July 17, 2025
My favorite use-case for AI is writing logs
One of my favorite AI dev products today is Full Line Code Completion in PyCharm (bundled with the IDE since late 2023). It’s extremely well-thought out, unintrusive, and makes me a more effective developer. Most importantly, it still keeps me mostly in control of my code. I’ve now used it in GoLand as well. I’ve been a happy JetBrains customer for a long time now, and it’s because they ship features like this.
I frequently work with code that involves sequential data processing, computations, and async API calls across multiple services. I also deal with a lot of precise vector operations in PyTorch that shape suffixes don’t always illuminate. So, print statement debugging and writing good logs has been a critical part of my workflows for years.
As Kerningan and Pike say in The Practice of Programming about preferring print to debugging,
…[W]e find stepping through a program less productive than thinking harder and adding output statements and self-checking code at critical places. Clicking over statements takes longer than scanning the output of judiciously-placed displays. It takes less time to decide where to put print statements than to single-step to the critical section of code, even assuming we know where that is.
One thing that is annoying about logging is that f-strings are great but become repetitive to write if you have to write them over and over, particularly if you’re formatting values or accessing elements of data frames, lists, and nested structures, and particularly if you have to scan your codebase to find those variables. Writing good logs is important but also breaks up a debugging flow.
from loguru import logger logger . info ( f 'Adding a log for { your_variable } and { len ( my_list ) } and { df . head ( 0 ) } ' )
The amount of cognitive overhead in this deceptively simple log is several levels deep: you have to first stop to type logger.info (or is it logging.info? I use both loguru and logger depending on the codebase and end up always getting the two confused.) Then, the parentheses, the f-string itself, and then the variables in brackets. Now, was it your_variable or your_variable_with_edits from five lines up? And what’s the syntax for accessing a subset of df.head again?
With full-line-code completion, JetBrains’ model auto-infers the log completion from the surrounding text, with a limit of 384 characters. Inference starts by taking the file extension as input, combined with the filepath, and then the part of the code above the input cursor, so that all of the tokens in the file extension, plus path, plus code above the caret, fit. Everything is combined and sent to the model in the prompt.
... continue reading