Your Public Inbox relays is where your followers/readers download replies to your posts.
If you added relays with porn (currently damus, primal, nos.lol) to your public inbox list, all your followers are getting porn.
Your relay list matters.
Discussion
nice. Free porn for everyone. What more could one dream of?
nostr:nprofile1qqsyvrp9u6p0mfur9dfdru3d853tx9mdjuhkphxuxgfwmryja7zsvhqpzamhxue69uhhv6t5daezumn0wd68yvfwvdhk6tcpz9mhxue69uhkummnw3ezuamfdejj7qgswaehxw309ahx7um5wghx6mmd9u2mk7fe When i look at my relay list on amethyst, there are relays for inbox, outbox, public, general, dm, etc. As a non techy, how can i know which relays are best for me, or non spam? And which category of relays? I do see a "recommended relays" section. Does this mean recommended to add to current lists, or to only use these relays, for all categories? Its all confusing to me.
The recommended section are for the general list and comes from your follows relays. If your follows use, we assume it could be good for you as well.
General are all the relays amethyst uses to get posts for you. That list is only for you.
Outbox is where your posts are stored, so other people can download them.
Inbox is where everybody should send replies to your posts and all your followers can download from them to see replies to you.
The other section are less relevant to this discussion.
Ok, thanks for that explanation Vitor.
Having to explain all this will turn off more people from nostr than the gay porn π
True. You can have any freedom if you don't control where your content is.
It's not just those three relays.
So I need to pay the price for what randos post on relays my follows have chosen?
If I follow you, and you choose Damus, then my client should look for *your* posts on Damus - not *any* post on there.
Correct for the write (outbox) relays on nip65. The inbox is to download replies to my posts. So, if I choose damus, the porn reply will post to damus and your client needs to download from damus to see all my post's replies. Which will make you see porn.
You can unfollow people if they are feeding porn to you. It's your choice. Losing followed is a good lesson to them.
This does not make sense to me: "your client needs to download from damus to see all my post's replies. Which will make you see porn." . If my client downloads from Damus YOUR posts, signed by YOUR pubkey, why do I have to see porn?
And I hope you are joking with the unfollow strategy.
The outbox (write) is for your posts.
The inbox (read) is to store all replies, reactions, zaps to your posts.
If people want to see replies, reactions and zaps to your posts, their clients need to download all references to the post from your inbox relay, that were written by anybody.
If the porn is there, ALL your followers will received the porn reply.
Yeah but it's on them to filter for authors they and I follow. Please stop shaming people into micro managing relays. Nostr has to filter first and foremost in the client not the relays.
Sure, relay operators will try to tell you otherwise.
There will never be a future where evaluating and choosing relays won't be absolutely necessary. There is no freedom if you don't know and control where your content is. If your content is in the same relay as millions of porn posts, your content inst any better than spam. Stop thinking that relays can be hidden. They cannot be.
Can you not bootstrap relays by using relays of people you follow?
And if the user blocks authors, identify from which relays they come and then ask the user: relay X has a lot of spam. Do you want to remove it?
That's what the recommended relay section in the relay list page does. But your follows are still using porn relays, which gets recommended to you. Unfortunately, users can't delegate finding good relays for themselves. This notion of using whatever relay the app recommends makes you a slave of the app itself. Not much different than the old Twitter.
And purging relays after you blocked enough users from them?
Why not? I really don't want to see posts by people that are completely outside of my WoT and the client should compute that, not the relay. Relays are a substitute for the "ether" we would ideally be using to broadcast messages in a censorship resistant way. Yes, CSAM is what 99.9% of all users want to avoid in their stream but what if you are into gay porn? Why should the relay decide to delete it *based on the content*. Sure, the current spam attack is posting porn out of context but don't tell me they allow them as top posts or reply to gay people posting gay stuff. And the yoda bot. Yes, you can detect and block it but you probably get false positives as soon as people quote the original post when commenting on it etc. and then people get shadowbanned on our censorship resistant nostr.
You are a client dev. Please don't shy away from fixing this on the client. Calculating a follows score is relatively trivial and to hide media based on this, too. What else is missing? If posts get into my WoT I only need some way to cut off the branch at whoever is open to follow spammers or score based on follower count among my follows. If Alice follows the spammer Eve but nobody else follows Eve, Eve should have a lower score than Bob who all my follows are following.
We don't need a contest for the best censors among the relays, so all agree on that to fight spam.
I am not shying away from fixing. WoT doesn't solve everything as it takes days and sometimes weeks for new users to show up in people's radars even on clients that helped onboard such new users. Imagine onboarding your friend and then trying to find the friend and your client blocking everything.
Regardless of the solution in the client, three is NEVER going to be a future where picking the relay isn't the most important thing users do. There is no solution, no alternative to it. You can pick your relay or be a slave of those who do for you. You cannot have both. Ever.
"never" is a strong word, especially when all-caps π
I bet in the very near future, people will rather have their personal AI assistant on their phone judge what is spam than have some central authority preemptively delete events.
Even if you use an AI to filter spam, the central authorities will still delete events. There is no nostr without relays. They are central authorities between you and your friends. They have all the power to shape your reality in anyway they want. Personalized to you.
If you are not careful of which relay you choose, even today, you don't see a lot of criticism to certain projects, for instance. That's why your relay choice matters. Because they can censor you regardless of what the client does.
Yes, absolutely.
I want a world where people have to choose *enough* relays to receive the "broadcasts" while still being manageable for their limited client. Adding more will gradually degrade their experience.
You encourage to pick a minimal set of "good" relays. Adding one wrong relay will occasionally totally ruin their experience and the experience of their followers.
>If people want to see replies, reactions and zaps to your posts, their clients need to download all references to the post from your inbox relay, that were written by anybody.
Curious why can't replies from untrusted people go behind a placeholder? Or at least have that option?
They can. We are building that option on Amethyst outbox. Still doesn't solve much though. And users will need to be very aware that new users are not showing up for weeks, maybe months to them.
Makes sense.
Thanks Vitor. Nostr is more broken than I thought.
Ecash stamps across the board. No stamp, no delivery. Keychat is from the future and was sent back in time to battle Ultron.
Do you mean attaching a payment to each note?
Yup, P2PK-locked cashu string in every individual event. And a network-wide minimum of 1 Sat at core NIP level.
So for 1000 sats I can gay-porn bomb 1000 replies, while making everything more complicated for everybody? I think money can't really solve this problem
Totally can.
1 sat is the minimum. If you use 5 sat stamp relays then it'll be 5000 sats.
(And actually 1000 sats is a pretty high price to bomb 1000 replies.)
E-cash should have been baked into Nostr at the start.
I'd even go so far as to say E-cash should have been baked into the entire internet at the start, would have saved us all all of headache, and the concept goes back to the 1980s.
Gayporn spam actors have orders of magnitudes more money/computation to spend than I do.
Publishing conditions are figured out at the local community level.
Not on a (non-existent) global protocol level.
Anything can be at the protocol level if agreed. That's what a protocol is, an agreed set of instructions.
Sure.
But you put something like the 10 commandments on that level.
Not the minimum price of beer.
What are these 10 commandments that are more important than a no-fuss financially self-sustaining relay layer across the board?
Why do you want to impose that a relay should be self sustaining?
You prevent companies from subsidizing relays, communities from running theirs for free, people who think ecash is crap from participating, etc.
Plus, it assumes a fixed set of relays for each publication.
And what about the (more expensive) blossom media?
Etc...
Stamps are great for letters.
They suck for :90percent: of other use cases.
Like how are you gonna stamp an :app: App Release?
Stamp an app release with 10,000 sats, scam users for 100,000 , sounds like a sustainable business
What can I say, financial health matters. Clearly where things are at now is not in a place of financial health.
You need a fixed reimbursement mechanism at the end of the day, for the lowest level infra only, but still there. Black and white. No confusion, no blind trust in the miracle of value for value. Companies and communities can subsidise on top of that, they can throw around relay-locked sats like confetti, they can invest in other things.
If not you get a situation where a few players invest in better rails and then capture the bulk of the revenue. This more or less always happens.
I don't think we can determine what is a high or low price. If I wanted to bomb you with porn or propaganda of any kind even for lulz I'm happy to put down $10
Ecash of what mint? What if the mint is down?
This is the challenge architecting in money after the fact. The thing is, money will always find a way in, and if you don't architect it in at the protocol level then it'll flow in elsewhere, as it pleases, via advertising, consolidation. It's fluid dynamics.
Look at Primal, almost no spam. Is it a coincidence that probably 80% (my estimate) of all revenue on Nostr (as in funds that could be declared to the tax agent as income) is gong to Primal? It's not.
Would be interesting to know where you pull those estimates from
Let's take users with a single daily driver app as our monetisation base. From numbers shared by Primal cross referenced with Nostr.band stats I'd say about half of these users are Primal users.
Primal has a proper subscription offering, which is fair, they offer what no other client offers in terms of UX, wallet integration, spam mitigation, backups, etc. That cache relay of theirs opens up business opportunities, it just does. So their subscription offers value. And you can pay via Apple Pay, which always helps sales.
Of those users not using Primal as their daily driver, most use an app that doesn't have a subscription offering, or has one that's not as compelling as Primal's on aggregate, and that is framed more along lines of support, for the cause, help out the team and such (versus tangible UX things).
Let's say Primal converts its daily-driver users to income at a rate 3 to 4 times the other apps. (I'd not be surprised if it was even higher, but anyway). That's your 80%. (There are also the paid relays, but even if you put that revenue all to the non-primal side it's not really going to move the needle.)
Not saying Primal is in the black, far from it I'm sure. Just a comment on market share. It is what it is.
Your estimate of 80% is way too high. Revenue-wise, Alby alone is probably bigger than Primal at this moment. Fountain, Wavelake, Satellite, nostr.build, nostr1.com, nostr.land, nostr.wine, jellyfish are also taking good chunks.
Grant recipients take home :90percent: of the economy here.
Correct. But he mentioned revenue. So...
Alby had a 5m round (.. two years ago?), so there is some VCs in the game as well.
Grant = private on chain zap = revenue.
VC is a different thing imo. Though still money that enters the economy, true.
Still pales in comparison to the millions going to grants, organisatiins like &otherstuff, etc...
Yeah, grants are also revenue. And it's revenue without users, so it tends to have no positive effect on expanding the overall market. Only business people wanting to actively earn income can get Nostr really going, IMO.
unfortunately the grant funded devs are a dead horse that isn't feeling the flogging.
Yeah, that's how centrally-planned economies always are; no response to market signals. You don't have to care about the market, when you're sitting on a literal pile of Bitcoin and can just subsist off of that, forever.
it even extends to VCs... because they have a big pile of money as well, the project doesn't have to have revenue for a long time. that's why there's no functional diference between damus and primal. both are not serving paying customers, but "investors"
Although to be fair if you're investing in Nostr at this stage you need a serious altruistic streak of one sort or another. Private equity running Toys "R" Us into the ground this is not.
The road to buggy clients and gay porn spam is paved with good intentions... and awe-dropping amounts of donor and investor money being wasted.
They're only wasted if the intentions are indeed good π
Bitcoin washing π
Deception, scams and blackmail is basically :90percent: of what the ΓΌber rich actors on the worldstage spend money on.
So if some of them come here, it'd be weird to expect something different.
Should've know that π€¦ββοΈ
I'm thinking it's all an AI play.
#plai
ai fools about 95% of the population, just like the fallacy of explosive nuclear fission
ok, nuclear fission i just always assumed that "the scientists know what they are talking about" but then i was prompted after reading Death Object to realise that the mechanics of it are infeasible
It's the protocol also I think, protocols have minds and intentions of their own. We currently have this incentive context wherein those who earn money from operating relays can earn more money when those who run free relays get spammed. And also where free relays will have to pay out to various API providers to deal with spam (visual stuff is LLM intensive), which means an urgent need for monetisation. And what do free relays have to sell besides user data and user attention? Or if the thing to sell is something in a platform paired with that free relay then the relay becomes a loss leader, which is its own dynamic.
I've spend a lot of time in both Hong Kong and Singapore. Hong Kong got the local housing incentives all wrong, while Singapore got them all right. That changes everything.
Well, maybe, but free Nostr relays hardly have to deal with spam. The rest of the Internet is much much spammier.
Our free relays aren't loss-leaders, as the relays themselves offer premium features or are part of a set with paid relays, so some people pay and that finances them for other people. We regard our relays as part of our product and our clients as the ideal viewer for our relays.
Yeah, it's such a weird space, here.
Completely defies all free market dynamics because it's aflood with charity. Charities are, by definition, not meant to solve the problems they address. They exist to perpetuate themselves by alleviating problems.
These aren't real investors. Real investors want a return on investment, and efficient development raises their returns. Why would an investor pay β¬500k for something that would be built better for β¬5k? He wouldn't, but they are.
The only real investors we have, on Nostr, are the individual people and companies zapping, crowdfunding, and making contracts with suppliers. We're increasingly off in our own economic ecosystem.
yeah, gitcitadel is doing stuff mainly for academic and git-based software collab, i want to focus on discussions, chat and calendaring
Yeah, we aren't really doing anything with that.
This sounds like you never had investors before. This "real investor" you talk about doesn't exist. It never existed. It will never exist. It's wishful thinking. A purely theoretical view of how things should be run. Which is ok for the books, and that's it. Its like saying that electricity is made by electrons flowing on a wire. Its great for high school science BS, but misses what actually happens, why it works or exists in the way it does. You will never change how investments (or grants) work, you can only learn to surf the wave for the benefit of your customers. If you think Nostr is unique in that regard, you are just suffering an extreme case of "the grass is always greener in the other side of the fence". This market mismatch that you describe that we have on nostr is not anything especial or unique to us. Or bigger here than in other places/fields. It is as common as shells on a beach.
It's common... in America. Particularly the West Coast.
I'm used to European investors, particularly German ones. They are much more attentive and critical because they are poorer and worry about making a loss.
American investors throw gigantic gobs of money at 10 projects. If one turns out, they break even. If two turn out, they turn a big profit. That works because they have such enormous amounts of money, and it means that they don't have to care what the other 8 do.
yeah, i'd prefer to work with investors with small cap and reasonable but attentive input on monitoring and having some input into the operations, but not a dominant role, of course. CEO is the one with veto power, and the investor has to trust them, primarily, and the CEO has to have a nose for bullshit in his contractors/employees
Yeah, attentive, reasonable investors are a blessing, for a business.
I am just going to live this here
https://klementoninvesting.substack.com/p/germans-are-bad-at-investing
Yes, and we're too poor to invest, anyway. Even the investment funds here are surpringly tiny, unless they are government-backed. We're more savers than investors.
But, people like to invest in *us*.
Everyone knows we got no game and are honest and hardworking, to a fault. That's one reason they like to hire us. We get fronted a lot of trust because we have everyone's favorite flavor of autism:
Can't lie, hyperfocuses, likes machines, good at maths.
indeed, that's why i'm going to go independent, build out the back end infrastructure and then save my funds to build out a decent proof of concept front end web app. or maybe i'll figure out a better strategy involving flutter.
this thing about not being able to make one UI codebase for mobile and desktop is bullshit in my opinion. i created a responsive layout for a PoW crypto web app that was acceptable and even not only between mobile dimensions (very small mobile like 800x480 display) it also scaled up to 4k. so idk what these people are talking about lol. lazy, is what i'd call it.
#Flutter you said :eyes: ?
yeah, flutter is the easiest for covering desktop and mobile, as far as i can tell. i mean, a lot faster than react native anyhow.
i've not got into UI stuff that much because there is basically no real options for keeping the logic in Go
Agreed. Is why I use it.
And Dart is ok-ish for what we do with it.
just curious, can it output a javascript/typescript/html type front end or?
ok, if you can build one codebase with reactivity to cope with small mobile screen up to 4k, i can sketch out the general spec of how the design will be on the 3 different sizes. i have already done a lot of work with this.
desktop apps that can shrink to mobile widths especially are handy when you have a dynamic tiling window manager, lets you stack little displays for apps that you want to watch but don't want big detail unless you want big detail
a great example is chat, btw. having a chat window in a side panel is so neat for collab, with a 4k display and sharing quarter sections of the rest of the screen.
there are so few really good UI designs in production unfortunately. i hate windows but the windows 8 paning system is so intuitive
Using Niel's stuff is more efficient than hiring a frontend developer. At most, you'd need to pay him to add some new components and help you with smoothing out the implementation.
well maybe i'll hire him to do the design and find a front end dev who can build in the necessary front end logic under it. it's a ways in the future i am still just building the shiny database at the moment
He is a frontend dev. All good designers are.
We're putting the infra together so you'll be able to "vibe code" it. Out in two weeksβ’οΈ
i suspect that dart is too technical for dumbass LLMs
i can't see a better option, i'd rather not be stuck in web app land and react native is a joke.
if i had a year all funded to just work on a GUI library i'd make the smokinest golang GUI library that ever did exist. i made a pretty decent one in 6 months back in 2019.
Why suspect, did you not try? Pretty much the opposite of what you said is true
:checkmarkgreen:
you are not speaking from experience.
i came up programming in BASIC at 9 years old, 40 years ago. i have already tested several basic mathematics and logic cases with LLMs and they don't get them right. i repeatedly had to tell it "this has bugs" "oh yes this has bugs" yeah, so why did you FUCKING give me this garbage.
in that time i could have written 20% of it and in the time it takes me to find and fix the bugs, i could have written the rest.
whatever. kurzweil singularity sucker. you won't learn today but you will learn.
hype destroys the capacity to reason. i don't buy into hype, nor do i use hyped up shitty languages like javascript or rust or c++
You are right, I also got frustrated. These tools are not perfect
they are good search engines, but they don't reason.
i mainly use it now exactly for simple searches and simple, well known algorithms that would take me longer to write because i hadn't written one before. and once i have that algorithm i can just copy it and reuse it easily for other things.
The confusion arises in the massive gap between
What American investors like
(flashy stuff, built for influencers, tiktokers, and advertisers, massive audience you can capture to mine their data for your AI)
and what PAYING users like
(elegant, efficient, reliable stuff, that does something useful and makes sense running in a LAN, and good customer support)
There is simply much more promise of fast π€, in the first one, but the second brings in steady income and is harder to dislodge with a competitor.
yup. modesty is better than audacity. when you have a way to save your wealth.
just look at felis catus sylvestris. the image of the perfect balance of modesty and audacity in a fluffy little bundle.
There's a reason why #Einundzwanzig is so big, in Germany. And #Nostr. They appeal to the modest.
i was in their telegram for a while ways back, but i liked "dvadeset jedan" group better, the serbian bitcoiners
Also #PurpleKonnektiv started here. Anything more modest than meeting up for tea π«?
>Only business people wanting to actively earn income can get Nostr really going, IMO.
π¨ nailed it
Alby, sure, but they're part of a lightning payments ecosystem that stretches quite a ways beyond Nostr, so I'm not sure how you'd classify their revenue. Nostr.wine, Nostr.build, etc. I'm skeptical their revenue is anywhere near Primal's, but without more to go on very hard to know.
How much do you think Primal is making a month?
Primal and Alby is likely taking up at least 50%. nostr.wine and nostr.land are very close, nostr.build is somewhat lower (their free uploads are propped up by nostr:npub10pensatlcfwktnvjjw2dtem38n6rvw8g6fv73h84cuacxn4c28eqyfn34f), Satellite/Nostr1.com would benext, then Jellyfish
Which puts Primal at about 25%? That would be my guess. Somewhere between 15% and 25% by just talking to people and seeing what they are using. I don't actually have any numbers.
I think Alby is fair because the main driving force of Primal is the wallet. The client itself is on par with many other options out there. So, it is wallet vs wallet.
Why do you think we all keep saying that we are surprised that nostr works at all?
If I use an aggregator relay (that does spam filtering) the client should never connect to other users inboxes. Could that be forced via a toggle?
Outbox ngmi.
On nostr, nobody takes responsibility for what others talk to them and only mildly endorse those they follow. Relays are just the broadcast substitute because we have no always-on-and-not-censoring-relay we all can just agree on.
If we give credit to your concept of "Alice picked this inbox. Alice endorsed messages *to her* on this relay" then maybe we should allow for Alice to actually see whatever she gets there but also consider "Alice follows Bob but not Carol" so maybe we should show potential spam videos from Bob bot not from Carol client side? And for the more concerned, only show own follows' content?
I am absolutely convinced that we don't need new masters. Don't make relay operators the arbiters of what is spam and what not. Let them mark stuff as spam as a service or what not but ultimately the clients have to get resilience independent of chosen relays.
As a user, I focus on the client. It is the client who displays the content and who can filter and give me tools to customize what I want to see. I don't see the relay as a content moderator.
Yes and that is how it should be. All other approaches would lead to a centralization in the relays.
At some point, users will have to pay to maintain the ecosystem. No one is going to finance servers and development for us forever. The more Nostr scales, the more necessary this will become.
One possible path could be a paid client with agreements with relay operators. From the client, the user can select which relays to use and channel funds.
I doubt most people are going to finely tune their relay list like this. easier to just do WoT and put people outside your trust network into the bottom of threads with media blurred.
It could be. I think WoT with some manual WoT bump-up is a great option for an Inbox relay. WoT is terrible for Outbox, though (a simple paid relay is easier). The key part is making it setup of the multiple relay lists we already use easier. If I sign up for Nostr.wine, for instance, there should be a tool that puts each of the relays they serve in the right list for me.
How do you onboard new users into WoT? New users won't be in WoT initially which still means their on-boarding experience is tarnished by spam. Paid relays help, but then that is still expecting new users to finely tune relay lists and pay for relay services to start with.
Sometimes freedom of content like this just requires the user to go through some hoops to see only the content they wish to see.
If each user has their own WoT inbox relay, and the WoT is just local of that server, then there could be a simple approval screens for new pubkeys that reply to the user. The user could approve each of those at night in some simple left/right screen for ok, not ok. Then something to mass reject all the spams. Maybe some little tool to evaluate the video and image content and already separate into a list for easy rejection.
If the WoT is shared, I think an active moderation could be very cool. Similar to what I said before, but for all accounts at the same time. Which I think is already happening in some form in existing WoT relays.
WoT doesnβt mean ban them outright, at least in our implementation. It just means put them in a low priority section of the app
Your description of Inbox is very different from the description in amethyst's menu.
Can you please clarify? 