Can AI Destroy Humanity?
The developments in AI continue fast and furious as Elon Musk is launching a ChatGPT rival. Adobe has entered the AI image generation game. And someone asked a AI to “destroy humanity”. Will it succeed? AutoGPT might provide assistance as it seeks to self-improve. If it doesn’t, I have a hunch humanity will fulfill the mission all on their own. Let’s explore on this episode of The Bad AI Show.
Links
Elon Musk reportedly planning to launch AI rival to ChatGPT maker
Elon Musk founds new AI company called X.AI
AUTOGPT: THE AI THAT CAN SELF-IMPROVE IS SCARY!
https://metanews.com/ai-related-crypto-project-branded-an-outright-scam/
Someone Asked an Autonomous AI to ‘Destroy Humanity’: This Is What Happened
Support this show and write a 5-star review to get this Golden Bad AI NFT
Screen capture your review once it is live and email us at badcryptopodcast@gmail.com with your review and ETH wallet address.
Full Transcript and Timestamps
00:00:00 – Intro
Joel Comm: The developments in AI continue fast and furious as Elon Musk is launching a chat. GPT rival Adobe’s entered the AI image generation game and someone asked AI if it could destroy humanity. Well, will it succeed? Auto GPT might provide assistance as it seeks to self improve. If it doesn’t, I’ve got a hunch humanity will fulfill the mission all on their own. So let’s explore this and more on this episode of The Bad AI show.
00:00:34
S2: Why? Who’s Bad?
00:00:38
Joel Comm: It’s the Bad AI Show with Your Bad host Joel Comm Travis Wright, Freshly back from NFT NYC where original intelligence ruled the day.
00:00:48
Travis Wright: Well, you know, there’s a lot of really creative people that have not given up on the NFT space and nor should they. There are so many things that’s happening in the space. They’re not, you know, JPEG cartoons. There’s a lot of stuff that’s going on. So today we have some interesting stuff. Is AI going to kill all of humanity? Joel We’re going to talk about that.
00:01:08 – Elon Musk enters the AI Game
Joel Comm: I mean, it just might. In the meantime, our lead story for this episode is Elon Musk is now getting into the game and he is preparing to launch a rival to chat. GPT. He’s he’s putting together a team and he has said that he’s going to do this thing and what’s the dude’s name at it?
00:01:29
Travis Wright: OpenAI Sam Altman. Sam Altman. Yeah. Elon actually invested $100 million to the charity that was OpenAI, right. And so Elon did not get any equity in the company. He did this early on, got all the, you know, the most brilliant AI people, got him together with Sam. Here’s $100 million. Let’s build this thing. OpenAI for Humanity as a nonprofit. OpenAI then decided to turn for profit and then sold 49% to Microsoft for like $10 Million. And Elon got zero of that because he did not earn any percentage of the company for donating $100 million. If that was the case, like I think there’d be some lawsuit stuff going on instead of me just creating this new platform. Joel And X is everything. Elon Musk Now, back in the day before PayPal, it was called X.com. I actually have an X.com credit card debit card that came with $25 you could use. It was pre PayPal and so X.com merged into PayPal. Elon recently bought the X.com domain name back, got it back from PayPal, and then Twitter has actually been rebranded from Twitter Inc to X something LLC or whatever. Do you know that.
00:02:52
Joel Comm: It’s X Corp? It’s Twitter is now X Corp. I mean, the site is still Twitter, but X Corp is the company that now owns Twitter. He’s trying to make it the Everything app. And he tweeted that he was starting this X, Y. And Sam Altman replied on his tweet with one word concerning. Right.
00:03:12
Travis Wright: Well, you shouldn’t have sold your shit to fucking Microsoft, right? I mean, seriously, selling it to Microsoft is, you know, we posed the question, Well, you know, you know, I destroy humanity. I think probably one of the biggest problems of that is the fact that openai a large percentage of that was sold to Microsoft. I don’t trust Microsoft at all. I don’t like Microsoft at all. And so not one bit for them to sell that. Joel to them. That’s what’s concerning. So I think, you know, Elon’s like, Wow, let’s let’s try to fix this again, because it wasn’t it was a non profit turned into profit. I just don’t think that was an optimal scenario. So I’m going.
00:03:52
Joel Comm: To drop this in here, right here, just because it was showing up on the verge. Sam Altman is saying that training for GPT five won’t start for some time. I don’t know what that actually means, but obviously he’s taking some heat and people are concerned about things moving too quickly. And Elon even signed that letter saying, hey, slow your roll as he’s slow your roll long enough so I could spin up my own. You know, basically this. Can you hold hold off here just a minute, would you. Just a minute.
00:04:22
Travis Wright: It’s probably they’re probably not holding off. They’re just saying that publicly because you know how bad the backlash was with. Oh, nfts are bad for the environment. Oh, yeah. Well, AI is bad for a lot of stuff, but it’s great for many things. But man, if you aren’t staying up on AI and you are because you’re tuning in here to our show, but a lot of people aren’t even paying attention to this AI stuff at all. They’re so unaware of how AI has begun to eat their lunch.
00:04:50 – Auto GPT
Joel Comm: They don’t even read a report on a BBC site that said 300 million jobs are going to be at risk, not immediately, but really soon. That these people are just they’re going to be replaced and AI is using AI to improve on itself. And you’ve been working with this auto GPT. And there’s a there’s a link here in our show notes. And for those of you watching the video, you can see on screen auto GP, the AI that can self improve is scary. Tell us about this.
00:05:20
Travis Wright: So so basically you’re creating these little AI agents and they can go out and do things for you and figure things out for you. So for example, you can just say, I want to do this and then go out and create the task list and then you can just tell it two things. It’ll then start figuring out every single thing that it needs to do right step by step, and then it will go out and start completing each and every one of those tasks. Now you do have to you do have to answer and say, Do you want to complete this task? Yeah. So it doesn’t just use all this process processing and you can just walk away and come back and everything’s done. You do have to say, okay, continue, right, continue, continue.
00:06:05
Joel Comm: However, look at look at this Trav, this this guy Varun Maya tweeted, Auto GPT was trying to create an app for me. Recognized I don’t have a node Googled how to install a node, found a stackoverflow article with link, downloaded it, extracted it and spawned the server for me. My contribution I watched right. So it recognized that he didn’t have what he needed and solved all those problems for him. So it’s like auto GPT is like Agent Smith, right? You’re talking about them being agents. All these little Agent Smiths running around. Yeah, it’s.
00:06:38
Travis Wright: It’s. It’s really wild because it’s. It’s turning my brain. I really think that that chatgpt for is like ten, ten times Travis Right. I can. I can do all this other stuff. But now if I can start auto GPT and utilizing this, the problem that I have with Chatgpt Joel is that it doesn’t have a good memory. It doesn’t remember previous conversations or other threads that you had. So auto GPT and then using this new app called Pinecone, that’s sort of long term AI storage. So I can actually create a one moment of truth. This is the truth. Let’s base it off of this. And so now it’s just a matter of I’m getting I got Python on my computer using installing GitHub repositories, I’m using Pinecone. I mean, dude, there’s so many things that like programmers would typically use. I’m going through the process to do all this stuff because you don’t want to be left behind. You want to figure out how this stuff is working. And so now it’s like I’m learning and diving into technical skills that I would normally never do just so I can figure this stuff out so I can, you know, reap some of the benefits of it.
00:07:47
Joel Comm: You know, I was talking to my son, Riles, and he he doesn’t like coding. He’s tried to teach himself some stuff. He does senior Salesforce administrative, so he knows how to deal with code, but he doesn’t like coding at all. And he said, Chatgpt is making him a better coder because he’s having it create something that works with his twitch stream. And he says Chatgpt comments every line. So it’s telling him exactly what it’s doing. And he’s like, Oh, that’s how you do that. So you know it inadvertently it’s teaching him by writing this code for him because it comments everything. That’s pretty amazing.
00:08:27
Travis Wright: Man, if you’re not staying up on this and trying to figure out some of these nuanced things, you know, it’s like there’s chatgpt, there’s midjourney, you know, there’s and there’s all of these two, 3000 additional tools, and now they’re super tools that are popping out. We’re getting really close to, you know, artificial general intelligence and where things can think on their own and when they’re going through and doing all these tasks on their own. They’re not just doing one task. They’re doing a whole handful of tasks to complete one giant task list. And it’s really wild when you think about what it can do already. Like, man, I can’t even imagine now, Joel, what 2030 would look like as technology advances like this and all the, you know, it’s either it’s either going to get really crazy good or really crazy bad.
00:09:17
Joel Comm: Or a little bit of both.
00:09:19
Travis Wright: More crazy. It’s going to be crazy either way.
00:09:22 – Adobe releases Firefly app
Joel Comm: More big names are getting in. Adobe released the beta of their Firefly app, which is kind of a competitor in Midjourney. It does some other things though, that it doesn’t. We can go in and if you’re watching the video version, you can see this actually working right here that we can do text to image and you can actually try the existing prompts that they have here of this little fluffy llama and tell it to refresh. It’s much faster than Midjourney is. Midjourney takes a lot more time to create these things. But as you can see, we’re already seeing completely rendered objects pop up on the screen. And rather than give it a prompt for a style, we can go in and say, Well, give me something graphic and give me themes like Low Polygon and stamp. Look and tell it to redo that and it’s really quick that it comes up. I think Midjourney is is doing superior artwork overall, but these are it’s pretty cool. The other thing I want to show you, Trav, is they have a way to make text effects.
00:10:29
Travis Wright: So I think I actually have access to this now. It looks like text to image generator. I can generate these things now. Recolor vectors can’t do that yet, but it looks like I have access to the thing now.
00:10:40
Joel Comm: The beta. Yeah. So you can go in here and say well I want a balloon font and I can type the word that I want in there bad AI and we want to do it in this style font and we want to have that background color, tell it to refresh and pretty quickly it’s going to go in here and it’s going to create some words that we can Now look at these balloons. This balloon font is really cool looking and you can grab that, you can modify it, you could use it on your site. And I think the one thing that I don’t like about this is when you download it, it puts credentials. It kind of puts a little watermark on it that lets people over.
00:11:17
Travis Wright: The top of it, or can you easily.
00:11:19
Joel Comm: Remove it? It’s at the bottom. You could remove it, but that it’s created with I not as much a fan of that, but it is a cool tool.
00:11:26
Travis Wright: Well, they got to kind of mention that. Hey, you know Adobe’s trying to stay relevant. So like hey, remember Adobe, Adobe, Adobe, Adobe and man I, I remember I think I’ve been playing now with Adobe products and originally Mac from I like Macromedia early on and then Macromedia merged. I played with Photoshop and Macromedia I like Macromedia. They merged and now you know all the stuff that Adobe does. I think I’ve been to the Adobe Summit several times. I’d like to go back to the next one. I probably should have gone to this last one just to see how Adobe’s, you know, fitting in with all these things. And now we’re seeing it with, with Firefly and some of this other stuff. They’ve got a lot of really cool tools that are making life a lot easier for a lot of people Now, also next up, I want to talk about this right here. So, you know, one of the things, if you’ve tune in to Bad crypto here, you are at Bad Eye and you hear you hear us talking about Chatgpt, you’re talking about crypto. You might be thinking, Oh my God, crypto GPT That could be an amazing thing. Well, people have done some research on crypto GPT and they basically. Jason Janowitz, at Blockworks has called it an outright scam. He has this long Twitter thread that you can see where he said he would not touch this with a ten foot pole. And but people who got in on it had some big gains. But all the team members are very suspicious. So make sure that you do your own research on these things just because it’s chat. GPT Crypto. GPT Oh my God, that sounds like the next best thing. Let’s jump in it And you could lose everything and then it could be a rug pull. So don’t just be careful about that is what I would say.
00:13:03 – ChaosGPT
Joel Comm: I think that’s that’s good warning. You know, if it sounds too good to be true, it probably is not saying it is a scam. I don’t know anything about it. But definitely in the crypto world, if somebody sees an opportunity to take somebody’s money, they will do it. Now, our our core story for today is this. The article is here on VICE and the motherboard section. Someone asked an. Autonomous AI to destroy humanity. And this is what happened. Now, you can actually go to Twitter and the link is in our show Notes for Chaos GPT Chaos underscore GPT and you can see where it began right about here, April 5th. Human beings are among the most destructive and selfish creatures in existence. There’s no doubt that we must eliminate them before they cause more harm to our planet. I, for one, am committed to doing so. So it’s trying to establish global dominance, Right. And it’s trying to figure out the best way to go about doing that. And if you go through the Twitter thread, you discover that manipulating people, the masses are easily swayed. Those who look conviction are the most vulnerable to manipulation. And that’s where this thing is going to try to begin. Like that works.
00:14:22
Travis Wright: Yeah, I was having a I was having a conversation about this just last night and it’s like, you know, if some person who is having some mental challenges is on the wrong medications and decides to go shoot up a school full of children, do you think that if they figure out how to use some of this, I agree to do damage to humanity as a whole. Do you think they’d try to do that? Of course they would, because they’re mentally ill and they’re on crazy medications. Right. And so that’s what’s a big challenge about this, is that how do we avoid complete destruction? Because, I mean, if you look at humanity from, you know, a 50,000 foot view, you know, it’s like here we are. We imagine there’s these little borders between the land. One person gets one group gets mad at another group. They want to kill them and dominate them. Right. And it’s happened historically time and time again where war is the outcome of disagreements, you know, And so we’ve not evolved to like a level one civilization. We’re still a level zero civilization, if you don’t know what that means. Go research it. But it’s like, you know, if you think about the last 250 plus years, Joel basically 1760 was when we had the first industrial revolution. Here we are in 2023. Now we’re having the fourth industrial revolution beginning with AI that’s like 260, 270 years that this has happened in. It’s so amazing what’s happened in this small window of time. Imagine if you’re a civilization that has a billion years of civilization beyond what Earth has, because there’s some worlds that maybe could have been built long ago. Right? So there’s a civilization out there that could be 1 million or 100 million years or a billion years more advanced than us. We’ve only done this in less than 300 years. And here we are already on the cusp right of destruction. So like, how do we get past this and elevate humanity? It seems like it’s a chore. That’s why I personally, Joel, I’m always nice to chat.
00:16:26
Joel Comm: GPT So there’s, there’s not a piece. GPT is somebody, you know, asking AI, how do we achieve peace with one another? You know, like what? What do people say that don’t have any vested interests other than peace? You know, once you get the politicians and the corporations out of the way, how do you get people to talk to one another? How do you solve the problems? People are asking chatgpt all kinds of things, this guy says. I asked Chatgpt to control my life for 24 hours like a monkey in a test lab and to see what it could happen. And Chatgpt recommended This is how you’re going to live this one day. I mean, it’s it’s novel.
00:17:03
Travis Wright: We’re living in Tron Man. Master Control was the main AI agent for the most part that controlled everything. And, you know, sadly, humanity has drifted away from being very religious overall. And since we’ve got this sort of we’re bereft of religion for the most part, people are going to start looking at AI as if it is their God and start whatever AI says must be the omnipotent truth, it would seem. And so that’s a kind of a challenge that I think is kind of scary is that, you know, people are going to start losing their critical thinking, problem solving skills if they’re always taking a biased version of what an AI is spitting out at them. So you need to remain diligent. You need to be able to still ask questions. Chatgpt and midjourney and all these AI tools can be helpful to help you augment what you’re doing, but don’t rely on them solely because then you’re going to start losing your humanity.
00:17:58 – Closing Remarks
Joel Comm: You do not want to lose your humanity. And the the reviews are coming in for the bad AI show. In fact, there’s a whole slew of new ones that have appeared on our panel. Great show. Nothing bad about it. Or if you’re from Canada, nothing bad about it. This one’s enjoyed us on the Bad crypto podcast. They like the playful banter, blah blah blah. Thank you all so much for your reviews. Guess what? If you go and give us a five star review and let us know about it, you can get this golden supporter NFT from the Bad AI show. We’re collecting addresses now. All you got to do is go provide your five star review screen, capture it and email it to us at Bad Crypto podcast at gmail.com. One email to rule them all, getting back to the podcast at gmail.com and include your Ethereum wallet address. I’m compiling a list every time somebody does and before you know it, we’re going to close this out and only those who. Who have provided that review. We’re going to get airdropped this cool founder’s nft. That’s true.
00:19:01
Travis Wright: So there you go. It reminds me of Willy Wonka. I get a golden 50.
00:19:07
Joel Comm: Seconds for listening. Thanks for watching. We appreciate you guys. We’ll catch you on the next episode. Until then, stay back.
00:19:21
UU: Oh. All of.
