I'm Genuinely SCARED of AI...
Boxmining
5 days ago
The summaries and transcripts on this page are generated with AI technology and may not perfectly represent the content of the video. Please use the information as a guide only.
Executive Summary
The video titled "DLMMing $GOLD with TokenGuy" features a live discussion and demonstration of liquidity mining (DLMM) for the $GOLD token from DeFi Dungeons on the Solana blockchain. The presenter, Seb Montgomery, is joined by TokenGuy to explore the token's launch, liquidity provision strategies, and the broader context of the DeFi Dungeons game.
Key Points:
-
Token Launch and Liquidity Mining:
- The $GOLD token launch was oversubscribed by 43 times, indicating high demand.
- The team encountered technical issues with Meteora, delaying the withdrawal of unused USDC from the vault, which impacted liquidity provision strategies.
- The token price experienced volatility post-launch, with some participants selling below their purchase price, creating short-term opportunities for arbitrage.
-
DeFi Dungeons Game:
- The game is an RPG-style idle game with interactive elements, set to launch by the end of March.
- Players can earn and trade in-game assets, including NFTs and the $GOLD token, which will be used for summoning heroes and other in-game activities.
- The game features an in-game marketplace, referral systems, and cosmetic customization for NFTs, aiming to create a dynamic player economy.
-
Liquidity Provision Strategies:
- The presenters discussed setting up liquidity pools (e.g., 100-bin step 1% pool) to earn fees while accumulating $GOLD.
- They highlighted the importance of monitoring volume and adjusting strategies to maximize returns.
-
Market Dynamics:
- The $GOLD token's demand is expected to surge closer to the game's launch, as players will need tokens to level up characters and participate in the economy.
- The presenters emphasized the potential for long-term engagement, given the game's depth and the team's track record from previous projects like "The Heist."
Airdrop Summary:
- Core Functionality: The $GOLD token is the primary utility token for DeFi Dungeons, used for in-game transactions, summoning heroes, and marketplace activities.
- Recent Updates: The token launch was highly oversubscribed, and liquidity provision is currently active on Solana-based platforms like Meteora.
- Reason to Farm: Early participation in liquidity mining could yield high returns due to the token's utility and anticipated demand spike post-game launch.
- Participation: Users can provide liquidity to $GOLD pools or acquire tokens directly from exchanges post-launch.
Transcript
Transcript
So this video is a little bit different from every other video we made here and I kind of want to talk about AI. So I mean obviously we've seen a lot of these news articles on AI that it's going to replace our jobs, just kind of new cool things. But I kind of feel like we need to really have a serious talk about this because it's not going to impact just the global world, but it's really going to have a huge impact on the crypto field overall. And I think instead of hyping up a particular project or funding it to help, I feel like it's better for me to just share with you guys how we're using AI, some concrete examples of what makes AI really useful for us. And in many ways I'm scared. I'm honestly scared because I literally studied in this office today. We actually have 20 people working here. And I'm just wondering what would it be like in three years? You know, three, five years that once AI technology matures, you know, who's going to be doing what? Because not everyone's going to have the same roles and objectives because I feel like AI is going to easily replace some of these jobs. And I'll show you how we're already using AI to replace a lot of these jobs. But what's also going to extend functionality? I think like it's a point where, you know, I feel like there's a big opportunity to grow as well here. Like AI is going to enable us to do so much more, like namely build an entire app. And it's something that I mean, I spent a lot of time on this. It's like, I'll show you guys this in a sec, but it's something that like AI just allows us to just write code and make something credible. So that's what I'm going to do for this video. Ron is going to share some experiences with you guys as well. And I'm going to have a secret in this video too. Not the classic box money style, but there's a secret here. There's a secret airdrop for you guys who pay attention. So somewhere along the video, there's going to be a QR code. I'm not going to give you too many hints of how to activate the secret because you kind of have to figure it out. Only 50 people will get to win this free airdrop. What it is, is a free NFT that contains ENGINE. So one tip here is you'll need the ENGINE wallet. You'll need a phone. And yeah, just be very observant. Something's going to be interesting. It's going to pop up and you're going to have to figure it out. 50 people can claim this. So it's free money basically. It doesn't cost you anything. Big! Let's go! Alright, Ron. I mean, we've come a long way. It's been four years and Ron's job role actually changed. Yeah, I was a writer when I first came to Vox Mining. And even with ChatGPT3, my job changed. You guys didn't need a writer anymore. No. We used to have writers to write articles and airdrop articles. That's how you started. It's still in the airdrop history, but now we used to write airdrop articles and we used to even hire external writers. And they'll write for like, I think it was like 10 cents per word or something like that. Yeah. So they'll write articles and it'll be like paying like $20, $30 for an article and airdrop. Now AI does that, you know. One click. In the LIM, send it and go, right? Yeah. And that actually scared me a lot because that was a concrete example of like, and I reminded you earlier, I'm like, Yo, Ron. Yeah. You know, things aren't looking so good in the writing space. Yeah. Gotta have to slightly shift. Yeah. Fun fact, I was actually very close to getting fired. And so I had to evolve. I had to adapt. And thankfully, Michael here gave me an opportunity to prove what can I do more. So I had to do videos. You had to. You had to. You had to. It was no choice, right? And I think now even research was so much easier. I think like with Grok 3, I think you were using this as an example, you know. Oh yeah. Grok is also really good. But for me personally, I do. Perplexity. Super good. It basically does all of the preliminary research for me so that it just cuts my research time in half, maybe even less, maybe even more. So it's insane. Yeah. So now they can start playing on your phone. I caught him like multiple times. He's very smooth about it though. So I mean like even simple things that's like what happened with the mantra that will airdrop. And it's up to date. You know, now the new AI, they're doing web searches for what happened. And then it tells you, you know, March 18th, there was a linear vesting, blah, blah, blah. So it gives you up to date, accurate information. So unlike before where it's like maybe telling you guys what's happening in 2023. Now it's up to date and it knows what's happening. It's quite scary. Yeah. It doesn't lie. Before the problem was the AI would lie to you. Yeah. Playing random dates, random information. And I had to cross check. And for perplexity, I've been using this for about six months now. And with ever since DeepSeq came in too, which forced every other to improve, I've been cross checking as well. So it gives preliminary research, but I always cross check the information, check the sources that they're correct. And so far about 97%. I logged this in my spreadsheet. Really? And so it's 100% accurate. It's correct. Yeah. That's scary. That's, uh, that's getting there. It compiles everything you need to know what the details are. And that's it. Then your job or rather my job is to find more strategies that the AI can't really think for you. Because what it's really good at right now is just researching. But for Grok, this is a little bit different. Grok has the think model. Yeah. The think model where it attempts to solve the problem for you rather than compiling research for you. Yeah. Video, think of some ways to improve it, right? Yeah. I mean, that could work, right? Let me see if that works. So it's like a thinking model. I want to, I need a short video like this one. I'm quite curious if Grok will actually, some ideas. Because I feel like this is something that AI, like I used to really bounce ideas off people. And, you know, like just sit there and like, yo, how do we improve? Like we'll have a brainstorm session. We're like, yo, Ron, how do you improve the video? We'll watch the video together. And we're like, yeah, we should do this. And now Grok can do this, right? Literally. It does take some time though. I think that's the current limitation of AI. Like if you want something really smart, it does take some time to figure it out. And it's scary actually how much work is like, there's a lot of templates out there right now. So say for example, you have a great idea for a startup, great idea for a meme coin or something. It can actually start, you can ask Grok to ask you questions. Kind of like a bouncing board to figure out, oh yeah, you know, like does this work? You know, what's your idea? Oh, here, stronger engagement introduction. Oh, I think our engagement's bad. Better context. Oh, wow. Clarity. It even explains why is it better. Oh yeah. Jesus. Why it's better. So, I mean, you can see that this is like stuff that we'll be doing, we'll have meetings for, right? We'll have meetings for this. But I feel like this is one of the situations where like, you can upload a video, already get Grok to, you know, write some feedback for you and you're good. You're like, oh yeah, just read some notes and then, okay, good, great. That's actually quite scary. And I think like, you know, the reason why I'm making this video is because this is like directly applies to what we're doing here. Yeah. Right. I mean, instead of giving some like non-concrete examples, I'm just saying, okay, here, here's what actually happens. That's quite scary. Actually, we should actually look into this, like how to do it. So, the other thing that I'm really doing recently is we're transforming this. You guys love AirDrops. We had this AirDrops spreadsheet for the longest time. Yeah. And it was getting really hard to edit. Yup. A lot of backlogs. Yeah. And then even editing a spreadsheet was a big pain. So, we have made an entire website called AirDrops. And it basically summarizes the information in a nice presentable format. And we've got the full guide on here. We've got opinions on here. We've got our videos all sent together. This was actually generated by AI. Yeah. That's the crazy part. It's built by AI. It's actually quite insane. I literally gave AI something like this. I was like, here's our spreadsheet. Make this cool. And it did take a lot more time to actually tune everything to make it work. But it was using a combination of cursor and windsurf. If you guys want to use and play around with this, basically you can just talk to the AI and run something. If you guys want to use it, there's some codes down below. I don't have referral codes for it. But what I was saying is that this is actually very feasible to do, even if you don't really have any programming experience, which is actually really scary. It kind of made all these components, combined them together. I even added this component here, which is the latest tweets by the project. So I kind of envisioned this as the one stop for all your AirDrop needs. It's got the latest tweets. It's got all the intel on Monad or whatever you want to look at. It's got the tier list so you can see what's happening with the tier list. And then on top of that, it's got these additional features here. One is a Twitter monitor, which I kind of wrote last week, which is like it monitors Twitter. They use AI to scan the contents of the tweets and then it marks down whatever is an AirDrop. Whichever is AirDrop related. Yeah. Yeah. Because the project might tweet, most of their tweets are engagement, GM, hello. Filter that stuff out. I don't want to see any of any GMs. I just want to see the break-in information. Yes. And there is a limit to how far they're down you can scroll on the next page. Right? So if you just want to get straight towards AirDrop related information, this Twitter monitor right here, the AI scours it. Yeah. And actually you gave a lot of suggestions for this as well. These are the people that you're monitoring, right? Yeah. So it's like monitoring people and projects for tweets. And I'm going to improve this in a sec, because I realize now like it's confusing giveaways as AirDrops, which is bad. But yes, this is tunable by next week, so it'll be fixed by next week. But it shows that it's conscious, right? Because it's thinking giveaway equals AirDrops? It knows. And the other one that's quite scary, I think this is the full use of the AI, its power. So this is also trying to solve a problem that we have, which is like, I want to figure out what other people are talking about when they're talking about AirDrops. Okay? For two reasons. One, to see what's popular. And you kind of want to gauge, like you don't want something, you want to see what other people are talking about. You want to see what the scenes like, the vibe check. The vibe check. But I honestly don't want to spend 31 minutes. Sorry, Ron. I know this is your video. But I don't want to spend 31 minutes watching this, right? Like, or I just want to have a quick overview. So this is all AI right now. So basically it scans for YouTube videos that's coming out, or came out, and it has a full summary. Look at that. That's actually quite, like, it actually has a full transcript and a summary. Yeah, this is very powerful, especially for experienced airdrop farmers. Think about it. Like, maybe your day one, day two, you'd watch videos. You'd have the patience to go through the videos, especially if you don't really understand the environment, the context. But once you're like weeks in, months in, farming, multiple projects, you just want to cut to the chase. Airdrop farming takes a lot of time every day. So, guys, still please watch our videos. But, you know, cryptocurrency slay... Ehhh! Just read it. And then, yeah. Just watch ours. But yeah, so this is like fully automated. But this was actually quite a big pain, because to get the transcripts, you need to use Whisper. Mm-hmm. So, it needs to download the video, use Whisper, transcribe it, then send it to the LLM. So, still a little bit finicky yet again, but next week will be much better. But guys, check it out. Give some feedback. I'm still figuring out things for this. I don't want to throw any shade on these people. This is just a use of AI. So, I just want to demonstrate how AI can be used for this. And in fact, the other channels that are here, we told AI to scour these, because these are really good airdrop or DeFi related channels. And we also watch them for our own research purposes. So, we respect them a lot, so we put them here too. So, it's just so much information if you want to have a better information flow. We should do a graph or something. Or a bubble? Yeah, a bubble. Like, everyone's talking about Monad, or like, there's some bubbles appearing around. Oh! And then you, like, like that. Yeah. But I don't know. We'll see. We'll see. But yeah, I mean, that's how we're kind of using AI. And I feel like we're at a point where we're seeing a lot of processes that were old. Like, and I don't even want to show this here. It's like, just like editing our descriptions, you know. Like video descriptions, video titles, using AI for that. And those are simple things. And those cut down the daily, you know, 5 minutes, 20 minutes a day sort of grind. Yeah. And then I feel like the biggest change here is like, for me, I think my biggest objective is to see how AI can be directly applied to what we do. So that way you guys have a job. Oh, yes! And then we can make more money! Let me listen in more closely now. Because otherwise, I mean, otherwise if you just let AI take over, it's bad. So I feel like, I guess in many ways, since I'm not trying to persuade anyone to do something like this, but I told everyone in the office, I mean, like, look, the AI's are creeping up in terms of what they can do. Yeah. Right? It's scary. I mean, especially if you look at the new Studio Ghibli stuff that was really popping this week. I was like, it's scary because like, it's so good. There's so much emotion. Um, I think, let me see where it works. Box money. Unbox money. Uh, well, if, first of all, that's me. This is also me. So, I sent AI this photo. And then it came out with its first impression was this. Looks like a Pringles commercial. Haha! Christ. Uh, I don't have anything to say. I mean, am I that fat, guys? Am I that healthy looking? Yeah. But anyways, um, still got a little bit of work to do, but, you know. Expectation versus reality. Yeah, I think so. What AI thinks, what I think, but, yeah, it's quite scary. So, uh. That Warhammer looks cool though. Yeah, it's actually surprisingly good, you know, like, uh, scary. Scary Marine, guys. So, guys, I think that's kind of our wrap on our take on AI so far of how we're applying it. I mean, there's still a lot of work in progress, but do check out AlphaDraw. Do check out, and also have some pain points. If you guys have anything that, for you guys, is like a pain point in your world, why don't we try to make some apps to solve those problems for you, right? Yeah. And why don't you make your own apps too? But, like, I can do it too. So, um, list some pain points I really want to solve your problems too. Put them in the comment section below. But with that, guys, thank you so much for watching. Shout out! Peace! Shout out! Peace!