This hits the nail on the head. There's a marked difference between a JSON parser and a real world feature in a product. Real world features are complex because they have opaque dependencies, or ones that are unknown altogether. Creating a good solution requires building a mental model of the actual complex system you're working with, which an LLM can't do. A JSON parser is effectively a book problem with no dependencies.
You are looking at this wrong. Creating a json parser is trivial. The thing is that my one-shot attempt was 10x slower than my final solution.
Creating a parser for this challenge that is 10x more efficient than a simple approach does require deep understanding of what you are doing. It requires optimizing the hot loop (among other things) that 90-95% of software developers wouldn't know how to do. It requires deep understanding of the AVX2 architecture.
Seems more like a "mega" Marshmallow Test. Instead of putting off a snack for 15 minutes they're giving up an entire year of birthday gifts for a reward years into the future.
Even if someone COULD write a great post with AI, I think the author is right in assuming that it's less likely than a handwritten one. People seem to use AI to avoid thinking hard about a topic. Otherwise, the actual writing part wouldn't be so difficult.
This is similar to the common objection for AI-coding that the hard part is done before the actual writing. Code generation was never a significant bottleneck in most cases.
reply