A bad developer is still a developer. Even if that developer is largely wrong, the majority of the employed developer's operating time consisting of unproductiv

A bad developer is still a developer. Even if that developer is largely wrong, the majority of the employed developer's operating time consisting of unproductive efforts that will always need to be remedied by expert problem solving and engineering. But the LLM is such a naughty developer that is free (insofar as the pennies worth of electricity and microprocessor metals for running a local LLM is virtually no cost). Not only is it free, it is also able to give you a huge volume of labor. So, an LLM can give you a volume of free labor. And within the probability distribution of that large volume of free labor, there is the chance that bad developer will give you something valuable that can be productive and make up for all the losses you endured dealing with the eager assistant LLM developer. The bottom line is that being a skilled developer will require the employment of artificial intelligence's presently unwieldy enthusiasm and capability. As a sort of necessary evil that has on the opposite side of the bad side, the good side, unexpected insights, massive productivity, and unprecedented development. LLMs are probably the only way to perform the work of a hundred men all by yourself. Unless what you're doing doesn't demand risk taking, perseverance, and a visionary's personality? In which case, we should call a spade a spade and, therefore, recognize that most (if not all) so called senior software developers are nothing more than state welfare recipients who are fungible to companies and economics that can get rid of the drag on their progress if they really wanted to. If you're not learning how to struggle with LLMs, you're probably not improving and growing as a software engineer. You'll then be limited to tentatively useful areas of the Internet. Like working for a FAANG, Tesla, or SpaceX or some other mundane shit.