Have you listened to this yet? He talks a bit about the mental framework they are using and even discusses Julia a bit. The main goal seems to forever be a superset of Python, but to add the necessary syntax to allow performance.

https://fountain.fm/episode/AMio9PPBozCJf5nngmfC

#[1]​ have you run across an explanation to the above question?

These resources look helpful too. Some additional info I didn’t know

nostr:note19eedz3lzaychcuefngwyy7gnj2w5wgza9wk8sjqafep4llnr76ksp5g7vl

In short, it’s mostly word of mouth and trial and error. More relays are good for a robust network, but they are very hard on mobile battery life, data usage, and in my experience performance. I think many clients get hung up when one relay is slow.

There may be some tools that help document relay statistics, but I can’t say I’ve found many that give me a very reliable picture yet.

Most clients allow you to look at other users relays and my understanding has come from following relay devs and occasionally testing a new relay. For example, right now the highest performance relay software is strfry and I mostly try to use relays running on that (I think all 3 I listed are). Two of the 3 relays I mentioned are also default relays for two of the most popular clients, which ensures better coverage.

As a paid alternative, the nostr.wine set of relays does a bunch of behind the scenes aggregating and rebroadcasting on their end before you get the data. It provides a great experience but obviously does have a monthly cost. There may also be a free trial now.

A few relay devs with following and I’m sure there are many others I’m missing: #[6]​ #[7]​ #[8]

You have to wonder if this post by Brian Armstrong was the beginning of the end for Coinbase. It’s a pretty crazy read 7 years later when bitcoin is still dominant and scaling and Coinbase is…

https://medium.com/the-coinbase-blog/what-happened-at-the-satoshi-roundtable-6c11a10d8cdf

This is a combo I will be repeating

Mediocre VR was already pretty fun in the right context. Good VR is going to bring mass adoption (once the price drops)

This sounds kind of harsh, but maybe Amethyst needs a way to view what your follows are reporting. If you follow someone who is reporting aggressively and you don’t like censorship then maybe you should unfollow instead of blaming the client settings.

Can someone explain to me what the capital restraints are on an ARK service provider? Does the ASP need to have 1:1 capital for any funds that enter the system? Do they just need enough capital to cover liquidity for a single block?

I am trying to understand how ASPs may fit alongside large pools of capital from systems like Fedi

Any TestFlight release yet?

Any TestFlight release yet?

There is no punishment in nostr. Relays are not really doing any filtering. It’s all happening client-side. Amethyst will hide a user if 5 of the people you follow have reported them. Other clients don’t do this so you can just use a different client. There is no problem here.

Albanese brand FTW

My god it’s so true.

I would reduce your relay list to ~5-10 and see how things go. I would make sure you have:

- wss://nos.lol

- wss://relay.damus.io

- wss://relay.snort.social

To get broad coverage and then select a few others from people you follow

I don’t think I’m having any relay issues

Mind link nostr:note1592khapk5er7v55gsz3lkn8f0ufr5qzat56cyfz3ggkypcg7nspqps9xm2

The beauty of nostr is you get to choose the interface that aligns with your desired experiences. Many humans putting a ton of their time toward providing great interfaces that you can consume for free. Be ready to move quick and try new stuff out as this is a blossoming protocol.

Don’t be an asshole. Not sure why that even has to be said.

Yeah, another good one that secure Scuttlebutt used was “hops” basically just how many follows away someone can be for them to still show up. I’m surprised I don’t see the option yet in Nos

I personally prefer a small list of very different, but easy to understand filters as opposed to a score just for the sake of transparency. The web of trust score is interesting, but it can’t be black box or there will continue to be user pushback (I would think)

Seems like a simple “moderation” settings menu with these parameters would solve 90% of the problem. Maybe it’s just the one parameter for now, but it’s easy to envision a modular set of moderation parameters that the user can optimize