A Gemini Success Story
I have found something Gemini is relatively good for. I asked it, step by step, to explain to me how to setup a Python program to run as a command line tool, and it did it. It took a little prompting to get it to explain wrapper scripts and virtual environments. Its initial answer didn’t fully answer the question, and its second answer required several steps to run the tool when I wanted one. However, with some prompting I got it to answer all my questions and show me what I needed to do. It didn’t write the code for me, but it was like having a more experienced Python developer sitting over my shoulder and answering my questions. Pretty helpful.
The answers weren’t perfect. Some of its steps were out of date:
DEPRECATION: Legacy editable install of move==0.1.0 from file:///Users/elharo/move (setup.py develop) is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to add a pyproject.toml or enable –use-pep517, and use setuptools >= 64. If the resulting installation is not behaving as expected, try using –config-settings editable_mode=compat. Please consult the setuptools documentation for more information. Discussion can be found at https://github.com/pypa/pip/issues/11457
However I can improve on that.
It also told me some things I didn’t need to know or didn’t care about. However the Gemini answer was still clearer, more concise, more complete, and more on point than any of the articles and blog posts about this I found through web search.
What I’m seeing so far is that large language models aren’t very good (yet) at writing code and developing software. Some might be better than Gemini. It’s also possible that if I use Google AI Studio or a more recent model I might get better results. Also possible that they’re much better at web apps or mobile apps than the sorts of programs I write. They’re also very good at homework problems, but that’s uninteresting throwaway work.
However, LLMs are quite good at summarizing existing knowledge from a wealth of websites and pages, many of which are poorly written or only tell a part of the answer.