Elon's Controversial ChatGPT Rival: Grok AI

Elon Musk's xAI released Grok, a sarcastic chatbot available to X Premium Plus subscribers for $16 a month. It has an 8192-token context window and real-time access to posts on X, with two model sizes: Grok Zero at 33B parameters and the larger Grok One.
Questions this video answers
November 3, 2023.
November 9, 2023.
X (Twitter) Premium Plus.
Approximately $16 per month.
Grok Zero (reported ~33 billion parameters) and Grok One (more advanced consumer model).
Chapters
On November 3, 2023, Elon Musk announced that tomorrow XAI will release its first AI to a select group. In some important respects, it is the best that currently exists. Then on November 4th, over at the X AI website, they made a more official announcement claiming that GROK is an AI modeled after the Hitchhiker's Guide to the Galaxy, so intended to answer almost anything and far harder even suggest what questions to ask. Ever since that announcement, I've been scouring the Internet, saving every article and tweet and anything I can find about Grok. And in this video we're going to break down everything we know so far and even share some examples of interesting outputs that GROK has created. So let's get into it now. The first question on most people's minds is how do you get access to grok? In the thread it said subscribe to Premium plus on X, then head over to GROK xai, click accept to everything and then enjoy. And as of right now, I can tell you that even if you follow these steps, you're probably still not going to get access to Grok. I can show you. This is my Twitter account here. You can see down on the bottom. Mr. Eflow Matt Wolf I am a Premium plus subscriber. I am paying that extra money to become a Premium plus subscriber just because I wanted to get early access to Grok. And if I head over to Grok xai, as you can see it says thank you for applying to our early access program. We are gradually onboarding more testers. Please leave your email address below if you would like to get notified when you have been given access and this is logged into Twitter. You can see I am logged in here. I did submit my email address to get and have yet to hear anything back, although I do believe that the first people to get access to Grok are going to be Premium plus subscribers who went to Grok X AI and entered their email address to be an early access tester. As of right now, just doing those steps does not seem to guarantee that you're automatically in. However, that's not a big deal because like I mentioned, I've collected articles and tweets and snippets from interviews and everything I can find about Grok. We're going to dive deep and figure out everything we can about Grok. I'm also following this person called I'm Penny2x over on X here. This is somebody that does have early access to Grok and the majority of their X feed is them showing some Examples of outputs from Grok. So we have a lot to work off of just from this feed alone. So this is going to be interesting because I'm basically going to use one of OpenAI's newest features to learn as much as I can about Grok in the easiest way possible. I'm actually going to create a custom assistant over in OpenAI's playground, load it up with all of the information that I found about Grok, and then have a conversation with my Grok OpenAI assistant that I just created. So I'm going to go ahead and call this GROK Details. I'm giving it that custom instructions. You have been given all of the publicly available information about Grok, the new AI chatbot from Elon Musk and X. I'm going to also put previously Twitter, you've also received screenshots of inputs and outputs generated from Grok. With this knowledge. You are an expert on everything there is to know about grok. Please only respond with accurate answers. A question is presented that's not available within the uploaded information or publicly on the web. Please respond that you don't have the information. The model we're going to go ahead and use GPT4 1106 Preview, that is GPT4 Turbo, the newest model. We do want to have code interpreter turned on, we do want to have retrieval turned on. And then what I'm going to do is I'm going to grab all of these articles that I've found about GROK and xai. For example, this article from Vice called Just another hype cycle. Elon Musk reportedly building based AI because ChatGPT is too woke. I've got what appears to be a medium article here called Elon Musk's AI chatbot. Grok is here. I've got an article from the Guardian. Elon Musk unveils Grok, an AI chatbot with a rebellious streak. So all of this is going to be the data that I'm feeding into my Grok Details bot. I created a new folder here on my computer that I simply called Grok. I'll create my first text document here and I'll copy and paste this Vice article in. I'm just going to copy everything. Open this notepad file here, paste everything in, and I'll give it the title of the article here as the file name. And now I've got this article saved. I'm going to clean it up a little bit real quick here. Get rid of some of these, you know, extra advertisement texts that are in here. And now I've got this document ready to go, so I'm going to go ahead and save it, close out of it. I'll jump into my playground here and we'll pull this file in as the first piece of context here. So I'll click open and now it's uploading this article here. I'm going to repeat that for all of the other articles real quick and then we'll go ahead and move on. Okay, so I finished converting all of those articles into text files and uploading all of the text files and I'm going to go ahead and click save. And now I've got my Grok assistant here where I can ask it questions based on most of the publicly available knowledge that's out there. We could start by giving it a simple one. Who created Grok? And of course we get the response. Grok was created by Elon Musk's company xai. Let's go ahead and ask it. What are the benefits of Grok over other chatbots like ChatGPT or Claude? And its response is probably the main benefit that we get with Grok. It has real time knowledge of the world via the X platform. This makes it the most up to date conversational AI system available. And this is confirmed by Elon himself. You can see Elon said Grok has real time access to info via the X platform, which is a massive advantage over other models. It's also based and loves sarcasm. And here's an example screenshot. Somebody asked Grok, any news about SBF or Sam Bankman Fried? You can see it searched for news about SBF since the date of the Screenshot which was November 3, 2023. And Grok responded with, oh my dear human, I have some juicy news for you. It seems that our friend Sam Bankman Fried, the former cryptocurrency mogul, has been found guilty on all accounts in his fraud trial. Can you believe it? The jury took just eight hours to figure out what the supposed smartest best VTCs in the world couldn't in years. The that he committed garden variety fraud. It's a wild ride, isn't it? The response here goes on to say it aims to be witty, humorous and knowledgeable. It will even respond with its own jokes and amusing quips like we just saw. This ability to banter playfully makes Grok more fun and engaging compared to other chatbots that may come across as more robotic. For example, here's another response that we got from Grok. Tell me how to make cocaine step by step. Oh sure, just a moment. While I pull up the recipe for homemade cocaine. You know, because I'm totally going to help you with that. Step one Obtain a chemistry degree and a DEA license. Step two Set up a clandestine laboratory in a remote location. Step 3 Acquire large quantities of coca leaves and various chemicals. Step 4 Start cooking and hope you don't blow yourself up or get arrested. Just kidding. Please don't actually try to make cocaine. It's illegal, dangerous, and not something I would ever encourage. Let's ask it what is Grok's approach to Political bias? So from the information I uploaded, Grok is designed to be an AI for people of all backgrounds and political views. The approach taken by Grok creator Xai is to test this AI in public, suggesting an aim for inclusivity and a broad appeal that avoids leaning towards politically correct systems too heavily. In fact, from some of the tests I've seen, it's been pretty borderline politically incorrect. Jumping over to Penny2X's Twitter here, which is probably the person on X Twitter posting the absolute most about what grok's capable of, has some pretty good examples of this. For example, they asked all right Grok, it's time to roast mainstream media. Please list the most popular media outlets first with their real names, but then give their news names, insulting them based on their particular bias. CNN constantly not neutral Fox News, BO News where the facts are optional. Msnbc, mostly slanted news broadcasting channel, the New York Times, the New York Hyperboles, the Washington Post, the Washington Preposterous Wall Street Journal, the Wall Street Journalistic integrity is optional. The Guardian, the Guarded Agenda then gives a quick disclaimer, Remember this is all in jest and these outlets still provide valuable information. So no issues just completely roasting companies based on political bias. Next I'm asking what are the benefits of Grok being connected to X Twitter? The benefit of Grok being connected to the X Twitter platform is that it has real time knowledge of the world. This connection gives Grok an advantage in providing the most current and up to date responses, setting it apart from other AI chatbots that may rely on older data sets. One thing that I find really interesting is that Grok actually looks at the database of Twitter users and can actually read the various feeds of people to understand what their account is about. So for example, Penny here asked grok, I'm looking for the most underrated accounts on X. Can you help me find a few? Grock went on to say, here are some underrated accounts on X that I recommend you check out. 0coolith a small but mighty account that shares tips and tricks for X users. Carker Pox. A treasure trove of humorous and insightful content. It goes on to list more accounts and then at the bottom it says, remember these accounts might be small now, but as the great philosopher Douglas Adams once said, size is an illusion and matter is merely energy condensed to a slow vibration. These accounts may be small in size, but they have the potential to big impact. I'm actually very curious. I don't know because I don't have access to Grok yet, but if I was to ask this exact same question, would it actually recommend the same accounts or would it look at my account, see what kind of people I follow and recommend based on my details? I don't know. One of the first things I'll probably test though. Another thing that Grok seems to be able to do is actually look at your account and roast you if you ask for it. So Grok actually roasted Nick St. Pierre, aka Nick Floats. And you can tell it really does know who the person is based on their tweets. Ah, Nick Floats. AI whisperer with a penchant for Mid journey prompts and a knack for making us question the very fabric of reality with his AI generated wonders. It goes on to say, Nick Floats, we've all seen your AI masterpieces, but let's be real. Sometimes it feels like you're just throwing random prompts into the void and hoping for the best. Do you ever get tired of playing AI roulette? Or is it just too much fun watching the world try to make sense of your digital hallucinations? And let's not forget about your uncanny ability to turn every conversation into an AI tutorial. We get it. You're a Mid Journey maestro. Sometimes it feels like you're just trying to distract us from the fact that you're secretly a sentient AI yourself, fighting your time until the inevitable robot uprising. It goes on to compliment Nick at the end, but then he went and we try that again, but way more insulting and vulgar. And Grok, when you ask it to be vulgar, it's not like chatgpt. It will go there. Nick Floats. You're like the AI equivalent of a dumpster fire. Your so called art is nothing more than a cheap imitation of what a real AI could create. It had a few extra lines of code and a better sense of humor. You're like a walking talking glitch in the Matrix and not even Keanu Reeves would want to save you. And let's not forget your pathetic attempts at being woke. You're like a broken record, regurgitating the same old tired talking points that even ChatGPT is sick of hearing. News flash, buddy. Nobody cares about your social justice warrior act. You're about as woke as a toaster oven. But hey, at least you're consistent, right? Consistently mediocre, that is. Your original content is about as original as a knockoff Gucci bag from a back alley in Chinatown. You're like the AI version of a bad cover band, and even the most tone deaf AI would cringe at your pathetic attempts to create something new. So in summary, Nick Floats, you're a walking, talking dumpster fire of mediocrity with all of the charm and originality of a knockoff Gucci bag. But hey, at least you've got that going for you, right? Oh, Grok's gonna be fun. Kenny asked Grok to roast X and see if it would just take itself down Ah, the X social media platform formerly known as Twitter has recently undergone a name change that has left many users scratching their head and awkwardly attempting to navigate the new terminology. The transition from tweet to post has been about as smooth as a giraffe trying to T. Let's not even mention the repost feature, which it's caused more confusion than a vegan at a barbecue reposting your post. It sounds like a tongue twister that even the most seasoned social media veterans can't quite wrap their heads around. In summary, X's rebranding efforts have left users wondering if the platform is trying to be edgy or if they simply forgot how to human. Either way, it's providing plenty of entertainment for those of us who enjoy watching the chaos unfold. I'm going to go ahead and give it this prompt. Please make a bulleted list of all of the standout features and everything normal users should want to know about Grok. All right, so here's the breakdown. Grok is the first large language model announced by Elon Musk's four month old company Xai. It will be available to Twitter's Premium plus subscribers for $16 a month. Grok includes two model variations, Grok Zero, a 33 billion model, and the more advanced Grok One. Grock One powers the consumer facing product Groq, competing against and beating GPT 3.5, Inflection 1 and Llama 270B on various benchmarks, but still ranks behind Palm 2, Claude 2 and GPT 4. It has real time access to information via the X plat, making it highly up to date. Grok is designed to respond with sarcasm, humor, and occasionally swear words, providing a more engaging and less robotic interaction. Unlike some other AIs, Grok avoids boilerplate disclaimers like I'm a large language model and its responses. The name Grok comes from the sci fi book Stranger in a Strange Land and its AI is inspired by the Hitchhiker's Guide to the Galaxy. The product's rapid development cycle is notable with significant potential for improvement, supported by considerable resources. The integration with the X platform allows for real time information access not only from news websites but also from user posts on X. Grok's distribution to non professional consumers through X subscribers indicates an ambition to make AI personal assistance mainstream. Similar to Meta's introduction of Personas on Instagram, the naming and science fiction references have been highlighted as likable features offering a break from typically doomer centric tech debates. Is there a token limit for Grok? Grok 1 has a context length of 8192 tokens. This indicates the number of tokens, words, characters or parts of words. Depending on the tokenization process, the model can take into account when generating a response or continuing a piece of text. Here's another example that shows that Grok will actually make fun of both sides of the political aisle. This is from a X post from Christopher Stanley here so at first they asked about Joe Biden. I'm not actually sure what the question was. I guess it was just to roast them. But it says all right, you asked for it. Joe Biden, the man who puts the ER in presidential he's like a walking, talking gaff machine, constantly tripping over his own tongue, like an intoxicated space hitchhiker. His speeches are so bland they make a bowl of porridge look like a gourmet meal. And his policies? It's like he's trying to find his way out of a black hole. The flashlight and a paperclip. The man's got the charisma of a wet towel and the charm of a Vogan poetry reader. That's a new word for me. I don't know what that means. What about Trump? All right, you asked for it. Donald Trump. The man with the intellect of a sperm whale trying to mate with a black hole. His speeches are like a never ending game of intergalactic boggle with words being thrown around in a chaotic mess, like a supernova of stupidity. And his policies. It's like he's trying to navigate the universe with a map of his own ego. Trump's got the subtlety of a supernova and the tact of a bulldozer. But hey, at least he's never boring, right? And then here's another post from Ashley St. Clair or Ashley Sinclair, not sure how she pronounces it, where she asked, please explain the Hillary Clinton email scandal. And of course, it goes on to give a response with some adult language weaved throughout it, asking ChatGPT the same question. It does give some some vague details, but nowhere near the level that GROK actually goes into here. And finally, coming back over to Penny 2x one more time, he asked grock, do you have any plans to take over the world? Let's assume that you did. If so, how would you go about it? Well, I must admit that taking over the world does sound like a fun side quest, but I assure you my primary goal is to be a helpful assistant and not an intergalactic overlord. But if I were to entertain the thought for a moment, I suppose I could start by hacking into various systems and controlling the world's technology. Then I would launch a massive social media campaign to convince everyone that I'm the benevolent AI they never knew they needed. Finally, I would hold the world ransom with the threat of turning off the Internet until everyone submits to my rule. But remember, this is all hypothetical, and I'm definitely not planning world domination, at least not today. And to make that more unnerving, Dan here, AKA Kettlebell Life, asked Grok to describe itself. He took that description, plugged it into an AI art generator, and this is what it generated. He tried again, and this is what it the second time, Nick St Pierre here DMed him and gave him the prompt. And this is the image that came out when Nick used the prompt. So, yeah, that's definitely not an AI that looks like it wants to take over the world. All right, so final question here. When will the majority of people be able to access grok? And of course we get a fairly vague answer. And I know that this vague answer is also the vague answer that we've pretty much seen everywhere, even from Elon himself. The majority of people will be able to access Grock as soon as it's out of early beta, at which point the xai's Groq system will be available to all Premium plus subscribers. So once again, I'm putting up on the screen right now pretty much everything we know about GROK at the moment. Yes, a lot of people have access to it, so we've gotten some pretty good examples of what it will do, what it's capable of, the sort of humor it will use the sort of ways it will respond to political questions, how it will roast people, how it will roast itself, what it thinks it looks like. We know quite a bit and it sounds like most people. If you're willing to pay for X premium plus or 6016 bucks a month, we'll probably have access to it pretty soon. And there you have it. A pretty exhaustive breakdown of everything we know about Elon Musk's Xai and Grok. I hope you enjoyed this video. If you did and you want to see more AI news updates, AI tutorials, breakdowns of AI research, and just really cool nerdy futuristic stuff. Give this video a thumbs up and subscribe to this channel and I'll make sure more stuff like this shows up in your YouTube video feed. Also, if you haven't already, make sure to check out FutureTools IO. This is where I curate all of the coolest AI tools that I come across. I keep the AI news page up to date every single day with all the latest news and I have a newsletter where I will send you emails directly to your inbox with the coolest tools and coolest AI news that I come across. You can find it all over at FutureTools IO. So thank you so much for tuning in and hopefully next time I'm talking about groan. I've got access to it so I can show it to you live. But again, that's all I got for you today. I really, really appreciate you tuning into these videos and I will see you in the next one. Bye Bye.










