> I think there is a section of programmer who actually do like the actual typing of letters, numbers and special characters into a computer, and for them, I understand LLMs remove the fun part.
I know you didn't mean to, but I think that description is a mischaracterization. I'd wager most of us "I control the computer" people who enjoy crafting software don't really care for the actual imputation of symbols. That is just the mechanism by which we move code from our heads to the computer. What LLMs destroy – at least for me – is the creation of code in my head and its (more-or-less) faithful replication inside the computer. I don't particularly enjoy the physical act of moving my fingers across a piece of plastic, but I do enjoy the result executing my program on my computer.
If an LLM is placed in the middle, two things happen: first, I'm expressing the _idea_ of my program not to a computer, but to an LLM; and second, the LLM expresses its "interpretation" of that idea to the computer. Both parts destroy joy for me. That's of course not important to anyone but myself and likeminded people, and I don't expect the world to care. But I do also believe that both parts come with a whole host of dangers that make the end result less trustworthy and less maintainable over time.
I'm definitely warming to the role of LLMs as critics though. I also see value in having them write tests – the worst a bad or unmaintainable test will provide is a false error.
I know you didn't mean to, but I think that description is a mischaracterization. I'd wager most of us "I control the computer" people who enjoy crafting software don't really care for the actual imputation of symbols. That is just the mechanism by which we move code from our heads to the computer. What LLMs destroy – at least for me – is the creation of code in my head and its (more-or-less) faithful replication inside the computer. I don't particularly enjoy the physical act of moving my fingers across a piece of plastic, but I do enjoy the result executing my program on my computer.
If an LLM is placed in the middle, two things happen: first, I'm expressing the _idea_ of my program not to a computer, but to an LLM; and second, the LLM expresses its "interpretation" of that idea to the computer. Both parts destroy joy for me. That's of course not important to anyone but myself and likeminded people, and I don't expect the world to care. But I do also believe that both parts come with a whole host of dangers that make the end result less trustworthy and less maintainable over time.
I'm definitely warming to the role of LLMs as critics though. I also see value in having them write tests – the worst a bad or unmaintainable test will provide is a false error.