I posted in the other thread too, but thought I'd answer the specific question asked here. Most of the discussion in this thread has been focused on speech to text. As someone who is blind, I haven't found speech to text very helpful on computers. I can type fast enough on a full-sized keyboard that I'm going to be happier than dictating. I can control the errors I make better that way.
I do use Google Voice on smart phones. While I can type on an on-screen keyboard with accessibility enabled, I find that voice typing and then going back to correct errors works better than my typing speed on those devices.
With regard to text to speech, I think the native accessibility tools in the operating systems are good enough that I wouldn't see a need to look beyond them. On Windows there's Narrator. In my opinion, that's good enough for most tasks starting with Windows 8 that I'd no longer recommend buying JAWS, Window Eyes, or any of the hugely expensive screen readers that we tended to have to get in the past. On MAC there's Voiceover (also on Apple phones and tablets). On Linux there's Orca for graphical environments, and YASR/Speakup/Emacspeak/a few others for text environments. Android phones come with Talkback and Chromebooks come with a version of Chromevox that is better than the one you get on a desktop browser.
All of these are good enough to get work done including reading and writing stories. None of these are perfect.
None of these are great at giving you information about visual layout of what you're producing. I mean you can in general ask if text is bold, capitalized, etc. But understanding table alignment, kerning and multi-column layouts tends to be fairly tricky.
All of the above require some real dedication to learning the tools and understanding how they work. A lot of people who become blind later in life find the frustration of learning a new environment more than they can handle. It all depends on how much you value being able to use a computer, on how bad your vision is, and on what tradeoff you make between doing something the way you're used to slower vs learning something new.
In the past I had some very helpful tapes and classes that helped learn the basic concepts behind screen layout and behind how screen readers present information.
Today I can pick up a new screen reader in minutes or hours, but it all depends on that initial information basis.
Unfortunately, all those classes talk about technologies that no longer exist and I wouldn't know where to find them anyway.
The expensive Windows screen readers (JAWS and window Eyes) probably do have reasonably good training material. They might be worth buying just for that. I've used both over the years, but it's been a long time and I don't remember their training.