[0:01]Layoffs. Layoffs. Massive job cuts at Amazon. Amazon set to announce massive layoffs. The tech job market has been brutal over the last two years. I was laid off from Big Tech just about two weeks ago. Computer science and engineering have some of the highest rates of unemployment. Layoffs are in the millions. And then there's the elephant in the room, AI. Is it just a tool? Is it just hype? We can argue about it forever, but truthfully, the job market does not care. Companies are hiring differently, interviews are changing, and the people that are actually getting offers aren't the smartest coders anymore. And so in this video, I'll be presenting to you all exactly how I would break into tech in 2026, in this new era of software engineering, even if I were starting from zero. Specifically, I'll break down how AI is quietly reshaping tech behind the scenes and the key skills you must develop around AI to survive and actually win in this new tech era. And with that being said, let's get into it. So the first thing that I want to bring up is this picture right here. This is a picture from the year 1900, New York, 5th Avenue. If you really squint your eyes and ask yourself how many cars do you see on the road? One, right? Pretty much everyone else has horse-drawn carriages. But this picture from the year 1913 features how many horses? Zero. There are no horses here. And this is a change of over 13 years, where we went from pretty much one car being on the road to pretty much no horses being on the road. And actually, if you were to rewind, go back to the year 1900 and ask any of the horse riders or horse sellers, "Hey, are the cars, are they going to be the new big thing? Is that, is that what's going to change?" They probably would have been like, no. Horses, camels, donkeys, human beings have rode those as the central mode of transportation for all of history. Cars at the time were pretty bad. They didn't have the Bugattis, the Lamborghinis, the Teslas at the time. They pretty much have literally whatever you call that. I don't even know what that is. The central core that I'm trying to get to is technology is constantly changing. And we as humans actually have a tendency to underrate a lot of technologies. When the Internet first came out, a lot of brick and mortar stores went away. A lot of those physical stores went away. You guys know Blockbuster? Like where are they now? For those who don't know, Blockbuster was a store where you would have to drive up, find a movie that you like, pick out the movie, drive back home, put it in your CD DVD player, whatever that is, and then play the movie. It was such an arduous process to go back and forth just to watch a movie. Then Netflix came out. They adopted the internet. They adopted the technology. And now we have the ability to watch any movie we want right on our phones. And the only thing constant about the world of technology is we are constantly in change. So no matter where we are, do not underrate any technology because it will come in super fast and change the whole world, as you can see by the time it was 1913. And the parallel that I want to really bring up, this is the impact that AI is having, but at a much faster time scale. You see, this is a chart that shows the differences between GPT 3.5 and 4 in comparison with humans on the certain exams they were taking. So these aren't the exact scores they had, but just in comparison to humans. GREs, the AP courses. The GPT 3.5 was all right. It could score rather decently. But then GPT 4 blew it out of proportion. GPT 5 blew it out of the water. Better than some PhDs nowadays on certain examinations. And so, the biggest changes from this, how I like to lay it out, is first, human beings started as farmers. Then came the technology, the industrial revolution, we got a lot of machinery, a lot of factories, tractors, and a lot of physical labor was changed by the technology. And then the game became, all right, so who's the smartest now? Who can use the technology the best? Who is the most educated? Who can write the best code? Who has done the most degrees? That's why people with degrees typically earn more than people without degrees. People with masters earn more than bachelors. PhD students end up becoming research scientists at some of the top companies earning a lot of dollars. But now, because of AI, we have turned into the world of agency. So no longer is IQ the most valuable thing. The most valuable thing is agency. Agency is your ability to use AI to succeed. You see, AI, you can argue it has a PhD in everything. It knows everything about Computer Science. It knows everything about Geography. It knows everything about History. It knows everything. You name the field, it knows everything about it, because it's been trained on the entire Internet.
[4:31]So no longer am I the most valuable person for knowing something. Your knowledge in a certain domain, that itself, is not the premise of the value. The premise of the value comes in your ability to use that knowledge. And personally, Bob the Builder, I think is the best example of this. You see, Bob the Builder, he knows exactly what he wants to construct, but he doesn't construct it himself. Rather, he has a team of construction cars, tow trucks, whatever those things are to help him actually construct it. So he puts those agents to use, and that's what agency is. So the question that you guys should really ask is AI gonna replace my job? What what do I do with AI? What what's happening? It's more of how can I use AI to reach my goals? How can I deploy AI to work for me? And that's where I come in to the best career that I think everyone who's interested in technology really needs to take heat of, and that is creative context engineering. Earlier we had a question of what should I do? How should I tailor my resumes so that companies in the future would like me? What do I do for interview prep? Like how do I prepare for the worlds to come? And that is creative context engineering. Everyone is using AI tools, and typically how it goes is you ask Chat GPT, "Hey, write me an essay." It spits out your essay. Make it shorter. Make it longer. Add some evidence. Make it sound not AI generated. Make it sound cool. Add in a little bit of humor. Come on, you're so slow. This sounds so dumb. Fix it up, right? So it's it's this constant back and forth, back and forth of prompting the AI, not getting what you want, prompting it over and over again. Context engineering takes it to a whole nother level. And here, context engineering, the central goal is for you to get the AI to give you the exact result on the first try. You want to prompt the AI so well and give so much context that it gets you the exact result you want on the first attempt. And so, I'll give you an example of how this works. So you guys remember back in grade school where you had those like simple five paragraph essays. You had uh intro, a three body paragraph, and a conclusion. Typically right now, like I just described, you ask Chat GPT to write you an essay, you go back and forth, back and forth, back and forth, right? Don't do that. The way context engineers would think is they would first prompt the AI to say, "Write me a five paragraph essay. The first sentence should be on an emotion-based hook. Then we need a synthesis of the entire essay adding in some evidence. Then three pronged thesis statement that cover the three body paragraph. Within each body paragraph, start once again with the hook. Add in evidence from this reading/PDF/website. Then do analysis on it and make the analysis follow X, Y, and Z and then ultimately have a conclusion." So my prompts at this point might be two, three, four, five pages long, but that's perfectly fine. In fact, that is actually better and you become more of an efficient engineer by prompting it effectively. Because right now, everyone has access to the same AI resources. So you are no longer impressive if you are just able to use AI. But it's rather effectively using AI and providing the right context to develop it. And on top of that, the way that this translates into code, so I give you like the writing example, the way this translates into code is you set up file structures. You create those comprehensive readmes of here's how the files are supposed to look. Here's how the architecture is gonna be. This is the testing framework you should use. You should have meet certain requirements. This is how the mockup looks. You provide as much information as possible. Because we have the tools to create websites and apps just like that. So how can you do it as efficiently as possible without getting into a vicious cycle of debugging? On top of that, I added the word creative. Creative context engineering, like holistically, is what will make you the best engineer going forward. And what I mean by this is AI can do a lot. It knows a lot. The one thing it cannot do, and the one thing it really sucks at, is being very creative. That is something so core to us as human beings, we are sometimes very weirdly creative. An example of this is, you guys know chocolate chip cookies, right? I'm sure we've all had one. You guys know how they were created? How the first chocolate chip cookie was created? So it started off, uh back in the day, they used to not have chocolate chip cookies. They used to have chocolate cookies, where they would pretty much create the dough. They would melt chocolate, pour it into the dough, mix it all together, bake it in the oven, and that would be a chocolate cookie. Then one time, someone didn't have enough time to, like, properly melt the chocolate and all that. And what they did is, they broke up the chocolate chunks, threw it on the cookie, and they're like, "Oh, hopefully it will melt in the oven and it will make the chocolate cookie." What ended up happening is the chocolate stayed separated, and boom, we had chocolate chip cookies. So, this creativeness and kind of almost randomness that exists with human beings, it's something that we enjoy. Like everyone enjoys chocolate chip cookies as a result of this innovation is uniquely done with human beings. And actually, this is one of the arguments that I try to make when some people say, "Oh, AI is going to take your jobs. AI is going to replace you." What you'll actually find is companies that use humans and AI in conjunction will always beat companies who only use AI. Because that creative innovativeness that exists, and the ability to architect and prompt effectively, exists with human beings. AI is an amazing PhD, but we limit it at its ability to autocomplete. Like it's it's it's has a lot of knowledge, but it doesn't invent new things. It doesn't add that human level of creativity. If I were to ask Chat GPT, "Create me a business plan to sell lemonade." It might give me a plan, like you gotta buy your lemons, you gotta price it like this, you gotta set it up here. If you were to ask how to set up a lemonade stand, it'll probably give you more or less a very similar instruction. But what actually makes the lemonade stand excellent? What what's the best business model? It's the one where you add your own punch, your own creativity. For example, I see a basketball park on my neighborhood and I sell lemonade there and I incorporate some logos on my cups to appeal to basketball fans. That's the creativeness that exists in this context. And so, what I want you all to start thinking of is stop thinking of yourself as this kind of confined software engineer, but think of yourself kind of as an AI founder. You see, with the ability to create apps and websites and projects at our fingertips, the only difference comes to our ability to creatively context engineer. And now, since everything has been democratized in our abilities, we all can effectively become our own founders. Because truthfully, the difference between a lot of projects and a lot of startups is the fact that there's a paywall. That's what businesses do. Businesses identify a problem, solve it via say technology or non, but like if we're dealing with tech companies, they use technology to solve a problem and then they charge money for it. Now, everyone here in this room has access to do the same exact thing. But it comes to you to creatively context engineer to achieve that result. So, overall, there are five rules to AI. One, always invite AI to the table. Once again, like those horses and the car scenario, you will vastly, quickly, very soon see the impacts of AI rampant across the industry if you don't adopt it. So adopt it in every part of your workflow. And I'm not just talking about coding. You know, there are so many AI tools that can make the mundane tasks of life that so easy. There are agentic workflows that you can set up to have Uber Eats deliver you food within a button. In fact, I recently started using this one tool called Whisperflow. I don't even type on my keyboard anymore. I just press a button and I talk. So there are so many tools at our fingertips that if we don't adopt, we'll once again be the horses compared to the cars. Second, be the human in the loop. Do not forget your creative abilities and how to creatively context engineer. And treat AI like a person, but tell it what person it can be. The most effective uses of AI are the ones where you say, "Hey, act as a product manager and do X, Y, and Z. Act like a marketing director, do X, Y, and Z. Act like this role." Tell it exactly what role you want it, and then it'll adopt it perfectly, because it'll get fine-tuned in that way. And fourth, assume that this is the worst AI that you're dealing with. Because AI is changing extremely rapidly. Do not get romantic with what you're currently doing or what has existed in the past.
[12:25]Because once again, you'll see everything changing very quickly. So assume that you're dealing with the worst model possible at all times. And lastly, do not date your AI. Thank you.
