Agreed. Plus if you're someone like the NYTimes with significant traffic and a people mostly hitting the same content (who wants to read old news?) then CDNs are great for the most recent stuff that everyone is hitting. (But not for the older archive articles).
But IMHO, most people who use CDNs would do better with properly configured HTTP/2. I'm just shocked that the default config of Apache + PHP is a bad config. That's a bit unforgivable - it's like the CDN providers paid them not to do it right.
And to me hearing a CDN provider say "we just opened a new POP in [insert name here]" is like fingernails on a chalkboard for me. You often actually want fewer POPs. The moment the POP has to go retrieve the content from an origin server or even a regional cache server, things will slow down. Better to just have a few POPs - it ups the chances the file will be in the POP's cache.
Re: CDNs…
I did tests a few years ago and found that HTTP/2 _properly implemented_ is as fast or faster than using a CDN.
The trick is proper implementation. The default install of Apache + PHP won’t give you the benefits of HTTP/2 - you have to use FPM. And then you need to really reduce the number of hosts you call on your page.
Substack CEO Chris Best Doesn’t Realize He’s Just Become The Nazi Bar by Mike Masnick. It’s important to look at this when we think about nostr. Now before you jump to say Mike’s pro-censorship, remember he wrote the very influential piece on social media: Protocols, Not Platforms: A Technological Approach to Free Speech*
Take a read, I’m curious what folks think? How do we create spaces where we can both have free speech and be able to kick the nazis out of the bar?
* https://knightcolumbia.org/content/protocols-not-platforms-a-technological-approach-to-free-speech
Yeah, until you’ve been responsible and seen people abusing what you’ve built you don’t get it. Until you’ve seen people advocating things that harm the community you’re trying to foster (even if they’re part of the community)…
Every one of the categories in NIP-69 comes from experience. Not saying it’s perfect, but it’s a “been there, done that” kinda thing . I can’t see anyone with significant social media experience having an issue with 69.
Content moderation preserves free speech. People have to feel safe to share their thoughts.
They fired the people who understood how things work. (Who uses what, etc.) Are you really all that surprised?
I was looking for how to speed up nos.social with loading images and I thought about proxying the images, so they're smaller and served through a CDN... turns out that Iris already does that!
Using https://imgproxy.iris.to/ and the https://github.com/imgproxy/imgproxy tool.
I didn't see a discussion about it, but i think it's interesting. On one level, it's a privacy issue, you take requests which were being logged at what ever url they find in a nostr post, which leaks a bunch of info... and puts it in a proxy. So it protects privacy but also shifts where your ip address is logged to iris's relay. This actually feels like it is a step forward for privacy, but it is a tradeoff. Getting smaller images served faster is the primary benefit. If nostr is going to compete with centralized apps it needs to feel snappy. CDN's help a lot with that.
Scuttlebutt directly shares media from peers, and it's very slow, often the no peer with the media you want is online when you're viewing it. Many apps, including planetary, don't do a good job at downscaling the images for resolutions needed to view in an app, so the files are bigger than they need to be as well.
https://nostr.build/i/nostr.build_95bb7ab1602652b152795511012747fafcbc040bf0adac220cd833cc5a0ff817.jpeg" class="embedded-image" loading="lazy">
#[2] does proxying? I know #[3] does it. I’m not sure what I think about it. Snort does crazy compression to the point that my pics look like crap. If Iris is doing it, then it’s great - things are nice and crisp. (Need to look at the URLs when I’m not on a phone).
The screenshot is why I think Iris doesn’t do it.
Would be nice to offer to #[4] users. But I’ll probably have to run a client on my domain to make that happen. Would be nice if the clients let you pick the proxy service.
That said, I do worry about the legal implications of being a proxy. It’s possible things like CSAM are (briefly) on your server. BUT clearly with proxy there’s no intent to break the law.
Are you vegan or just stuck eating vegan?
Me: 🐄🐖🐓🐑
I wonder if it was taken down by Apple due to it pretending to be the official Midjourney app.
FYI… NIP-27 says it’s a “draft”. While NIP-08 says it’s “final” (instead of “deprecated”).
Now can you imagine if images were stored on relays and Nostr couldn’t delete? At least now people can delete their image mistakes off whatever media server they used.
I’ve got the “Midjourney” iOS app installed… ❓🤨🤷❓
Question for those who use Midjourney… Do you use it on the web or do you use the iOS app?
On the web it seems to be at https://www.midjourney.com and it has a login that requires Discord. The iOS app however doesn't have a Discord login, and when you go to some of the legal pages they talk about "AI Art" which is presumably their corporate name.
The pricing is different as well - $96/yr for the web, $80/yr for the app.
Those of us in NYC might beg to differ 😁
I get the difference but doesn’t filter read and write to nostr.wine? So isn’t specifying nostr.wine redundant if you’ve subscribed & set up filter?
One more question… If I've specified wss://filter.nostr.wine as a relay, is it necessary to also have wss://nostr.wine? Seems like that would cause double the traffic for you…
I ask because I repeated saw the instruction to set up wss://nostr.wine and nothing saying not to if filter was set up.
Just signed up for Nostr.wine 🍷
I’m getting the sense that even if you run your own relay you’ll need something like filter.nostr.wine to make sure you can communicate easily with the maximum number of people.
I was most interested in filter, but apparently you can’t sign up for that directly.
#[0] - may I suggest additional “buy now” links? It wasn’t easy to figure out how to pay if you’re specifically looking for filter. Oh, and it’s unclear what period of time the 18,888 sats covers. Would be good to add that too!
Y'all remember when Arthur Miller wrote "The Crucible" as an allegory for McCarthyism?
There's a modern Witch Hunt among us: cancel culture.
I've been a Harry Potter fan for about 18 years.
I recall years ago (2007) when JK Rowling claimed that Dumbledore was gay. This information didn't dissuade my enjoyment of the franchise.
Now Rowling is under fire for claiming that sex exists and women are women.
This leaves me wondering who's side is she on? Is she a supporter of the LGBTQ etc. movement? Or is she opening her eyes and seeing that the left has gone too far?
Folks, that's what I'm trying to figure out.
I'm listening to this podcast with the hopes that I can hear her perspective first hand.
https://podcasts.google.com/feed/aHR0cHM6Ly9mZWVkcy5tZWdhcGhvbmUuZm0vUlVOTUVEOTkxOTE2Mjc3OQ?ep=14
Regardless of her stance, I won't be cancelling one of my favorite authors. I'm just trying to understand.
She supports #LGB, not #LGBTQ+. There’s a huge difference.
Just let people define themselves. It’s their business, not yours…
Umm… all they have to do is write a Nostr client to pull from the major relays. They’ll monitor Nostr just like the others.
Unfortunately they went a bit too hard on focusing on a Twitter product and not developer community.
If you could only have one, would it be Bluesky or Nostr?
We've seen that bots can be trained in different ways. I could see trusting a bot to have a voice in how content is moderated, and then have humans to spot check the bots and deal with problems that are escalated. Bots are really the only way to do this at the scale that will be needed.
Personally Nostr gives me hope. If we can come up with a individualistic, culturally-neutral, bottom-up approach to content moderation and show that it can work. It's a framework that could be put into law. Other aspects of the tech world could be brought into the picture as well. For example IETF's Privacy Pass could be used to validate device-based content preferences and pass them onto sites which would allow parents to "moderate" what happens on their kids' phones.
I really love that Nostr exists and we can have these discussions and possibly set an example of how things should be done. It's not something corporate America would ever manage to do properly.
In Utah and Arkansas social media sites with >100,000 users from that state now have to do 3rd party age verification for ALL users from that state.
There are all sorts of questions about how that might or might not apply to Nostr. But it's just one of many things that can trip Nostr up legally.
(For the record I think age verification is a horrible idea on many levels. But the laws exist, so you can't ignore them.)
IMHO, deletes are like same-sex marriage. If you’re not in favor of same-sex marriage, don’t get married to someone of the same sex. If you don’t support deletes, don’t delete anything. Problem solved (x2).
The user deleting their own content is not “censorship”. Relay owners can already delete events by running a SQL delete. So I don’t understand the point here.