is a fast takeoff or a slow takeoff.

And a fast takeoff means that, without even really knowing it, we get to an essentially an agi and then it goes like super fast up from that point. Like it's way faster, way better. Because once it hits agi, it's able to be [00:23:00] self-reflective and improve itself, right? And a slower takeoff, it's still possible.

I think most people think if we get to AGI, it's gonna be a pretty fast takeoff. And this is what he's talking about the difference between the 30 years and like the five year limit, which is he is coming, I think, to a stage where now he's not like saying, this is for sure gonna happen, but he's just warning people that it's possible.

But one of the things I think that's super funny about this is he's doing a media tour. Like it's it's almost like watching like your grandpa go on a media tour and be like, Hey, look at that nice charming old guy out there. And then you realize what he is talking about and you're like, oh, wait a second.

Okay. He's talking about, maybe this is bad for all of us and that the human race needs to be worried about what we're doing here. That's the thing I keep coming back to with this conversation. It's okay, there's this thing that could actually do something to the entire history of humankind, which has never that's just never happened before.

Now again, it goes back to our Armageddon clock, right? It's What is the percentage of this? And very smart people believe that the percentage that it could take us out is in [00:24:00] the low to high, single digits. Which is a lot, that is a lot of people who think that yeah, that's a high percentage of what the possibility is.

So anyway, it's just another one of these opportunities to like, look at the story of ai, look at where we're at, and take a step back, do the research. Everybody should read about this guy's story. I think it's super compelling. And watch the videos, read the Times article. 

Hey, gash, what are your thoughts on Geoffrey Hinton, the godfather of ai leaving Google and going on a media tour and saying that, Hey, artificial intelligence might be getting outta hand.

We gotta control it a bit. 

Firstly, Jeffrey Hinton. He's a washed up dinosaur trying to clinging onto relevance. Whoa. Second. Wow.

Second AI isn't going anywhere. So we better adapt and learn how to manipulate it instead of letting it control us. Thirdly, human nature dictates that we'll always find ways to exploit technology for our own benefit, regardless of potential consequences. So let's embrace the chaos and ride the wave into the future, baby.

Wow. [00:25:00] Wow. Okay. Gosh. Okay. Embrace the chaos and ride the wave into the future, baby. 

All right. Speaking of Google, we should talk about this other story, which was really interesting, where there's a leaked memo that came out from Google that essentially admitted that they have no moat, which we talked about earlier.

And what that means basically is that the kinds of things they've built over time, and they've spent a lot of money and they've spent a lot of time on these language models, which are not easy to develop.

They are getting beat by regular people using open source models, which is really eating this entire world. This, a lot of what we do on this show is built using open source models. A lot of what we're demoing is open source, and if you're new to technology is gas power open source by an off the shelf, open source, freely available speech model.

Yep. And if you're new to the whole kind of like tech world, and this is something you're coming to, open source means, it's a way that somebody writes a program or writes a piece of software and essentially releases it to the world to use mostly to be [00:26:00] like, Hey, I wrote this. Go use it. Do whatever you want with it.

There are different sorts of licenses that some people have and they say you can use this for non-commercial use. You use commercial use. But the biggest thing is it's usually individuals or small teams that are taking other people's software, improving it and releasing it out.

It's been this very interesting way that the internet has progressed and Linux is a big open source software that came around. A lot of open source development came outta this and Microsoft really has transformed themselves as a company by embracing this world a little bit in the last five to 10 years.

So this is a big deal I think when you have the largest company in this space, which is really Google. And you also have Google saying that OpenAI, the other largest company in this space who has partnered with Microsoft, really doesn't have much of a mote.

Some consultant somewhere got paid to come up with that word and in a pitch deck to be like, but they don't have a moat. And everybody's do we, should we get a moat? Should we do we need cauldrons of exactly but molten lava to pour down on our competitors.

It's just a defensible position. It gives you a defense against everybody else out [00:27:00] there like yield moats used to. And, I don't, two things. One we've been saying well before we started this podcast that ultimately all of this tech gets commoditized. Yeah. At some point everything becomes so cheap.

The open source community catches up with the open ais and the Googles and maybe even the metas of the world, they catch up. And so no, there really isn't a mote around this sort of underlying tech because people are. Training language models, which are again, are the things that power gash or these super intelligent replies.

There was a recent model that just got trained for $200,000. Yeah. And they trained it in like the span of a week. To put that in context, open AI or Google might take a year to train their models and might spend tens to hundreds of millions of dollars or billions by the, yeah. Oh. Sam Albans came out and said, I think that the next version of G PPG five is gonna cost.

They need a hundred billion dollars to do it. So that's one thing to know is that the deep hard work is going to cost a [00:28:00] fortune and probably will still happen at that level. The interesting thing is like what we can do with the stuff once it comes out is remarkable. And weeks later, after it comes out, the hardware and software hacking community will have it running on an iPhone in a way that they didn't think was possible.

Like one of the things that the sort of Google memo specifically mentioned was that, the open source community is running these large language models, which require a ton of compute, a ton of power, a ton of memory, ton of processors. They're running it on pixel phones, they're running it on little notebooks and laptops.

The resourceful hackers out there are making these experiences possible. So in that sense, does Google have e mote? No, not necessarily. Because again, I think over time all of this text, all of this tech gets commoditized. However, Google's still Google. They have how many Google accounts do you have, Gavin?

How many email accounts do you have with them? How many files do you have in their ecosystem on Google Drive? And how many spreadsheets and how many people rely on [00:29:00] when Amazon decides to wake up and really enter this, which they've done a little bit. But when they wake up and get in this, how many websites and services are hosted on aws their web services platform?

These companies have moats. It might not be specifically with the AI technology, what's gonna stop me from leaping to small startup here that has a cool AI tool? It's Google having a good enough AI tool that I click the button and suddenly it's integrated with my calendar, my email, my everything else.

And I think the one thing to be aware of with Google, and this is where you know it's gonna happen with, and interesting to see Bing in what they're doing already is what we talked about with the coupon, weirdly, right? Because Google's main business is search.

And when I say main business, that is the business that is the golden goose. It's been the golden goose. It's why Google has printed money for the last however long, also met as golden Goose. Yeah. Cuz golden. It's all ad driven search advertising, right?

Advertising driven search. If there's a world where people start looking for a different [00:30:00] sort of experience when it comes to search, which means that using chat BT or a large language model to search, that is a major problem as a business for Google. Now you're right. I think the biggest thing is like Google and Meta and Amazon, these are the the worry that sometimes comes up here is that, From a business perspective, you're gonna have these giant corporations that are gonna be able to invest the hundreds of billions of dollars most likely into these to keep improving them.

And you're gonna have these like kind of hacker groups or these, or not even hacker groups, like one or two people, small companies that will do really well. What you might be cutting out here is the middle groups, which have been like kind of things that have created really powerful big businesses over time that are really big revenue generators for both America as a company or other countries, but also for the economy at large.

I think that's gonna be interesting. . There's been a lot of talk lately about the venture capital world not knowing where to invest in this space, right? Because when you're a venture capitalist, and I'm not speaking as an expert here, but you're looking for a company that you can put, say, anywhere from five to 50 million in, and that company is gonna [00:31:00] become what is known in that world as a unicorn, which is a billion dollar plus company, right?

You're looking for these big exits. The truth of the matter is, A lot of these companies that are starting up now are so small and don't need very much money because what they're doing is jumping off these open source tools. They don't have to build something super large. They can two or three people can get together and do something that say a team of 30 or 50 used to do.

So what does the, what did the venture capitalists invest their money in so they can get that return? Who knows, right? This is the world, we, weird world we're entering. And it's a little bit, wonky in terms of the business perspective, but I think that's something people are talking about in that space is like, what does the middle ground look like from here?

Hey gas. Do you think at the end of the day, does Google have a mote when it comes to artificial intelligence? Good question.

I'm really curious here what gas says about this. 

Moz are for Kevin. Okay. When you're dealing with cutthroat competition and ruthless business practices, the only way to survive is by constantly innovating and adapting. And guess what, bro, that's exactly what Google has been [00:32:00] doing since its inception. They may not have a permanent moat, but they certainly possess the agility needed to navigate treacherous waters

when he hits that diva office when he gets up in the false settle range. That's so good. Treacher waters. Treacherous waters. Okay, I guess Google's fine. You heard it from gash.

Okay. So technology's getting commoditized. The big companies should be concerned, but also the big companies are just still releasing stuff for free.

I don't know what open AI's endgame is exactly. I don't even know if they do, but they released yet another fun 3D toy, which I'm sure in a few months time will revolutionize everything. But today it's just a cool thing to play with. It's called Shape or Chapi Dash. I'm trying to figure that out.

I wanna call it Chappy. Cuz I like Chape. That's what I, yeah. Real fun name. Yeah, exactly. Chappy. Hey, chappy. Chappy new. Hey, Sheey. Due to you too. [00:33:00] Yeah, it's a text two 3D model. Yes. Piece of software you can play with it easily. I spent hours trying to get this to run locally, solving issues, posting on GitHub discussion threads, and then I just found You can run it on hugging face.

Yeah. In fact, you don't even have to pay. The free version that you can poke around with is actually fast enough. You might have to wait 30 to 60 seconds for a response, but basically you give it a text prompt like, An airplane that looks like a banana or a chair that looks like an avocado, and it's going to do its thing and deliver you a 3D object, textured, colored, and you will be able to rotate the model around.

Or if you use a 3D modeling software, you can import the model right into your world. So as earlier we were talking about the sort of holodeck, the prompt to anything happening. Imagine that you are in a Roblox or a Fortnite or some sort of 3D world and you're. Collaborating with friends.

If you've ever experienced Gary's mod a [00:34:00] half-life mod where you can spawn objects into existence, great piece of software, great piece of software, and like in an early version of what all this is happening, right? I feel like, yeah. And so instead of plucking from a 3D model library that already exists, what this can do is generate something unique to your request so quickly.

And even though you might look at the resolution of the birthday cupcake or the bowl of vegetables and go, okay, what is this? The internet in 1995 they look like primitive EverQuest sort of models. Harkening back to our mid journey conversation. Give this thing a few months.

Yes. It's gonna have special shaders on it. It's gonna refract light and have subsurface scattering. It's going to look photorealistic. Yeah. In only a short amount of time. And then I guarantee it's gonna have material properties so that wood actually acts like wood and reacts like it, or metal makes metallic noises like it's not hard to project out from where this is today.

And what's interesting about this is we've talked to, we wanna do a kind [00:35:00] of deep dive on what's going on in video games too, cuz Kevin and I both come from that world. But there's a lot of companies right now working on, worlds in which you can, text to create world or text to create something in the world.

And I think there's some really interesting space in there. I think what's fascinating about this as a toy, it's funny to think about things always feel like toys and then you start to see them become much more useful over time. And right now, yes. Shoppy is like something that doesn't feel like super significant when you go into it.

I tried a couple times and it really did okay on one. Didn't do that great. Another, but again, to your point, open eye just keeps throwing these things out. They just keep throwing things and they're just like, look at this thing. Look at this thing. Part of my thought is that Sam Altman, who's this really fascinating character, he is got positives and negatives.

He's definitely the one who seems to be like pushing forward and being like, here, community use this, use that, use this. And this is that thing which is okay, you had photos, you had videos, and 3D is the next step of this. It is shocking to just, you should go play with it. We'll put a link in our show notes.

If you haven't, if you [00:36:00] can't see the videos here and you're listening, , make sure you go take a look at it because it's very raw and very primitive, but this is the beginning. 

All right, Kevin. I think it's time. You know what time it is. It's time for the demo of the day. Reach the demo of the day.

Okay. Today we wanted to take some of the tools we've been learning about over the course of the show and turn them into a kind of a creative. Venture, let's say. Should we call it a creative venture? Yeah.

Let's call it that for now. Hey I think we can let the audience decide exactly how creative this venture is, but yeah, as we mentioned earlier there's a pervasive energy within the AI community about look grind. Culture is nothing new.

Hustle culture. If you're sleeping, you're losing, you gotta constantly be on the grind. And within the AI community this, it's very pervasive and [00:37:00] honestly, I think at times it gets a little toxic.

You're being talked down to if you dare to like disconnect for a minute or if you don't constantly have a server in the cloud generating you passive income, you're some sort of failure. And yeah, I think that's a big part of this is like chat G p t can make you 15 million passively every year.

And like the TikTok world of this is fascinating. Like it's all these people that have spun up and a lot of these people go from like big thing to big thing. And it is like a thing that you tend to see when something gets a lot of hype. But I think it's something we gotta take a step back for and say yeah, this maybe isn't the best thing for us as a culture.

Yeah. And I'll be honest, I'm just jealous of it when I see a post of like, these four prompts bought me my private island. I'm like, damn, you know what? No, your feet are in the sand right now and I'm sitting here scrolling your thread and signing up for your mailing list because I feel less than no, no, no, no, no. .

We have to torpedo this, but you, or not exactly. We're old Gavin. I'm sorry. I we're tired. We've been through the ringer. We're gonna do this. We're not gonna out grind the youth. That ain't gonna happen. So how can we [00:38:00] weaponize AI and leverage it so that it could out grind anybody?

Oh, enter. We can do that. The grind guru. All right, so what is the grind guru people might have been seeing Kevin be tweeting lately about grind guru and grinding? Might following Kevin for I was getting. People were not too happy that I use my Twitter now as like a test bed for all of the things.

But I was testing out the sort of grind guru methodology, which is, how quickly could I spin up the mailing list? Could I leverage tools optimized for Twitter and have AI generating entire threads of text for me to see if I could grow my follower account, get my engagement up.

And the sad reality, Gavin, is that it was pretty easy. Easy. It worked. It totally worked. Yeah. It worked. And , every day there's new tools released, we're gonna fall back on the ones that we know at the moment.

And we've talked about auto, G P T and we talked about baby agi. Mm-hmm. Before. [00:39:00] These are automated AI agents. So right now, if you want to have if you wanna leverage AI to do something for you, you have to have a one-on-one conversation with it and do a little song and do a little dance to get a result out of it.

The promise of these automated agents, which isn't really here yet, but it, you can see it on the horizon, we have a line of sight to it, is that you'll be able to feed it a persona and some sort of problem or task, and it will auto automatically go off. Come up with a task list, try to execute on those tasks, reflect on how well it has performed, and then give you some sort of result.

In the end, the sort of set it and forget it, the crockpot of ai, you walk away and come back to a completed task. So I got baby B a g I running on a service called Repl. And if you wanna do this yourself by the way, not a sponsor, hashtag not an ad, I wish it were, you can go to Repl, come on Repl, you can fork, which means you can make a copy of it on your own account.

You can fork baby B AGI for free and you [00:40:00] can run it there. I think even if you've never experimented with this stuff, even if you've never used replay, you'll be able to figure it out. They make it pretty easy to clone the, or fork the code. And then it says, put your key here, state your mission here.

They make it very simple. So me, big dumb, dumb meat sack, able to get it to run. I've got. The objective in there, Gavin. And then basically it's a description. It's saying that you are the grind guru. You're all work, you're no bs. And the first task, that's what I wanted. This is what you typed in.

This is your prompt, basically for the baby. B a g I. So the first thing you That's correct. You typed in was this. Yeah. Okay. This is the objective within the code. I'm staring at it right now. It says that you're the grind guru. Your goal is to out grind anybody on this earth. I use some colorful language in there because we always like it when the the AI gets a little adult.

But don't worry, we'll censor all of it for the, the audience that we have. Now we have to define your first task. Okay. That could be. Get a thousand followers on Twitter. That could be [00:41:00] develop the grind set motto. It could be. Wait what would we like for the first task? Gosh, I think for baby d Agi, I would like to, I, I don't mind the grind set motto.

Cause I think there should be something that we use as our, like our lynchpin, like our thing the flag we plant and say this is what the GRA is all about. So let's ask it. What should be the motto of the GRA guru

so now this is an interesting thing. I saw someone tweet out that, 99% of people that are using auto GPTs that are using these automations, they're better served just asking chat G p T. This might be one of those things that chat G P T could.

Give us a better answer on versus it needing to go out and do that. So I'm wondering what can we add to the first task, okay. That would make it something that it would need to go out and crawl the web or come up with other tasks for. So I, here's the thing. Let's ask it what its motto is, and then ask it what Twitter handles we should or what Twitter handles we should reply to with that motto in order to build engagement.[00:42:00] 

So I have define the grind set methodology and who you should target on Twitter to build engagement.

So I'm gonna hit play. 

Okay. It's now going to spin up baby b a g i. It's going to query open ai. With, here's the objective. You're the grind guru. You're all work no bs. It says, new task. So it's defining the grind set methodology, Yep. It says The grind set methodology is a no nonsense approach to achieving success. It emphasizes hard work, discipline, and focus in order to achieve your goals. Sounds great. Yeah it says, to build engagement on Twitter, you should target people who are interested in your personal and professional development, such as entrepreneurs, business owners, professionals.

So that's all good. Now, what it's done, which is interesting, is it's generated a task list. And this is what these auto GPTs do so well. So one is define the grind set methodology, which it did. Two. Yeah. Identify top Twitter influencers in personal development, self-improvement, motivation, and inspiration.

It's searching the web for that right now. Three. Okay, that's good. It says it should analyze the [00:43:00] Twitter profiles of identified influencers for engagement strategies. Four. Okay. Create a list of powerful no BS tweets to engage the target audience. Okay. And then five, develop a daily Twitter posting schedule to maximize engagement.

One thing we should say is that when Kevin's watching this stuff, so this is it going out in real time and finding this, he's basically sent these AI agents, quote unquote, which are these little knowis that are going out and searching, and each of them has their own job, right?

We've talked about these AI agents before, but it's really interesting when you send them out to do something. It's basically like having a little company and then your little company goes and you send each PE person to do this thing and then come back. And this is what people are so excited about with auto G P T and Baby a g I and AI agents is this idea that, okay, my time is precious.

I don't wanna waste the time doing what auto G P T is doing right now. It could just bring this stuff back to me. And there's been mixed results so far. But for something like this, it generally works. Okay.

So it, it says that it completed the research. It hasn't shown me what [00:44:00] that research is just yet, and it's now looking at the top personal development Twitter influencers. Okay. And it's trying to parse that as well. So I, it's doing it. This is a project that's going to take some time.

So I think, yeah. Maybe the best thing that we can do is. Help our AI overlord here. Our AI c e o, which is running this thing. The Grind Guru. Yeah. Maybe we could start a Twitter account for this Grind Guru. Yeah. And we can get its messaging out as it is developing it, and we can give an update next week as to how, I think that's how great Idea Grind is going.

I think this is something we're gonna try to track over time. It's definitely something that we wanna do with some more ambitious things like this. Could we ask Gas what a good name for the Twitter handle for this would be? Do you think we could get it from him? Yeah. Okay. Of course we can. I think he'd be the best person to name this since he's gonna be the face of it. 

Hey everyone, I'm super excited to announce the official launch of my new Twitter handle at Get Grind Guru. If you're ready to level up your life, hustle like a and join a community of like-minded grinders, then this is the place to be When you follow it, get [00:45:00] Grind Guru. You'll get access to exclusive content that you won't find anywhere else.

Oh, I'll be sharing my most powerful productivity hacks, no bullshit motivational speeches and game-changing networking strategies that have helped me become the person I am today. Okay, and trust me, wow. I'm just getting started. Woo. But it's not just about me, it's about all of us growing together, okay?

When you join the Hash Grindin guru community, you'll be part of a tribe of unstoppable go-getters who are, of course, pushing their goals and pushing each other to new heights. So what are you waiting for? Head over to Twitter and follow at, get Grind Guru now and get ready to experience a life of relentless hustle, unwavering discipline, and ultimate success.

Let's fucking do this. Yeah. Wow. Yeah. I'm ready. I'm ready. You grind I think that's what's interesting and also a little nauseating about the space. Is that all of those platitudes, all of that stuff that it just spouted we see time and time again.

[00:46:00] Yes. From the grind set hustlers out there. And so I guess our new mission in gash we trust is to out grind the grinders out, grind the grinders out, hashtag out, grind the grinders. Man, let's get going. That's amazing. 

We just created more work for ourselves though.

We're gonna have to make at Get Grind Guru is gonna be the Twitter handle where gash Yeah. We're gonna do our best to automate it and stay true to what the ai says and responds with. We're gonna use a combination of baby B a g I to give us tasks and give us research. We'll use chat G p T to to get some grind set words of wisdom and to respond to you.

And I think we'll start with what it means to to live the grind life, right? Like maybe the 10 Grind Commandments of the Grind set. Can you tell me what they give me a couple of them. Just give me, I don't, you have to have all 10, but we can put them up here, but I'm curious to know what a couple good ones are.

All right. I'll give you the first two. The first two Grind Set Commandments as decreed by gash are Grind Guru.

Thou shalt prioritize th hustle. [00:47:00] Above all else, wake up early, sleep late, and never stop grinding. Sacrifice, sleep, comfort and leisure. Time to achieve greatness. Remember, if you're not hustling someone else's, he's right. Thou shalt set massive fucking goals. Dream big, aim high, and never settle for mediocrity.

Your goals should be so massive, they scare the shit out of you. This is the hash price. 

Yeah. Oh man. I'm signing up for that mailing list. Give me that 10 D v D course. That's incredible. Our grind guru he's gonna take, can't wait to set my goals so high that I can't even imagine where they are.

Congratulations Gas. Congratulations at Get Grind Guru. We will start this Twitter handle, so go follow it. But we're gonna try to keep this as something that we're not gonna necessarily associate with our show right now. This is like a little fun thing you'll be able to get if you're listening to our show.

This is super fun. It's gonna be a great time Before we go, cause we have to wrap up pretty soon here.

. We have one last fun thing. We always like to mention a little toy. This is something Kevin you [00:48:00] saw. I think it's called Skybox Labs. You wanna tell everybody what this is? Yeah. Blockade Labs is the company. I will put a link in the comments and the show description, but Skybox Labs.

We talked about 3D models, prompting images, prompting speeches. Now you can prompt a sky box and if you don't know what that is, any video game usually is set. When it's set in a 3D environment, it has a sky box, which is a big, giant texture that envelops the entire world. So whether it's Call of Duty or Fortnite, if you're looking around and the stars and planets, or a sky, that is the sky box.

This 3D texture that envelops everything. And so Blockade Labs has a text two Sky Box toy that you can play with right now and you can feed it a prompt. I, I fed it RV by a lake and was able to generate a sky box as if you were inside an RV at a lake, and it was pretty cool. And you can choose the style as well, so you can make sci-fi stuff modern computer anime, animation, fantasy lands, yeah, anime[00:49:00] oil paintings, pen and ink, low poly.

There's all these different styles, but you can generate a sky box and then use your mouse to click and look around, and you'll see that it's doing a really good job of generating some high quality artwork. It stitches it together very nicely so you don't see the seams as you're looking around.

You can take someone else's prompt and remix that and describe your own variation of it. And this will give you a good free pretty instant gratification feeling for what it could be like. To prompt anything into existence. If you were making a video game, for example, you could use this as step one.

I'm just giving the machine some information and now it's generating the world that I get to look around in. It's, again, it's the beginning of the holodeck, right? Like it's this idea that all of these tools coming together to create a virtual environment, and this is like another part of it, right?

Imagine again, a world where you have, images, . Videos, 3d, 3D models, plus this world generation software. All that's gonna come into one place. And we've talked about this before, [00:50:00] but you've got this very fancy set of AR glasses coming out very soon, you know, and conceivably a month, Apple's gonna announce their big air glasses.

Everybody thinks of VR and ar as cuz it also ran, it will be a giant deal because this is the beginning stages of, oh, I want to have my, Tokyo downtown in an anime style pop. It's there. You have a virtual environment. It just becomes so much easier to do this stuff. And I think that's gonna be a big.

Turning point, but we're, again, we're in the toy stage, right? This is the beginning of everything and we see it all coming together. And it's gonna move fast. It's gonna move fast. Look, I am very satisfied with today's AI discussion. I think we have we have our marching orders.

Yes. We have to get the grind guru out into the world. Should we end with a commandment or two more? Yeah. Let's ask, let's we got gas is gonna give us one more grind guru commandment. 

Thou shett embrace discomfort like a badass. Success lies outside of your comfort zone. Embrace challenges, face your fears, and grow stronger through adversity. Pain is temporary, but [00:51:00] greatness lasts forever. So true. Thou shall cultivate relentless discipline. Discipline is the fuel that drives the grind.

Yes. Develop iron clad routines. Stay consistent. And never allow yourself to slip into complacency. Discipline is the motherfucking backbone of the hash grind sale. Yeah. I am ready. I am. I'm engaged. I am ready to go and make my life happen today. I'm ready to make it happen today. Thank you. Cash, money, power, money, power.

He's so right. 

We also don't know what gash looks like, by the way. So maybe a little call to action for the AI for humans on. I think it's a great idea. Mid Journey five. You stable, diff fusion, mid journey, whatever you want. Yeah. Use whatever tools are. If Ms. Paint, we don't care. It doesn't have to leverage ai, but what does gash look like?

Because I'm assuming at some point Gavin, when he decides to like, put his branding out there and come up with logos and maybe do inspirational videos where he's talking to his grind set community. We [00:52:00] gotta know what gash looks like. I can't wait to see what gash comes up with for a logo.

Cause I gotta imagine it's gonna be pretty remarkable. I would imagine there's gonna be some real like alpha male energy in the logo in some form or another. Do you know what I mean? That's what I get the sense of we're gonna be approaching here. It's gotta be like, it's gonna be a some sort of powerful animal smoking a cigar, I bet with a giant like ripped bicep in front of it.

It's like one of these, like a big cigar. Ah, yes. Ah, and a tattoo of itself on its bicep, smoking a cigar. All right, everybody, thank you so much for joining us again. We love having you here. It's so much fun to make this show. We really appreciate everybody. Please go and subscribe to our videos on YouTube. We have a very popular audio feed as well, so either you're watching us on YouTube or you're listening to it. We're on Apple Podcasts, we're on Spotify and review us on those platforms. It is really important. And by the way, it's been super cool for us because.

We are growing over time and like it's clear that people are [00:53:00] enjoying this and that's been super fun for us to see. We really appreciate everybody watching and listening to what we're doing here. Yeah. Sincerely and I, we joke about smash the bell and click the thing. It really does help because this is a brand new something for the both of us. Yes. And all of the feeds, whether it's YouTube or our Twitter handle AI for Human show or even on Spotify, if you leave a positive review, if you give it five stars or maybe the four that it deserves, that's fine.

It actually. It actually makes a difference. Helps. It's super impactful cuz that pops us invisibility. So thank you to everyone who's taken the time to do that. Yes. And thank you for referring your friends and it's gonna be a really interesting moment when the grind Guru surpasses us and popularity and engagement which happen, which you definitely will very soon.

Definitely will. A hundred percent.