AI, Data Annotation, Jobs

rewrite in my voice in paragraph form: I imagined that data annotation would change the world like it changed my world. This is why I wanted to start a small data annotation business and help provide jobs. In this way I could see how the industries surrounding AI, such as data annotation might also help with wealth inequality as it creates employment. I imagined creating a UI so simple and easy to use that anyone could use it on their phone. This would open up the role of data annotator to more and more people. Why would we want data from less tech savvy people? Linguistics. The study of linguistics is limited by the lack of written forms of language which comprises a large portion of spoken language. Through a simple data annotation tool which takes the speech and analyzes it for linguistic research and development, you could help preserve unwritten languages and empower underserved communities. Unrealistic is what this is though. It was just a pipe dream. So instead I am going to try to make something which will work. Instead target higher quality data from trained professionals. I was thinking about helping to improve coding models by presenting difficult coding problems to be solved, or output of LLM generated code being corrected and analyzed in order to improve the models. That would at least give work to out of work software developers. Maybe they would hate AI less. Then all you would need to do is study software development in order to get a job, or at least be good enough at it that you can help train the models. No small feat but it would still maintain the meritocracy that tech jobs allows. That was one reason I started studying SWE was to pull myself out of poverty. It is also my idea of fun, but part of the reason was to improve my situation, I imagine that is a reason for a lot of developers. With all the layoffs happening in tech now people are losing hope. But I think there will be plenty of jobs for people who have a true understanding of SWE. It just might not be as much money as it was before. Yet even with the little I have taught myself, I have been able to pull myself out of poverty and I hope that others might also follow my path. In that way I can hope that AI will help reduce income inequality by offering more employment opportunities. People who say AI will replace all SWE I think are mistaken. Vibe coding is fun and all, and a great way to learn how to do things you could not do before, but at the end of the day, production level code still requires an experienced SWE to solve difficult problems and a well trained one can write the code faster than you can generate and debug the output. At least for now.

I once imagined that data annotation could change the world the way it changed mine. That’s why I wanted to start a small data annotation business—not just to build something useful, but to create jobs. AI and all the industries surrounding it, like data labeling, have the potential to reduce wealth inequality simply by providing employment. I pictured a UI so intuitive that anyone could use it straight from their phone, opening the role of data annotator to people who might never have considered working in tech. But why would we want data from people who aren’t tech-savvy? Linguistics. The study of language is severely limited by the lack of written records for so many spoken dialects. A simple data annotation tool that collects and analyzes speech could be a game changer for linguistic research, preserving unwritten languages and empowering underserved communities. It was a nice idea—maybe even a revolutionary one—but ultimately, it felt unrealistic. Just a pipe dream. So instead, I shifted my focus to something that might actually work: targeting higher-quality data from trained professionals. I started thinking about how to improve coding models by presenting them with difficult problems to solve or by having experts correct and analyze LLM-generated code. At the very least, that could provide work for software developers struggling to find jobs. Maybe it would even make them hate AI a little less. If nothing else, it would mean that learning software development still leads to employment—if not as a traditional engineer, then at least as someone refining the models that will shape the future of coding. That’s why I started studying software engineering in the first place: to pull myself out of poverty. Sure, I enjoy it, but part of the appeal was the potential to change my own situation, and I imagine that’s true for a lot of developers. With all the layoffs happening in tech, people are starting to lose hope. But I believe there will always be jobs for those who truly understand software development—it just might not pay as much as it once did. Even with the little I’ve taught myself, I’ve managed to escape poverty. And if I can do it, others can too. That’s why I still believe AI could help reduce income inequality by creating new kinds of employment. People who claim AI will replace all software engineers are missing the bigger picture. Sure, AI-generated “vibe coding” is fun, and it’s a great way to explore new ideas, but when it comes to production-level code, experienced engineers are still indispensable. A well-trained developer can often write and debug code faster than an AI can generate and fix its own output. At least for now.