Ask HN: How do you solve the speech to text at the command line problem?

In a world so dependent upon working on your phone, speech to text and command line, how can you possibly utilize all 3 at the same time? How do you get speech to text smart enough to know you need SQL not sequel?