I’m writing this with a tequila bottle in one hand and a five dollar cigar in the other. That’s not totally true, I guess, or even true at all in the technical sense. But I could be writing with booze and smokes in my hands if I wanted, and every writer in literary history would envy me. Today for the first time I’m attempting to use speech to text software to write a real thing that real people might read.
I’m trying this in anticipation of a neurological rebellion that might hold my hands hostage, like socialist guerillas occupying a power plant, but I’m finding it a problematic exercise. For example, in the prior paragraph the speech to text software thought that the word “for” should be “from.” When I tried to edit the word, the software obtusely led me on a Maypole dance through four or five incorrect commands. The most entertaining was when I said “select four words right,” and the software interpreted it as “Open World of Warcraft.” I am not making that up. I haven’t logged on in years, so it was a surprise. Also, it’s really hard to get this program to type the phrase “Open World of Warcraft” when it thinks you want to open the program World of Warcraft whenever you say those words.
The preceding paragraphs took me five minutes to write. They took seven months to edit, otherwise known as 30 minutes, but anyone who challenges the seven months interpretation can put on this god damn headset and try it themselves. It also aggravates me that the stupid software doesn’t understand the word “obtusely.”
A quick experiment has just shown me that this program understands almost no profanity. That is a F you see Kay I in G shame, and I expect that’s going to slow down my words per hour considerably.
Holy frijoles! (I just found out it doesn’t understand Spanish, and I had to type “frijoles.”) I don’t know why, but all on its own this software just tried to take something I said and post it to Twitter. I hope it wasn’t “F you see Kay I in G.” That’s a little bit scary if you ask me.
The biggest problem I’m having is that I’m not verbally oriented. I have a hard time learning things by listening to people, especially if they’re really boring people like most of my college professors. I learn things by doing them. That’s handy when you write by typing on a keyboard with your actual fingers. But in order to speak the words I want to write, I have to stop and think about every phrase before I say it, so that it doesn’t come out sounding like a Neanderthal on Quaaludes. (Holy crap! This program understands what Quaaludes are. I bet that’s because the people who use this program have to take them a lot.) So, for these few paragraphs that would normally take me about half an hour to write and edit, this program has demanded an hour and a half, a liter of Diet Coke, and a surreptitious pull off the Cuervo bottle. (I see it understands Cuervo too.)
Despite all that, I admit that this would be better than not being able to write at all. So I’m going to call this test successful, maybe have a party, and definitely have a celebratory bag of peanut butter M&M’s. Technology is a damn fine thing, but I will say that I never want to have software integrated into my body, no matter what technologists say. It would take me 45 minutes to pee.
Photo from Wikimedia Commons.