AI voice isn’t enough. Natural voice has more bandwidth to encode nuance with tones of voice etc compared to text.
Made a good old fashioned blog post. The previous one was about nostr 😁
IRL meetings (even online): it's expensive to gather everyone together and make them pause everything else when only one person speaks at a time. You cannot speed up or speed-read the conversation to get to the point and save time.
https://www.petrikajander.com/airchat/
#airchat #asynchronous
Async voice chat with profile pics and subtitles is vibes… feels like I get to know people better, and misunderstand them less due to hearing their voice inflections.
I like being able to switch back between this and nostr/Twitter. Sometimes I struggle to explain something simply with writing, so now I can try with subtitled voice chat at my own leisure.
The Path not Taken… 
When one door closes… another opens.🚪 nostr:note1g8rp3r9dh8de8pseyvk7ad2x2tr0yxsd6crajha3gy7dvv9tt02s8vkk32
I just can’t forget it along the way. 💜 The come-up is blinding sometimes.
Gotta start somewhere, yup. But there is huge room for improvement and no one has suggested anything to address this bootstrapping problem.
I’m tempted to release our answer to this Relay Discovery Problem once our merkle tree storage system is ready to deploy.
I was going to wait until we coded everything, but maybe the new merkle trees will be enough to encourage everyone to help us code the rest.
Agreed… Although, the “default relays” set by apps like Damus is a similar problem #nostr suffers from.
BlueSky reminds me of those old Byzantine consensus models that failed before proof-of-work came along. History, but useless nowadays in terms of costly-to-attack trust minimization.
Optimizing the merkle trees. ⌨️🌲
Is Damus glitching or something? This is the only note I’ve ever reposted that shows up on my Posts section. All other renotes go into my Replies section. Spooky…