The first question that comes to mind is if this is a passive device, paired with your phone, then it is just a microphone. Why build a new device for that?
Ask HN: Why hasn't anyone tried building an AI wearable wristband?
11–20 of 20 posts
Re: Ask HN: Why hasn't anyone tried building an AI wearable wristband?
#12So your requirements are:
1. Reliable speech to text from multiple possibly identifiable speakers
2. Long-term knowledge storage and retrieval
Speech to text is a solved problem, AFAIK with a caveat: single speaker. You'd need to train a local AI to identify all these different voices reliably. No easy feat.
Assuming you have done that, you have to feed that data into a vector database to retrieve it when you're talking to the AI. You can't use it to train the AI because it would be too expensive. But then you hit another roadblock: you either have very good querying capabilities for that database so you're able to retrieve what matters and feed into the prompt; or your context window is huge. The latter is expensive.
Some commercial LLM implementations are already implementing some form of learning based on previous chats, so it might be doable from a cost perspective.
I think you can't fit the necessary computing power into a wristband today. It needs to take care of speech to text (again, multiple speakers), uploading all of that to some cloud, and do it for hours and hours non-stop.
Maybe it could just be a smart microphone that uploads a constant stream of audio to the cloud for processing? A privacy nightmares no one is willing to touch most likely. Would you have to ask permission from anyone in the room before you enter with your microphone?
Re: Ask HN: Why hasn't anyone tried building an AI wearable wristband?
#13So your requirements are: 1. Reliable speech to text from multiple possibly identifiable speakers 2. Long-term knowledge storage and retrieval Speech to text is a solved problem, AFAIK with a caveat: single speaker. You'd need to train a local AI to identify all these different voices reliably. No easy feat. Assuming you have done that, you have to feed that data into a vector database to retrieve it when you're talk…
> Speech to text is a solved problem, AFAIK with a caveat: single speaker. You'd need to train a local AI to identify all these different voices reliably. No easy feat.
The OP asks for “a device that passively listens to your conversations”, so even if single speaker is solved perfectly (I wouldn’t know, but have my suspicions, certainly for a device worn on the wrist, which means it can rotate, be covered with a sweater, etc), that isn’t enough.
Re: Ask HN: Why hasn't anyone tried building an AI wearable wristband?
#14How about some smart headphones or ear buds? Or like a pin you can wear on your lapel or chest?
not a fan of the pin form factor, and i think ear buds are cool but i personally would rather have another form factor
How about a choker then (like a throat microphone)? Or perhaps bone conduction headphones?
Re: Ask HN: Why hasn't anyone tried building an AI wearable wristband?
#15So your requirements are: 1. Reliable speech to text from multiple possibly identifiable speakers 2. Long-term knowledge storage and retrieval Speech to text is a solved problem, AFAIK with a caveat: single speaker. You'd need to train a local AI to identify all these different voices reliably. No easy feat. Assuming you have done that, you have to feed that data into a vector database to retrieve it when you're talk…
thanks for the explanation! I'm thinking of creating this just as a project and for me
Re: Ask HN: Why hasn't anyone tried building an AI wearable wristband?
#16So your requirements are: 1. Reliable speech to text from multiple possibly identifiable speakers 2. Long-term knowledge storage and retrieval Speech to text is a solved problem, AFAIK with a caveat: single speaker. You'd need to train a local AI to identify all these different voices reliably. No easy feat. Assuming you have done that, you have to feed that data into a vector database to retrieve it when you're talk…
> Speech to text is a solved problem, AFAIK with a caveat: single speaker. You'd need to train a local AI to identify all these different voices reliably. No easy feat. The OP asks for “a device that passively listens to your conversations” , so even if single speaker is solved perfectly (I wouldn’t know, but have my suspicions, certainly for a device worn on the wrist, which means it can rotate, be covered with a sw…
does that mean that the only viable wearable is smart glasses? you can say the same thing about necklaces for example (they'll be covered).
Re: Ask HN: Why hasn't anyone tried building an AI wearable wristband?
#17How about some smart headphones or ear buds? Or like a pin you can wear on your lapel or chest?
I want it to look like a monkey who sits on my shoulder
Re: Ask HN: Why hasn't anyone tried building an AI wearable wristband?
#18Earlier quoted context omitted.
This line made me think about a future where people have and use these kinds of things. It would mean that I could no longer really engage in conversations with people without vetting them first to know if they're doing this sort of thing. I'm not sure how to navigate such a world.
i genuinely believe its the world we're heading into. privacy is practically dead, and i think if the device has enough utility, people won't give a shit about it listening to you (look at TikTok). plus, with proper marketing, you can show that the recordings aren't being stored.
About recordings being stored -- there is literally no way of knowing if that's happening or not. But in my view, it's not a relevant point anyway. The important data about the recordings is being stored (otherwise why make the recordings?).
I object to the notion that it's OK for companies to spy on my without my consent, and that objection is exactly the same if it's an individual doing the spying rather than a company.
Your prediction may end up coming true. If so, I need to figure out how to exist in a world where I can no longer communicate with people candidly in any setting.
We really are building a dystopia here.
Re: Ask HN: Why hasn't anyone tried building an AI wearable wristband?
#19Earlier quoted context omitted.
i genuinely believe its the world we're heading into. privacy is practically dead, and i think if the device has enough utility, people won't give a shit about it listening to you (look at TikTok). plus, with proper marketing, you can show that the recordings aren't being stored.
About recordings being stored -- there is literally no way of knowing if that's happening or not. But in my view, it's not a relevant point anyway. The important data about the recordings is being stored (otherwise why make the recordings?). I object to the notion that it's OK for companies to spy on my without my consent, and that objection is exactly the same if it's an individual doing the spying rather than a com…
I get you, and you're right, the important part of the recording would be stored (the context). however, I see it being like a summarizer and at least it won't be recorded and fed to you verbatim.
I can't lie, it does sound dystopian, but I am excited for wearable AI. I used to be a smartglasses hater but I see the value of it once we can get it down to normal glasses size + GOOD AI capabilities.
Re: Ask HN: Why hasn't anyone tried building an AI wearable wristband?
#20The first question that comes to mind is if this is a passive device, paired with your phone, then it is just a microphone. Why build a new device for that?
the new device would be to avoid the constraints of Apple. Apple wouldn't allow an app that is always listening to you to run, unless it's always open which isn't user-friendly at all.
there's a reason why all these companies are coming out with their own hardware. Apple severely limits your ability to create these as apps.