5 ms·
I’m working on free open source for non speaking people. In the early user testing stage. If anyone is interested in getting involved please reach out. Demo he
by scosman 3y ago
I’m working on free open source for non speaking people. In the early user testing stage. If anyone is interested in getting involved please reach out.
Demo here: https://twitter.com/scosman/status/1608173613340917763 https://twitter.com/scosman/status/1608173613340917763
Top level ideas:
- LLMs to reduce typing. Expand single words into complete sentences easily, and predict next sentence without typing.
- Voice to text to LLM: listen to ambient conversation, and offer contextual replies with zero typing?
- New UI for text input leveraging the dynamic touchscreen of the iPad. No need to move hands, kinda like modern and user friendly Morse code.
Source: https://github.com/scosman/voicebox https://github.com/scosman/voicebox
- abecedarius 3y agoI was curious about the new UI for text input, but didn't see more about it in the README. Are you familiar with the Dasher design? https://en.wikipedia.org/wiki/Dasher_(software) https://en.wikipedia.org/wiki/Dasher_(software)
- scosman 3y agoI haven’t gotten around to documenting it yet. There’s a subdirectory “Morse” with an early version used in user testing, but nothing self explanatory. Plan is a 2-button (or 2 switch) design like Morse code, but without the learning curve and with auto completion. Hadn’t seen Dasher but checking it out now!
- scosman 3y agoUpdate: Dasher is super neat, but definitely for people with fine motor control. I'm working on the opposite, where people have very limited fine motor control.