Subscribe
Sign in
General
p/generalShare and discuss tech, products, business, startups, or product recommendations
Start new thread
Atsuki Hasegawa

What kind of jobs will be replaced by LLM?

by•

Numerous products now utilize LLM across various business domains, reshaping the job landscape. I'm curious about which jobs or tasks are most likely to be replaced. Feel free to share your favorite LLM tools, recent LLM news, or any other LLM-related topics to fuel this discussion!

Add a comment

Replies

Best
Alex Nix
Most of the text-based jobs can be replaced in the next 10 years. Also marketing, for example in analysis. And this list can expand every year. I posted an article about this topic, what is all made by AI (we tested our upcoming research engine), you can check it here: https://medium.com/@info_44834/a...
Rob Chapman
Great topic! I worry that LLMs are going to create the tech equivalent of 'the death of the middle class.' Why would a company hire a junior dev when their tasks can be done with AI? Or by their existing senior devs, made ever more efficient with AI tools? ...but what happens when this environment continues over several years and the senior devs start moving on, or retiring (or becoming bartenders, for some reason. there's a bar near me where two of the bartenders used to be developers, which isn't a crazy number in the grand scheme of things, but is still noteworthy).
Atsuki Hasegawa
@rob_chap Certainly, it's a good idea to focus not just on the types of jobs, but also on the levels of those jobs! If junior developer roles disappear, and the path to becoming a senior narrows, there might be another way to view this issue. In other words, since there's no need to do junior-level tasks, one could potentially start working on more advanced tasks earlier to become a senior. The impact of LLMs on the number of senior developers is indeed very intriguing.
Atsuki Hasegawa
First, I consider this from two perspectives. The first is whether the task is high-context or low-context. The second is whether the task's output is text-based or not. In my opinion, LLM excels in low-context tasks where the output is text-based. What do you think? I'd especially appreciate any abstract insights regarding tasks that LLM excels at.