In 2030, will most human-computer interactions happen through a LLM-interface?

Prediction market on manifold. In 2030, will most humans who interact with computers, primarily do so through a LLM-like assistant interface? Like people say "open browser", and a persistent LLM like thingy opens the browser. Or they say "remove so and so image from my computer", and the llm locates it and deletes it. "LLM" interpreted very loosely. Doesn't have to be based on transformers. Can be highly multimodal etc.

Liquidity: $1,045. Resolves: 12/31/2030.