Hacker News Reader

Comment by hprotagonist | Hacker News Reader

I have nothing against the major hearing aid companies per se. Their approach to product development is hidebound and stuck in about 1987, though, and they are going to find out in short order that that doesn't fly any more.

Personally, my money is on Apple and Google squabbling over who gets to buy Cochlear, Oticon, and Starkey for their IP and regulatory know-how, and then just mopping up the floor with the rest of the players in the space.

It's already begun, actually: https://www.wired.com/story/how-apple-is-putting-voices-in-u...

Wired fundamentally does not understand how crappy CIs can be, so it's too enthusiastically reported, but I'm pretty sure this is the opening salvo.

View on HN

The way I see it though, the limitations of CI implants are primarily driven by the internal electrode interface that limits the number of available frequencies to the person. This isn't something that any of the "big tech" companies are positioned to innovate in. I'd be more interested if Hewlett Packard got involved because they actually have a significant and successful micro/nanotech group.

Hearing aids though could be ripe for disruption by improved computational power and large supply chain leverage. So it will be a matter of whether or not Apple/Google/ back off once they meet the intractable reality of regulatory compliance in a limited market a la their infrastructure plans.

I think it's most likely that they will buy up those medical companies for expertise, then spin off the medical device manufacturing into a new separate company or otherwise sell it off but keep the key engineers to roll into other random consumer products that are less regulated and more profitable. This is what Google has more or less done with robotics after buying basically every major robotics research company in the US a few years back.

> Wired fundamentally does not understand how crappy CIs can be

Just curious, can you explain what you mean by CIs being crappy, or point to an existing explanation?

Briefly, you get 16 frequency bands if you’re lucky.

A healthy ear has about 3,000.

It’s more complicated than that, but in a nutshell, the fidelity with which CIs can represent the world is very limited. They work for speech, mostly, in quiet. There are simulations of what they sound like on YouTube, somewhat.

"Crappy" is one way of putting it. Those 16 frequencies, for me, are the difference between a life worth living, and a life in crippling, hopeless isolation, ended by suicide at an early age.

Hum... I'm pretty sure my life wouldn't be much different if I could only communicate non-verbally.

just to be clear: i’m not knocking their utility! many normal-hearing people just seem to believe that CIs are bionic ears with perfect performance, though, and that’s manifestly false.

the CI users know are glad to have them, and they are, as you say, a lifeline for many. i’m glad my research ultimately contributes to better ones. we just have a long way to go.

This needs to be more widely known. You'd think from reading marketing material that you would be listening to Mozart and jazz the week after implantation.

The reality is that you have basically no pitch perception. 16 frequency bands across, say, the four octaves that music is interesting? You're lucky to discriminate half-octaves let alone semitones.

Temporal aspects of music come through alright. I suppose you could be a drummer.

You would be surprised. Closer to 100 or 200 bands actually, overlapping. There is a plenty of redundancy. The number cited is about the number of neurons.

I am not sure I agree with you.

Given that you get ~16 electrodes, and some greenwood-like frequency-place map along the basilar membrane, and given that the place specificity of stimulation from those electrodes is significantly larger than the width of an inner hair cell, I'm not really seeing "200 bands".

Reply on news.ycombinator.com