The whole 'AI taking programming jobs' reminds me of the '3rd world countries taking programming jobs'.
Don't get me wrong, its great for filling in parameters of functions and APIs, it sometimes good at figuring out errors and solutions. Its bad at coding anything other than generic algorithms, and its typically bad at that too.
That's pretty much what most of these firms do.
Perhaps there's a startup idea here: identify the way communication is done with these teams (PRDs? I have no idea...), and create a service that takes that as input and creates results using LLM? There for sure should be room for fine-tuning models for it, as well as learning to have the LLMs create and run tests for all the code, according to requirements/specs.
Is FAANG laying off employees in Sweden as well?
Do you think most of the Swedish engineers working for FAANG would be replaced by GPT4 soon?
Are you planning of doing it in your startup as well?
Should the employees be worried?
As for FAANG layoffs - I think it is happening all over, but also that it is overreported. The graph of hires over the past few years should be references whenever number of people layed off is mentioned. Over the past 1-2 years I think the tech giants are way up in head count.
So many of these folks fill up seats in code factories or do glorified data entry.
I don't think the world is ready for what's coming.
The level of human interaction required to distill the abstract and implicit requirements of the customers into a machine-implementable format far outstrips the capabilities of current LLM based systems, and I say that as a daily GPT and Copilot user. It will just re-balance the workforce, turning 10x programmers into 50x programmer-project-managers, hence just ever more scarce.
I don't understand the math here: Why would it make programmers more scarce?