Peter Todd is wrong that Nostr is inherently centralized or that it needs a protocol change to become what it has always purported to be. He is in fact wrong today, because what is written above is not valid for all clients of today, and if we drive in the right direction we can successfully make Peter Todd be more and more wrong as time passes, instead of the contrary.
Discussion
Peter is totally right. Nostr currently relies on domains and on relay servers. Even worse, notes are totally exposed between the client and the first relay.
This was a highlight from fiatjaf's article, not sure if it's displaying correctly.
Any improvements to the nostr protocol will tend to start out as additions, not replacements. The protocol needs expansion and I would like to make upgrades so good they replace old parts, but if the old parts keep running in parallel instead, so be it.
Peter isn't totally right, but he's not as wrong as he could be, so I mainly liked the part at the end about making him more wrong.
You might be interested in my thoughts on redesigning the core of the protocol
nostr:naddr1qvzqqqr4gupzqamkcvk5k8g730e2j6atadp6mxk7z4aaxc7cnwrlkclx79z4tzygqqfkummnw3ez6v3dd3jhs6trdahz6amfwqavne0u