Recently, I mean today, I have made the discovery of a lifetime and it's so profoundly changing my perspective on Artificial Intelligence..
I could go through the whole conflict and resolution structure, explaining my ethical concerns for AI and human replacement, but you might miss the point so let's fast forward, what I really really hate more than anything, is to be silenced by dyslexia, "I can't find the words today" that's a real and frequent issue.
Did you know that for my 100 plus posts, all of them where wrote on a phone, and do you know why? Because I type the words and a selection of approximate suggestions fills in the gap, how do I spell that, irrelevant.. although actually, that's not entirely true, words I really want to use can be hard to spell phonetic ... Aly, can you see my issue, just then, dumb technology failed me, spell it how it sounds and ironically, no dice.
So then Grammarly came along and I fell in love, I don't care if this thing is sucking up what I type for espionage (spelt that one) I get my voice back, and that's worth the risk. At work, even for a slack message, I copy the thing and stick it into Grammarly and ultimately look as smart as I intended to sound, (big brain smarty brain me 👏😎)
Soo I cheat, actually as a programmer, it's all about cheating, cheating is about finding a loophole, something to bend you around a problem after all.
What's my problem now? It's live demo's... I have been told often to write a script, but can you imagine, for me at-least this is catch-22 hard, write what you want to say, don't know how to write it, don't know how to say it! And yet I could read a script and yes it would help.
I thought about using dictation to read an idea then feed that back into itself to iterate until I have a punchy iteration, but... Where do you start, then there is my voice! Do I really sound like that?
Does it really matter that my voice is the reader after all, if somebody else read my words to me then I could hear the issues and fix them... But short of hiring someone to help me all day everyday, maybe just maybe I could use AI!
Descript you are my non-free (after several hours) saviour, descript has a pretty great voice to text service which I started using today, I read a script aloud full of awful pauses and issue, then I scrubed through my recording, cutting out the pauses and ummmm sounds (brain buffering), the result was great but still me... I need someone else! Enter the overdub feature, you may know that Descript has some ties with mocking bird AI, an awesome and now vapourware AI which you could train with your own voice, it would then clone what you sound like and scare you to death with the frightening AI future unfolding in your ears, that feature is now called Overdub and it's got some pretrained voices which not only are passable, they breath! I mean you play this to anyone in the room and you may only get a second or two of uncanny valley.
What is a boy to do? Write something, read it out, overdub it with the fantasic preset voice, play it back, run it through Grammarly, play it back again... Eventually your sentence will sound so unlike you, and so much better than you can write, that's the stuff I want to read and hear, stuff that's not mine, except the core idea, the philosophy, that is mine, then I can say, without a shadow of a doubt,
I beat dyslexia with AI and publish it...
I plan to formalize this workflow into software that can help you create amazing tutorials, I could use a hand ✋
Oh by the way, this post is 100 percent me, no AI or smarty big brains used here, hope you find the potential... inspiring.
Top comments (0)