Who knew that the killer task for LLMs would be "recreate existing things that exist without me mentioning them by name, based on a worse alternative as the example".
You're seriously suggesting that an electron app you vibe coded in two days is more performant than native code that's been developed and improved and tweaked over literal decades?