AI Is Not Going to Take Your Job
A lot of people think AI is going to take their jobs.
I’ve been working in machine learning, deep learning, and now the GenAI domain for the past 6 to 8 years. I’ve also been working with LLMs for the last two years, and I can tell you LLMs are not going to take your job.
Here’s why: LLMs are trained on publicly available data, which is not 100% accurate. Because of this, every LLM has at least a 3% hallucination rate, meaning they will never become completely deterministic systems.
No one is going to give an LLM full control to make changes to their codebase or manage systems without human intervention. If that were to happen, the model would generate, say, 100 files at least 3 of them would have issues. No company is willing to risk their final release containing critical bugs for no reason.
Instead, LLMs will be used primarily for rapid prototyping. It’s similar to when WordPress introduced templates before, building a website took time, but with templates, you could have one ready in just a couple of hours. In the same way, LLMs can generate code for you, allowing you to build a proof of concept (PoC) or minimum viable product (MVP) in a week. But when it comes to developing a full-fledged system, you’ll still need to write the core functionality yourself.
LLMs will help generate code for new features, but they won’t build entire projects on their own. This isn’t new before, we copied code snippets from Stack Overflow or other websites. Now, we’re copying them from ChatGPT. The only real difference is speed. What used to take 1 to 2 hours can now be done in 5 to 10 minutes, making development much faster.
Since everyone will be using LLMs, the pace of development will accelerate across the board. The only challenge is that technology will now evolve much faster. In the past, major frameworks changed every 10 to 20 years now, we see shifts in just 5 years. That means developers will have to learn and adapt more quickly than ever before.
Here, ChatGPT will help you learn new frameworks faster. So, there’s no need to worry about the rapid changes in technology.
Talking product sense with Ridhi
9 min AI interview5 questions
Still this is increasing dev productivity. Can the number of job openings decrease due to this?
If producing software becomes cheaper, then more companies will upgrade -> resulting in more jobs
As people can perform many tasks, there will be layoffs in companies with a large workforce. However, since people can handle multiple tasks, many startups will be able to break through, creating new job opportunities. This is how we will see a balancing act. I expect 2025 to be a digestion period.
The whole post itself lists all the reasons why AI will take our jobs, Eg. Something which was done in 1-2 hours will be done in 5-10 mins. Which literally corresponds to 1 person doing 12 people worth of work. 3 out of 100 files error is such an amazing result, have you ever worked with junior software devs? At least 60 out of 100 files have errors, that's now even counting readability, cleanup and anti pattern changes.
No company is willing to give control to AI now, but as time goes by trust will come. The same reason it took 20 years and a pandemic to move to remote work which was possible since the 2000s.
Coding is just one small part of software engineering. A software engineer does a lot more than just writing code. They need to test their code, fix bugs, and make sure everything works well in different environments. They also have to design the system, review code, and ensure the software runs smoothly after deployment. Once a feature is live, they monitor it, check logs, and fix any issues that come up. They work with other teams like testers, designers, and DevOps to make sure the final product is reliable and high-quality. So, being a software engineer is not just about coding, it’s about building, testing, maintaining, and improving software continuously.
The time we save in coding will be used for model development and training. AI is going to be a part of software engineering in the future like big data
W take,
So much fear mongering on X has people confused
People just need to focus on double down on their strengths
Interesting. What’s your view on interview process in the age of proliferating AI? Will grinding on leetcode ever go away?
No, it will not go away. LeetCode is used primarily to assess how logical you are. If you can think logically, you can solve any problem that is the core idea behind LeetCode tests.
It’s completely understandable to feel that way given the rise of AI, but the key to staying ahead is upskilling! AI and automation are reshaping many industries, but they also create opportunities for professionals who can work alongside these technologies.
If you're looking to secure your job and boost your career, I recommend focusing on fields like Data Science, Data Analytics, and Data Engineering.
If you're interested in learning more and connecting with others in these fields, you can join the iSkills WhatsApp group, where you can stay updated on the latest trends and opportunities.
Toh fir kya lega AI? 🤖😂
I agree with your central idea but the way AI is going to take jobs is by replacing 3 human resources with 1. Since LLMs are help improve human speed and efficiency, no is going to pay for so many resources.
You've already answered why AI is taking over SWE industries. Imagine being able to have your project scaffold be presented to you quickly (despite of the fuzziness of the requirements), a lot of time goes into just understanding what all pieces should be assembled to give the desired scaffold/chassis; if AI does this then that's two edged; also if AI can start providing bug fixes by being told where to look and stitch together a good solution, that makes swes even more redundant
