The 5 BIGGEST Lies You've Been Told About Claude

The hardest thing about learning AI is knowing who to listen to. So, in this video, I'm going to break down the five biggest lies you've been told about claude, productivity, and building in the AI era. These are by far the most common traps I see people fall into. Now, if you don't know me, I'm Austin. I'm not just a content creator. I ran an engineering team at JP Morgan that replaced Palunteer. I was a COO of a tech startup worth over $25 million. And now I work directly with top AI brands like Anthropic. So, I'm going to go through each lie that you've been told and what you should actually do instead so you can build 10 times faster with Claude. Line number one is you need to stay up todate with the newest AI tools. I speak with business owners almost every day and the biggest problem I hear is I feel like I can't stay up to date with AI tools and I don't know what I don't know. And I'll cover how to fix this problem tactically. But first, should you actually be worried about this? Well, for starters, I can't blame you for thinking this because this is what everyone is telling you. For example, here is a clip of Ali Abdal who has over 6 million YouTube subscribers telling his audience how important it is to do exactly this. >> Is where the conversation is happening about AI. If you're serious about leveling up an AI and you're not on Twitter like looking at it every day, I think that's a mistake personally. >> Now, I'll prefix this with saying Ally is one of my favorite YouTube creators and he's actually the reason I started creating on YouTube. He's my YouTube goat, but I just couldn't disagree with him more here. And I'll provide data for why, as well as from anecdotal experience. So, first, if checking the feed every day worked, the people deepest in the AI space would feel the most on top of it. But it seems like the opposite is actually true. Andrea Karpathy, the guy who essentially taught half the internet how AI works, tweeted that he never felt so far behind as a programmer. So if Karpathy feels behind, then no matter how much time you spend online, I'm sorry, you are going to feel the same way. So this is a losing battle, but feelings aside, does it actually help your productivity staying up to date with the newest trends? So Boston Consulting Group surveyed about 1500 workers as part of their research found that one or two AI tools will generally boost your productivity and at three tools your productivity peaks and then at the fourth it starts to drop. And the reason is that each new tool means new settings, new prompts, and new workflows. And at some point, you just start spending more time managing these tools than actually doing the work. So, you don't have to be on the bleeding edge of every new tool. And in a lot of ways, it can actually be counterproductive. So, what should you do instead? Well, there's a famous story from Mark Zuckerberg when he first started Facebook. He would wear the same outfit every day because it allowed him to save his mental capacity for making more significant decisions at work. He literally thought choosing his outfit was taking away from the bigger decisions. And now this is extreme, but you need to do the same with AI tools. Identify your AI stack, call it one to three tools, and be confident in your decision. For me, I use clawed code, which acts as a central hub, and I plug every other tool into that. And a lot like Zuckerberg, removing the decision of what outfit to wear, you'll remove the decision of what tool should I use or what additional tool should I try out. Then second, when it is time to explore different options, don't AI window shop. instead search for specific solutions. So AI window shopping is the process of scrolling social media and being like, "Oh, that's cool. I want to try that out." And then you go, you try it out, you go down the rabbit hole, and eventually you're like, "Wait, why did I even do that?" So instead of doing this, go looking for a specific solution to a specific problem you have in mind. A simple way to do this is go to YouTube and just search for what you want or use Twitter's advanced search tool to find a specific topic. Generally, I want you to think about the emotion you feel when you use specific platforms. For me, I literally get anxious when I use Twitter. So, I just don't use it unless I'm looking for hypersp specific things. And even on Twitter, I've just blocked people who I know just make these ridiculous clickbait claims because it makes me feel FOMO and like I'm behind because I'm not making a billion dollars with zero effort. And I'll touch on this last part in the fifth lie, which I really am passionate about. But make sure that you're blocking all the noise and you're just focusing on the signal. Once you get past this first lie, I guarantee you're going to start feeling better about yourself. But this next lie is something that just isn't correct. Lie number two is that you can automate your work instantly and be entirely hands-off. We have all seen this on social media, right? Claude code works for me 24/7 so I don't have to work or I automated my entire sales process in 30 minutes. And you can't blame people for using titles like that and tweets like that etc because it works. But there's a famous line from Charlie Mer which says show me the incentive and I'll show you the outcome. So let's look at these incentives, right? Obviously, these social media algorithms reward outlandish claims. So that's the reason for making these. And now I'm not saying that these people are intentionally lying. The reality is that automation specifically is very much on a spectrum. A task getting completed isn't really binary. It's a wide range of quality. But the reality is you won't get clicks by saying, "I automated my entire sales pipeline to respond to leads with very shitty responses." So your feed will show you the claim, but not the quality. Now, for your own sanity, use a framework I learned in 11th grade IB history. It's called OPVL. Origin is who made this. Purpose is why does it exist? Value is what is it useful for? And limitation is what's missing, biased, or conveniently left out. This will help you evaluate the source and take what's valuable and what's noise. And you can use that across anything that you're watching. It's just a valuable thing to keep in mind. Okay. But we know that you can automate some things. So, it's not a,000% a lie. So, where should you actually focus your energy to create an automation that works? It's pretty simple. You have to earn the right to automate. And you earn it with focus. I call this Sauron's eye, which shout out all the Lord of the Rings fans out there. But what you have to do is pick one system and get very specific. Put real energy into making one thing work. If the scope feels too big, reduce it until you have one clear problem that you can solve. One of the reasons that Elon Musk is such a strong entrepreneur is he obsesses over efficiency. And so when reading his biography, and by reading I mean listening to an audiobook, there are story after story of him getting into absurdly small details of a system and then finding a way to improve it. That's Sauron's eye in practice. Pick one system, understand the details, and focus everything there. And while you're doing this, I urge you to use the five Y technique. [music] Essentially, just keep asking why until you land on the root of the problem. Here's something I recently did with a client of mine. They said, "I spend too much time on email, so I asked why. I sift through noise all day to identify where I have to respond. Why? Well, I see a lot of the emails that I don't really need to act on. Why? Well, they all come into the same inbox. Okay, so what appeared to be an email problem was really an email filtering problem, which is much more solvable. Now you know where to start and focus your energy. Before we get to the next section, this whole video I'm asking a lot of hard questions, but there's one simple question that just keeps getting harder to answer. It's a simple one. Who are you? This question was hard enough to answer before AI, but now with AI, it's even harder. So, when Anthropic, the team behind Claw, decided to sponsor this video, I really thought about this question and realized there's actually hope in asking hard questions. You see, if I ask someone, "Who are you?" they'll likely say an accountant, an engineer, a teacher, and we introduce ourselves with a label, the box we spent years earning. But with AI, the question becomes even harder, right? If AI makes it so that anyone can do their taxes, why is an accountant needed? Or if AI can makes it so that anyone can write code, why are engineers needed? AI can call into question our whole identity. But the more I think about it, the more hopeful I become because every one of those labels used to be a boundary. I'm an accountant but not a creative. I'm an engineer but not a designer. The label told you who you were and quietly who you weren't. But today AI has erased that line and now you can be anything. And honestly, this is scary. So I sat down with Claude to try and figure this out for myself. AI is changing everything. Interview me to help me figure out my future. Now, the thing about Claude, and I'll cover this in the next section of this video, is it doesn't just reassure me, it pushes back on my ideas, and it helps me think. It asks me what my skills and what my goals are. And so, I answer it and I tell it where I want to be in 5 years and the current challenges I'm having. And I soon realize that, yeah, I am a software engineer and my industry is changing. But when one door closes, another door opens. I can now start the business I've always wanted. I can learn the skills I always needed or get access to the information that I never had. So to answer the question, who am I? Well, I'm Austin Markeesy and I'm a lot more than a software engineer. So in partner with Anthropic, I want you to think about this question and comment below with what you're hopeful for about AI. Which brings us to line number three. More agents equals more output, which equals more value created. I know that you have seen this person with a screen with like a million agents running. But the truth is that more agents and more output most certainly does not mean more value created. Let's look at a couple of proof points. Starting with Uber, they rolled clawed code out to thousands of engineers and ran internal leaderboards ranking them by usage. And they burned through their entire 2026 AI budget by April. And when the COO was asked whether all that usage was reaching customers, his answer was the link is not there yet. So all of those tokens being used has yet to actually translate to a better customer experience. And we're seeing this trend pop up a ton. There's a growing number of memos from founders saying AI slop is killing the orc. One quote that I love from the co-founder of Hello Patient, he had to call out his team for exactly this. He said, "Faster is not better. More is not better. My rule of thumb, if it takes the recipient longer to consume it than it took for you to create it, it is likely slop." And when I see 10 agents going at once, I see an AI slop factory. I tell people all the time that I work with, anyone can use Claw to produce a report or create an automation. That's trivial. But not everyone can use Claw to produce something that creates value. So before you think that you need to have 10 AI agents running at once, otherwise you're a complete failure, you need to measure the value you create by the impact you have on a business, not necessarily the overall output. So use my post AAI impact equation as a guiding principle. Your impact is equal to a solutions identified time solutions delivered divided by time. This is where solutions identified is your ability to think critically, spot real problems, and create a plan to solve them. Solutions delivered is your ability to take that plan and bring it to the finish line. And the time is how quickly you can get it done. And the key with all this is that solutions is the word here. It's not output oriented, it's solutionoriented. In order to identify solution, it requires deep thought and precision. So to help you orient yourself, use this prompt, which will interview you and use the post AI impact equation to help you identify what you're working on, what real problems there are, and how you can deliver effective solutions instead of just overproducing AI slop. Now, this fourth lie looks way more innocent than the first three. But it's important that you're aware of it. Lie number four. Download these 100 clawed skills and your workflow is transformed. If you've looked online, you've likely seen people who offer hundreds of clawed skills to transform your workflow. But it just doesn't really work like that. Imagine someone just handed you a book on how to start a business and was like, "Yep, there you go. I just transformed your business." Having the book alone changes nothing. You have to take the learnings from that book and apply it to your specific situation. Skills work the exact same way. For a skill to help you, you have to understand how it integrates into your workflow and what the output is actually supposed to look like. When you download a massive conglomerate of skills, you inherit what I call testing debt. When you get 100 skills that you've never run, what does this one do? How does this one work? Does it even work for my business? An unfortunate side effect is once they're installed, Claude could start invoking these skills on its own. So skills that you don't even know how they work, what you thought was a superpower, is now a mystery in your system. and you don't actually have any idea how it works. From Anthropic's own guidance, the most effective skills start as a couple of lines and then are tweaked as the team uses them. From my experience, the skills that really change your workflow are the ones that you really hone in. And the reason in my videos when I share prompts to help you create these skills is that when you run it on your own system, by default, it will take in your context and then you will go through the process of creating it. This process of creating it is a feature, not a bug. Now, I have videos on my channel going into exactly how to use Claude skills and optimize them just like Anthropic. But if you do want to download external skills, which is totally fine, do this before you actually install them. First, if you have a skills folder, upload it to Claude and say analyze these skills, which would be helpful in my system and why bias towards less, not more. This will help you with the curation process. The second is once you actually upload these skills and you're creating this robust skill directory, you have to audit the skills that you currently have. The more skills, the more noise Claude has, and the more tokens and money you waste every prompt you send. So, use this prompt to audit what you have, what you can remove, and if any overlap can be consolidated. Now, before we get to the last lie, which is the one that boils my blood the most, if this is your first video, welcome to the channel. But if this is your second or more, you know the drill. This is our anti-Slap agreement. The visuals, the testing, the hours of research that went into this, this is entirely built for you, not for AI agent scrapers. So, all that I ask is you subscribe as part of this agreement so that I can keep making content like this. Also, every video I give away a Claude Max subscription. Last video's winner is Llama Sunshine01 for building a project management system. For this video, make sure to let me know what you're building so you can enter the next giveaway. Now, getting to line number five. Make $1,000 a day on autopilot using AI. This is just one of the general claims, but you'll see some variation of this on the internet almost everywhere. To understand why this is a lie, let me explain with a story from when I was working with a professional gambler in 2021. This is a wild story and I linked to the original Twitter thread that I posted, but I'm going to keep this brief. I worked with a professional gambler to algorithmically spot arbitrage opportunities across different gambling websites. And what this meant was that we were able to guarantee profit every bet we placed. We were able to do this for 3 or 4 months, and we were essentially just printing money. But then the books caught on, and they kicked us off. At this point, and only after getting kicked off did I decide to write a Twitter thread about how we were doing it. Now, why did I wait? Well, I love everyone watching. I appreciate all the support on this channel, but if you think I'd give a money printing strategy away for free for some views, you are out of your mind, respectfully. And it is the same for every person on the planet. So, bring that back to these broad claims about making money. Do you really think if a concept was so damn easy and so damn effective that someone would be shouting it to the world? Now, are they intentionally lying to people? No. Are the posts likely based in truth? Probably. Maybe. Depends who it's from. But the reality is that it's most certainly is more than a couple of button clicks before you start printing money with AI. So, if you want to make money using AI or help your employer make money with AI, you have to understand how it actually works. First, AI is not a business. AI is a tool and a fulfillment engine. You can't just start an AI business. That essentially means nothing. You need to use AI as a way that helps you provide value. So, if you're a content agency, you would use AI to help you create better content for your clients. Second, when you're thinking about ways to do this, try and focus on where your current domain expertise is so that you can provide real value. Use this prompt, which is built on Nathan Barry's Ladders of Wealth Creation, which is my favorite article on the planet on how to make money. Go through that. you're going to be happy you did. It really makes you think about your skill set. The third is that nothing is truly on autopilot. I have build partner.ai, which has over 2500 signups. And sure, I set it up so that I don't have to do anything dayto-day for people to get value from this. So, the fulfillment side of that business is automated, but the marketing sure as hell isn't. If I stop creating content like this, it would go to zero. So, yes, you can automate a very specific part of your business with all of the caveats I mentioned previously, but not the whole business. Now, this last lie, as well as the previous four, it reminds me a lot of a scene from the Matrix. So, Neo, the main character in the movie, a woman walks by him in a red dress. When it happens, he stops and stares at the woman in the red dress, and this disrupts him. And this woman in the red, this stands for distraction, temptation, and illusion. Across all five of these lies that we covered, which are visible on screen, there's one clear through line. In the AI world, there is a sea of women in red dresses that will distract you from being productive. But if you understand the truth behind the situation, it suddenly gets a lot less sexy. So go back and rewatch whichever lie you find applies most to you. And then for that specific lie, follow the steps that I outlined so that can go from a weakness to a strength. Now, if you like this video, you will love this video where I dive deep into how you are the problem, not Claude, and the five tactical changes you can make to 10X your productivity. I'll see you over there. Peace.