A wake up call for aspiring data analysts

Everyone's talking about AI and everyone is scared, but honestly, most of the conversation is completely missing the point because the thing that is gonna end your data career before it even starts, it's not what you think it is. And I'm gonna show you exactly what that is. And more importantly, I'm gonna show you what to do about it. If you are working a career in a job that you hate and you're trying to break into analytics, this is for you. If you've been studying for months, maybe you got Google certificate, maybe even watch hundreds of hours of YouTube tutorials, but you're still not able to get traction, this is also for you. If you are confused about AI or even scared about the job market, market due to AI and wondering whether it's even worth it to get started, this is for you. But I need you to be ready to hear some things that you might not want to hear because I'm not here to make you feel great. I'm here to get you hired ASAP. Now, if you're a subscriber, welcome back. But if this is your first time here, I'm Kedeisha, and 6 years ago I was delivering pizzas and Amazon packages for about $8 an hour. And today I'm running a community of over 80,000 current and aspiring data professionals, and I've helped thousands of other career changers break into data analytics without going back to school and wasting money on courses that just don't work. And if you're ever interested in working with me to land your first in data analytics, click the link in the description. So let's get into it. All right, so everyone is terrified AI is gonna take their job. You go on LinkedIn and you see every other post that some guru is telling you that prompt engineering is the only skill that you need today, that data analysts are obsolete, or that AI is gonna do everything in 2 years. But I'll tell you what the headlines aren't saying. There was a study by the Remote Labor Index that did something that almost no AI benchmark has ever done. They didn't test AI on controlled sanitized exams. They took the best AI models in the world and tested them on real freelance work, real jobs, real briefs, real deliverables, the kind of work that real humans get paid real money to do on platforms like Upwork. And you know what the result was? The best AI model in the world, Claude Opus, at the time achieved a 3.75% success rate. That means it failed 96% of the time on real paid work. Not 30%, not 50%, 96%. Give the best AI in the world 10 real tasks and it will fail at least 9 of them. Now, why does it fail? Because AI gives you corrupt files, It'll give you incomplete deliverables, and it'll contradict itself within its own reports. It produces work that looks professional and great until you actually read it, and then you realize the executive summary says revenue is up 12%, but the data table shows it's actually 8%, and it doesn't even know why it's doing what it's doing. It just generates output at will. And here's the thing I need you to sit with for a moment: most of you are acting exactly like the AI. Here's what I'm talking about. You're passing the benchmarks, but failing at the real job. You probably have all the certificates, you have all the Coursera badges, you completed all the boot camps, you probably watched hundreds of hours of YouTube tutorials, but the second that you are faced with a blank screen or a messy dataset, you freeze. You're trained to the test, but you're not trained for the real world. You can define what a left join is because you memorized a flashcard, but you have no idea when to actually use it in a real business setting versus, say, an inner join. You can tell me the definition of a window function, but you can't look at a business question and figure out which function to actually apply. You know what that is? That is exactly what AI is doing as well. It's passing the benchmarks by looking up answers, not actually reasoning. And some of you are doing the exact same thing without realizing it. I've been coaching students for years and I see this pattern constantly. Someone comes into my program with all the certificates, all the master's degrees, and when we ask them to open a blank SQL editor and solve a business problem, say live for example, they freeze. And it's not because they're not smart— obviously they are— but because they've been trained to follow along and not Think. The reality is companies don't pay us to pass multiple choice tests. They pay us to deliver complete and accurate trustworthy work a stakeholder can use to make real decisions with real business dollars on the line. AI cannot do that consistently. And if you've been spending all of your time collecting credentials instead of building real skills, you won't be able to either. So let's be real for a second. The certification graveyard is real. I just saw a Reddit thread from hiring managers and senior data engineers, and one manager said, and I quote, You need to tailor your resume to the job posting. Today's AI engines will help employers sift through hundreds of applicants. This removes the irrelevant applicants, and believe me, there are lots of irrelevant applicants. Another senior engineer said that they actively filter out candidates who rely heavily on AI by asking them to deep dive into a personal project, because then the AI crutch becomes obvious instantly. So this year, what I want you to do is close the tutorial, find a messy dataset online, something real, something that has missing values. And word formatting and columns that just don't make obvious sense and build an end-to-end project from scratch. No YouTube tutorial, no AI generating queries for you, just you and the problem. That's really the only way to build the muscle, and that's really the only way to go from someone who can follow along to someone who can actually think. And thinking is really the only way you're going to get hired today. But here's the thing too. Even if you do that, even if you build a project, there's still one place where everything else falls apart and most people don't find out about it until it's too late. All right, next point, the technical screen trap. All right, so let's say you manage to get an interview and let's say your resume gets past the ATS robots and you actually get on a call. There's one place where your AI crutch is gonna snap in half and it's gonna happen on live video where a hiring manager is watching you, the technical assessment. Now you're probably thinking, Kadeisha, I use AI to help me study, what's the big deal? And I hear you, I really do, because there's a version of using AI that is completely fine and there's a version of it that actually makes you better, but that's not what most people are doing. There's a massive difference between using AI as a learning tool and using AI as a replacement for your brain. One makes you sharper and the other one makes you dependent. And the technical screen will expose which one you've actually been doing every single time. I was talking to a student in our program. They were preparing for a phone screen with a real company, a real job on the line, and they came to us wanting to book a mock SQL interview. And in that conversation, they had a realization that stopped them cold. They've been using ChatGPT to write their queries too much, not to check their work that they already tried or to understand a concept that they were generally stuck on, but to generate answers from the start every single time. And when they sat down to practice their mock live interview with us, they couldn't open ChatGPT where you can't copy and paste from a tutorial where a real human is on the other side watching you think. They couldn't do anything. They knew what the code looked like, but they couldn't understand the logic behind it. They couldn't explain when to use a window function or really explain any single line of code they'd written. And that's the moment that it hit them. They had been collecting outputs but not really building true understanding. They had been outsourcing their thinking too often to a machine. And now when a real opportunity was in front of them, they had nothing to stand on. Now here's another contrast. Another student in our program, obviously same program, same tools, same job market, same course that they were taking, recently passed a grueling technical assessment and made it all the way to the final round with the CTO. Not the recruiter, not an HR screen, all the way to the CTO. She was sitting across from the technical decision maker of the company, and she held her own. She passed because she had done the real work. She built her SQL from scratch. She practiced explaining her logic out loud, and she had done the hard things over and over again, even when it was uncomfortable. Her exact words after she passed: Since when did getting a job got this hard? I'll tell you when it got hard. It got hard when the bare minimum died. And the bare minimum died because the market got smarter. Employers had been burned. They have hired hundreds of people who all look great on paper but could not do the job. They hire people who used AI to build their portfolios and their resumes, and they couldn't answer a single follow-up question in the interview. So they tested harder. They ask you to explain your methodology. They ask you to solve problems live, and they ask you to defend your work. And if you built your work through AI too much, you cannot defend anything because it was never really yours to with. So if you're using AI to write all your queries, build all your dashboards, or build everything, you are building a house on quicksand. You're outsourcing your thinking, and when you outsource your thinking, you outsource your value. So from now on, for every single SQL query that you write, I want you to practice explaining it out loud. Not to me, not necessarily to a mentor, but to yourself in the mirror, or to someone who has no idea what SQL is. Do this for every single type of concept in data analytics that you're learning. Pretend you are talking to a non-technical, you know, marketing director who needs to understand what you did and why it matters to the business. If you can't clearly explain why you wrote the query the way you did, or make your dashboard or chart the way you did, and how the results impacts a business decision, you don't know it well enough yet. So go back to the fundamentals, and this time don't let AI do the thinking for you. It's going to be uncomfortable, but you will thank yourself, okay, when you're getting the interview and you're not going to be stuck and you're going to be able to actually pass. All right, quick pause. Now, if you want to land your first job in Data analytics and avoid the lengthy mistakes that I did. Click the link in the pinned comment where you can find the exact steps that my students use to land their job in data analytics without going back to school. All right, now let's get back to it. All right, so the next point: the vibe coding disaster. This overreliance on AI isn't just ruining your job search; it's actually creating a massive problem inside companies right now, and most of the media isn't talking about it. This is a term that was dubbed by Andrej Karpathy. Andrej Karpathy, the former director of AI at Tesla, coined this term in a viral tweet. A while ago, he said, there's a new kind of coding and I call it vibe coding, where you fully give into the vibes and embrace exponentials and forget that the code even exists. This overreliance on AI isn't just ruining your job search, it's actually creating a massive problem inside companies right now. Most of the media isn't even talking about it. The vibe coding epidemic has ruined the desire for most people to actually get good at developing mastery of their skills and their craft anymore. And you think it sounds great, but the idea that people think that they don't don't need to learn Python anymore, or they don't need to learn anything anymore because AI will be able to do it for you. So let's see what's actually happening. When you make AI build your infrastructure, you're building a system that nobody understands. You don't understand, right? You don't know the assumptions that the AI made. You don't know the edge cases that it ignored. You don't even know what happens when the data changes or when the business requirements shift or when the system scales. And then when that system ultimately breaks, and it will break, You're completely helpless. You can't fix it because you never understood it in the first place. So what do you do? You ask AI to fix the mistake that the AI made, and now you're in an endless loop. The AI is debugging its own hallucinated code. Your production system is down, your company is losing money, and you're sitting there copying and pasting error messages in a chat window hoping for a miracle. This is what I mean when I say AI is creating even more problems than it is solving for people who don't have the fundamentals. And what the media isn't saying is that companies who went all in on this vibe coding and AI-generated infrastructure are now sitting on mountains of technical debt. And the engineers who actually have to maintain this code, they have a name for it. I was reading a thread on Reddit where senior engineers were venting about this exact thing. They call it stop code. One engineer wrote, I see a growing amount of junior devs copying and pasting AI code that looks okay, but it's actually Nonsense. Worst part, they don't even understand the code that they're committing. They can't debug, they can't maintain it, they can't extend it. And this isn't just happening at startups or just junior dev problems. This is actually happening in the biggest companies in the world. Microsoft proudly announced that 30% of their code was written by AI, and now engineers are warning that when 30% of a code base is machine generated, tiny error rates compound across the entire system, making it genuinely dangerous in production. So when I say make humans cool again, I really mean make humans cool again. The analysts who are going to win in 2026 and beyond are not the ones who are going to know how to write the best prompts. They're the ones who are going to understand the fundamentals deeply enough to know when to use AI as a tool, not a replacement for their brain. They're the ones who are going to be able to look at AI-generated outputs and say, man, this is wrong, here's why, and this is how we're going to fix it. That's not a skill that you get from a tutorial. That's actually a skill that you get from doing the hard work. From building things from scratch, from making mistakes and figuring out why they happen in the first place, from raising your own bar over and over again, even when it's comfortable. As Kobe Bryant said, my favorite player of all time, the best never get bored with the basics. And if you're going to look for a shortcut every single time, you're gonna find one, but that shortcut will cost you ultimately, and it always does. So remember, always learn the fundamentals deeply, not surface level, not, hey, I watched a video about it. Deeply. I mean the kind of deep where you can explain it to someone else, as I said prior, when you can apply to a problem that you've never seen before, and when you can know to debug things when it breaks, because it always ultimately does. Don't just learn how to generate charts in Tableau and Power BI. Learn the principles of data storytelling— why certain chart types work, how color directs attention, how to remove clutter, and when to make a single insight impossible to miss. Don't just ask AI for a Python prompt. Understand the logic of the code and understand what each line is actually doing and why. Because the day you're in a meeting and a manager ask, how did you build this? You need to have an answer. Be the person who knows how to fix the machine, not the person who just knows how to turn it on. Okay, so one thing that people also are not talking about is that this market is going to snap back, and it's actually already happening. Gartner predicted by 2027, 50% of companies that fired workers to replace them with AI will hire those workers back. Half. That's not a fringe prediction. That's a mainstream forecast from people who actually advise Fortune 500 companies. And we don't have to wait until 2027. To see it. It's already happening. Like I said, recent report called The Great AI Layoff Boomerang found that 55% of companies already regretted their AI-driven layoffs. Look at Klarna, who famously laid off 700 customer service workers to replace them with an AI chatbot. Just a few months later, their CEO admitted that they went too far and acknowledged that AI resulted in lower quality. Who would've thought? And then they hired their humans again. The analysts who are gonna be winning in this new type of job market are gonna be the ones who know how to bridge the gap between raw data and real business decisions. The ones who know how to bring their unique human experience and combine it with the rock-solid data fundamentals and position themselves as a person in the room who actually understands both sides. I call this being an insight architect. You're not just a tool operator, you are a thinker, you are a communicator, you are someone who can walk into a room, look at the data, look at the mess, and tell a business what to do next. That is what's going to get you hired, that's what's going to get you promoted, and that's what's going to make you irreplaceable in this job market. We're living in a moment where conventional wisdom is telling you to take more shortcuts. Let AI do the thinking. Just copy and paste these prompts. Do the bare minimum and hope that's enough. And I want you to challenge that directly every single time. Conventional thinking just says AI makes everything easier, but the truth is AI just makes the easy stuff easier. It still makes the hard stuff harder. It lowers the floor. Anyone can now generate basic SQL code. Anyone can produce a passable chart. The floor has dropped dramatically, but the ceiling has risen. And the bar for what it means to be a truly excellent analyst, someone who can think critically, communicate clearly, and make their own decisions, has never been higher. The judgment, the context, the trust, the storytelling, the communication, all that ability to say, here's what the data means and here's what we should do about it, that's still entirely human. And that's worth more than it has ever been. So here's my challenge to you. Stop chasing easy and stop looking for a shortcut. Stop doing the bare minimum and hoping that's enough. Embrace mastering your craft, commit to constant improvement, and raise your own bar every single day. Charlie Munger famously said, those who keep learning will keep rising in life. It's not just about mastering tools. Tools change every single month. Heck, even AI changes every single day. There's always a new tool. Someone's always copying someone else. It's about mastering thinking nowadays. It's about developing the kind of analytical mind that could pick up any tool approach any data set, any business problem, any ambiguity, and just figure it out. Nothing worth having is easy, and the people who are willing to do the hard work, the people who are willing to build from scratch, to fail, figure it out, raise the bar, and do it all over again— those are the people who are going to win. So always be the person who can think and problem solve and embrace ambiguity. So if everything I said resonated with you, let me know in the comments the most important thing or the best takeaway that you got. If you want to go to the next step and know exactly what separates portfolios that get callbacks and those that get ignored, I have a video that breaks everything down on how to build your next portfolio that actually gets you hired. Click on that video right now, and I'll see you in the next one.