During my undergraduate years, I had heard a lot about Python, but I had never actually used it. At that time, my university primarily relied on MATLAB for computation and simulation, so I didn’t really feel the need to learn Python. However, once I entered the workforce, I quickly discovered the appeal of Python. At my company, there was only one MATLAB license, which meant people had to share it or use it on a common PC—obviously limiting. That’s when I realized that the simple image processing tasks I had been doing in MATLAB could be done just as well, if not better, with Python. And the best part? Python is free!
Beyond its built-in modules, Python offers countless additional libraries available for download from the internet and GitHub. Anyone who has used MATLAB will know that it sells add-on toolboxes—like Signal Processing or Image Processing—almost like game DLCs. By contrast, Python has a far broader ecosystem of libraries, many of which are open-source projects contributed by developers worldwide. Unless you’re using Python code directly for commercial profit, these libraries are free to use.
In my field of computer vision, most real industrial applications are implemented in C or C++. The reasons are clear: these languages are flexible, have abundant references, deliver strong performance, integrate well with vision systems and other hardware, and allow much finer control over memory management compared to Python. For this reason, Python has rarely been used directly in industrial applications—except in very low-level or performance-light scenarios where speed is not critical.
That said, I’ve noticed a major shift in recent years. More and more clients are adopting Python—not for direct deployment in industrial environments, but for automating the setup, optimization, and testing phases before applying solutions in C/C++. And that’s the real charm of Python as a scripting language. While it may not be the best fit for production-level industrial environments, Python is incredibly useful for setup and test stages, where flexibility is key.
For example, Python-based interactive environments such as Jupyter Notebook or Spyder allow you to execute scripts cell by cell. Unlike C/C++, where you need to recompile and debug after every code change, Python lets you edit or add code and simply re-run the specific cell you want. This makes it easy to immediately check variable states and data outputs. In optimization scenarios—say, when you need to iteratively input estimates, extract results, and refine those estimates toward optimal values—Python shines.
Personally, I use Python mainly for image processing tests. Image processing often requires fine-tuning parameters such as kernel sizes or thresholds for different images or environments. There’s no fixed formula for this—you have to adopt a trial-and-error optimization approach, comparing input parameters against the processed results. Using interactive environments like Spyder makes this much more efficient, since you can quickly re-run only the sections you want and monitor variables in real time. It’s an invaluable tool for testing code and handling data, and I highly recommend others to give it a try.
(Anaconda also provides an option to install these tools selectively.)
Vibe coding makes it seem as if anyone can build software now, but beneath that promise lie new constraints: token costs, platform moderation, market concentration, and relentless quality competition. This essay looks at the future of vibe coding from both optimistic and skeptical angles, and argues that what we really need is not a fantasy of easy money but the ability to adapt to the AI era.
An experiment in conversing with talkie, a language model trained only on texts written before 1930. From lunar travel and Korean independence to predictions about 2026, it is a fascinating record where the biases and insights of a century-old worldview collide.
Labor song recommendation channel - An era where anyone can create, yet some are still chosen.
Feb 27, 2026
With AI-composed concentration music overflowing in YouTube work playlists, a sense of 'familiar yet unappealing' fatigue has emerged. As this happens, channels that create song selections and videos manually become even rarer. Even in an era where anyone can create, there is a clear reason why certain content gets chosen.
