When we started down this path to build the app, it was initially to find whether there was an existing solution for a realistic representation of my voice that could be used in the event that mine would no longer be available. I have shared in a separate post about that journey and what I learned. What ElevenLabs has done with their technology is remarkable. To be able to recreate a realistic representation of a voice is a remarkable thing for someone at risk of losing theirs.
Will this work for me when…?
In the first version of Talk to Me, Goose! we sought to leverage this amazing AI technology to give people their voice back in an easy-to-use, human-centric (and, yes, hopefully, fun-to-use) solution that they could carry with them whereever they went. Building that version 1.0 was a tall order, admittedly. We learned a lot, but we set out to build something that was scalable, robust and secure. We built it with the lens trained solely on the user, always with the question in mind – “Will this work for me when…?” But, one of those questions that kept coming to mind for me was “Will this work for me when my hands might not work as well as they do now?”
The Awkward Pause
In thinking about that question, and in returning to The Couple Shift podcast and listening to Jules talk about what its like to communicate as a person living with ALS who uses a speech generation device. What he shared was beautiful and poignant, but it also pointed out the challenges. He talks so eloquently about the need to be patient and to get comfortable sitting in the awkward pause that is the time that it takes him to type what he wants to say into his speech generation device with his eyes. This really sat with me in my mind. Here we were building a text-to-speech application that was going to require typing the text to generate the speech and while we found an amazing technology to render the voice, was there anything we could do to reduce the awkward pause. Could we reduce the need to type?
Welcome Merlin
Inspired by the need to reduce the awkward pause and a desire to mitigate the fatigue from typing, we wondered if we could leverage AI here as well. Thus, was born Merlin, the wizard who works for you to craft messages for you. While Merlin is not going to be perfect, we’re hopeful that he’s going to be helpful.
Right out of the box, we’ve found him to at least be fun. You can give him a topic area, a tone (or mood) and a starter set of words or ideas/needs, and then he can go off and generate a set of suggestions – three options of full text blocks that can be synthesized into speech directly on the iPhone or edited before synthesized on the iPad. Merlin is there to help you, we hope. And, we’re hoping he is helpful. After all, he was a wizard, and he does have a magic wand.
What’s next?
Now that Merlin is there to help, we think we can give him a brain. Right now, he’s bit like a goldfish with a memory of about 3 seconds. He gives you the suggestions that you ask for when you ask for them, but he won’t remember anything about them later.
We think Merlin will be even more helpful if he gets to know a little bit more about you, and if he uses what he knows about you to continue to make the suggestions that he makes even more relevant to who you are. That’s going to take a little bit more work, and a little bit more time. But, we’re confident that with more context, with a better understanding of who you are, that Merlin’s suggestions will only get better and more relevant and more impactful for you. We’re excited about what Merlin can do. I hope that you are as well.