Better voice to text. I’ve been acting like a boomer a lot more recently by dictating things into text fields and judging from the accuracy of the LLM models like whisper, I could seriously see myself using voice to text for everything that doesn’t need to be spelled out.