Over the years her hearing aids have improved in battery life, gotten smaller, and improved in things like filtering background noise. Every new hearing aid would start by sounding distorted and weird but her brain adapts to the new sounds after a couple of weeks.
The single largest life improvement I've seen for her was about 10 year ago when she could finally get a hearing aid with bluetooth. Suddenly she didn't have to decide between the "whistling" (feedback) or not hearing to talk on the phone. We got her an iPad and she stopped watching TV almost entirely and now watches YouTube and Netflix. And she can do her yoga over Skype.
Her biggest challenge, at this point, isn't a hearing aid problem but a bluetooh one. Her iPhone inconstantly switches between her hearing aid and her Volkswagen (pre-CarPlay).
I agree hearing aids are stuck in the past, but I'm also convinced most of the problems are lower hanging fruit than a neuralink :).
It's difficult to evaluate hearing aids because they are proprietary. Reading about the features in a brochure for a hearing aid is like reading the features for a TV set or stereo. Who knows what those trademarked terms really mean when it comes to how the code works?
There are a lot of challenges with putting this in a hearing aid: low latency inference (10ms), small power budget, consistent performance, fault tolerance, etc. Still a fair amount of R&D to be done, but change is coming.
For now it looks like it's going to be a behind-the-ear hearing aid. Maybe in 5-10 years it could fit in your ear canal.
Glad you're excited about it!
I guess ultimately you're limited by your interface to the brain (something maybe neuralink could address?), but even still shouldn't there be some "magic" machine learning stuff going on in this space? Not at all an expert in audio processing, but I would imagine that we could do a decent job distinguishing voices from background noise, automatically thresholding sound in different environments, and possibly modulating sounds into different frequency bands when there are e.g. multiple speakers.
For all I know, stuff like this is already going on.