The article recounts the author's journey into the realm of AI-assisted programming following the launch of OpenAI's ChatGPT, which successfully generated a WordPress plugin. Testing multiple large language models (LLMs), the author found that while some excel at creating code snippets and debugging, many struggle with larger tasks, with only two of 14 models consistently producing working plugins. The exploration emphasizes the cautious approach needed in relying on AI for extensive programming, suggesting that while some free models can suffice, paid versions ensure better reliability for significant projects.
"When I first asked AI to create a WordPress plugin, I was surprised when it worked, leading to my exploration of LLMs for programming."
"Of the 14 LLMs I tested, only two can reliably create working plugins, while five still struggle to deliver any usable code."
"AIs can excel at writing short code snippets and debugging rather than creating full applications or complex programs—caution is advised."
"The free versions of certain chatbots perform adequately for basic tasks, but I wouldn’t trust them for significant programming projects."
Collection
[
|
...
]