Nostr is a stupid simple P2P protocol that works, built by builders
https://github.com/nostr-protocol/nostr
Fun facts about Nostr:
* Nostr stands for "Notes and Other Stuff Transmitted by Relays". It is an odd acronym, but I like it.
* Nostr uses websockets and relays to build a really simple P2P network. We also steal a few ideas from bitcoin (ECDSA ids, schnorr-signed events).
* Relays are simply dumb data stores for events that clients publish and subscribe to.
* Clients don't trust relays to be honest, so all events are self-signed. Your pubkey is your userid.
* It is stupid simple to build a Nostr client. You can easily do it in less than 400 lines of JavaScript. And it runs in the browser.
(shameless self plug) https://github.com/cmdruid/nostr-emitter
* Nostr is powerful enough to host chat apps very easily. Here is a rip of Telegram, running on Nostr:
https://anigma.io
* There's a lot of fun things you can do with Nostr. Check out all these cool projects!
https://github.com/aljazceru/awesome-nostr
* We are constantly discussing how to improve the protocol. Come join the conversation here:
https://t.me/nostr_protocol https://anigma.io https://damus.io https://github.com/nostr-protocol/nips
Thank you for reading my nostr shill post. I did not create nostr, nor do I get any monies for promotion. I just think it's really cool and I have a lot of fun building stuff that punches though nats.
If you have any questions about nostr please feel free to ask.
Also, Happy Thanksgiving to everyone! I hope we're all feeling fat and sassy today. :-D
131 comments
[ 7.3 ms ] story [ 256 ms ] threadSee FAQ, second question.
You can request very broad subscriptions from relays! For example, here is a site that subscribes to everything, showing you a gods-eye view of events streaming into a relay:
https://nostr.info/relays
Events have different "kinds", so you can filter this based on the type of traffic you are looking for (like public posts or user profiles).
Platforms like damus.io are more user-friendly, and offer better tools for discovering users and content.
You can subscribe to a user's feed via their pubkey, so discovery methods typically revolve around learning pub keys.
I doubt anyone has ever been successful into signing up on any social platform and just followed the big names that are suggested automatically at the beginning or based on some "key interests" you select.
But hey, if you want that, it's easy for a third-party website to grab a ton of public Nostr data and build custom recommendation lists and whatnot.
There are still features that many apps will need such as tying multiple devices to an identity, abuse prevention for relay operators, etc.
- Proof of work: computing some hash, which is not enough to be onerous but enough to reduce spam
- micropayment over Bitcoin lightning network
Bitcoin consumes 111TWh annually, the power consumption of the Netherlands, and emits 62Mt of CO2 per year, the same as Belarus. It also yields 42kT of e-waste per year.
Each transaction produces 650kg of CO2, consumes 1160kWh of power (as much as 40 days consumption for the average American home) and produces 450g of e-waste (about the same as hucking your iPad into the garbage can each time you transact on-chain).
97% of all Bitcoin mining hardware will never successfully produce a single block in its entire useful life, going from factory, to space heater, to garbage can - while about 60% of all the power consumed comes from oil, natural gas and coal. So whoever sold producers the offsets I assume you must be alluding to better have replanted the entire Amazon rainforest by now. (Quick spoiler, carbon offsets are also a scam, generally speaking).
To say you're going to need to get some sources is an understatement about as large at Bitcoin's environmental footprint.
You can find all this in [1] or you can just reverse it yourself from the specs of the latest AntMiner and the current hash rate. Some napkin math is all you need.
I'm honestly amazed people still believe something so trivially falsifiable, but with everything else going on in 2022...
[1] https://digiconomist.net/bitcoin-energy-consumption
it only consumes as much as the complexity required to roll out new blocks. As its price goes down so does the energy required to find new blocks. It is not static.
tl;dr: price per coin represents something of a bounding function on consumption of resources (including mining hardware and electricity) but it's not as tightly correlated as hoped.
My point though is the idea all that consumption has been papered over by what I assume is buying some offsets - and now it's magically carbon neutral - is silly and obviously wrong.
1) miners profit from energy surplus. They do not create new power plants. Nobody is firing up a new coal plant to mine bitcoin. That is absurd.
2) miners have incentives to find and use wasted energy. For ex, flare gas recycling. Which actually helps the environment.
3) miners can turn on/off at will, and produce energy loads on demand. Which means they balance out the energy grid. Especially from erratic energy sources like wind and solar. This is already being deployed in some states.
You are fighting on the wrong side buddy.
(2) burning flared methane is better than not but not flaring at all is best? There’s no such thing as stranded power, there’s only missing transmission infrastructure which is being incentivized not to come online through profitable waste at point of generation.
(3) raising the baseline usage level through waste then turning it off when there’s a brownout level crisis is just an asinine plan to manage load that costs everyone money. Active demand management is an actual solution. So is grid storage. Also, “some states” is Texas, famously the worst energy grid in the US, home of the blackout. Meanwhile New York is banning wasting fossil fuel energy on miners.
Anyways you seem to bring no data to the table other than a “nuh uh” and some obviously flawed apologism.
No word on the ewaste?
It appears to me you’re using the “facts u dislike” acronym expansion for FUD.
[1] https://grist.org/technology/bitcoin-greenidge-seneca-lake-c...
The question is how much so called stranded power is there, what percent of Bitcoin energy usage does this account for and is there really nothing useful we could do with it? Finally with “stranded coal” is it better to just not use it?
This is just a talking point used to distract from the actual issues. In reality the percentage of renewables wasted on bitcoin mining is at an all time low. The question at issue was whether all this is carbon neutral, and no, it’s definitely not.
In the case of flare gas, it is a byproduct of gathering oil that is unavoidable. The oil is economically worth collecting, but the extra gas is not. I think this can be more finely split into natural gas vs methane-- sometimes the natural gas is worth collecting, but the methane never is.
You could argue that there should be laws to force the methane to be collected and used elsewhere as a condition of accessing the oil. That might be overall better. But prior to bitcoin, that perfect solution was not really implemented. In this sense, bitcoin mining of flare gas is a step forward.
Note that the argument is not just that economic value is created by mining bitcoin, but that running the gas through a generator leads to cleaner combustion and thus less greenhouse gas emission than simply lighting the gas on fire. In this sense, the bitcoin flare gas mining operation by itself could be considered carbon negative (again, surely all bitcoin mining is not). The bitcoin mining pays for the generator. So bitcoin mining is giving the industry a subsidy that enables reduction of environmental harm. It is not perfect, but it is an improvement.
Prior to bitcoin they flared it - or didn't. Now there's an economic incentive not to find another solution. Now the oil and gas companies are incentivized to prevent that better solution because there's something in it for them not to. So we make it harder to solve 100% of the problem because this gets us 5% of the way there and pays us not to.
This is a really disengenuous point. Mining works probabilistically, and mining pools payout based on smaller units of work that probably have some probability of finding a block for the pool. The fact that a block itself is a large parcel does not make the system less efficient.
Value judgement aside the question was “is it carbon neutral” and the answer is a resounding no.
What is your source around Hal Finney? I have not heard that.
The fact we have a system where 97% of all the equipment isn't used to ever do anything useful and instead "increase security" is an insanely wasteful system. Nobody has quantified what level of "security" is required. There's no feedback mechanism to pull back based on need because the need is fundamentally unquantified. It's a grey goo style uncontrolled positive feedback loop.
Each time more miners come online it's lauded as "more security is better" - but how much security do you need! The answer isn't "as much as you can afford period no follow-up questions."
You don't have 75 seatbelts in your car because "security." You don't use a dump truck to take your kids to school because "security." And you don't use a global army of computers consuming 100+TWh/yr to process 2-3 tx/sec because "security."
Ultimately it's a half-baked security model, and Bitcoin is a half-baked proof of concept that escaped the lab and gained a cult following.
> What is your source around Hal Finney? I have not heard that.
My source is Hal Finney [1]
[1] https://twitter.com/halfin/status/1153096538
> The answer isn't "as much as you can afford period no follow-up questions."
You are right, and this isn't how bitcoin works. The amount of mining should be determined by the block reward, market price, and fees per block. Mining in excess of this is not economically rational, and shouldn't happen. So it is not "unlimited security" at any cost"
Where in that tweet does Hal characterize bitcoin as ridiculous or unsustainable?
Sure it would, it would change that resource cost per transaction.
> You think bitcoin is a total waste, I get it. But that statistic doesn't help your argument.
I disagree, if 97% of servers at AWS were there for some hand-wavey notion of 'providing security' without quantification I think Amazon would be roundly mocked.
> You are right, and this isn't how bitcoin works.
Yes it is. You just re-stated my position with slightly different wording.
I said the security model was "as much as you can afford." You can afford block reward plus fees times price. The security model is "spend as much of that as you can without regard for what you need to achieve security." The issue is that block reward plus fees times price is not a function of how much security is required.
> Where in that tweet does Hal characterize bitcoin as ridiculous or unsustainable?
I never said 'ridiculous' - obviously I don't think Hal would make that claim, so let's stick to what I did say :) I think it is not unreasonable to extrapolate from his tweet that he believe that at the limit CO2 would be an issue. CO2 itself is an issue of sustainability. You can disagree with that interpretation, but I do not think that an average unbiased observer would find my reading unreasonable.
You seem to be trying to win an argument at all costs, putting words into my mouth, running with uncharitable interpretations and arguing in bad faith - in this thread and the other. If you read carefully you'll find I paid close attention not to move goalposts. I'm going to cut it off here. It's particularly silly because you've already admitted you agree to my premise that it is not carbon neutral.
Have a good evening though.
Again, you fail to understand that the number of a machines that find a block could be changed by making the blocks smaller. You could do that by also reducing block size to keep the total number of transactions the same. But the larger point is that mining pools make it so that mining machines are paid for partial work, even without finding a block. This enables the utility of mining to be measured by hash rate, rather than by number of blocks solved.
You called it a gray goo style positive feedback loop. It is not, there is a upper bound on how much will be spent. There is no feedback loop as well, if the price goes up, more will be spent on mining, but that does not feedback to make the price go higher as required for a loop. Moreover, the block reward is exponentially decaying.
Your reading of the tweet is that a person saying a system should use less CO2 implies that person believes the system is unsustainable. I don't think that is reasonable, e.g. an Airline CEO launching an initiative to reduce the airlines CO2 usage would not believe that the airline is unsustainable. I did misread the other part of your statement-- you said bitcoin is ridiculous, not that Hal said that.
Also he’s a central banker, so not exactly motivated to fix it.
I agree, except for the bit about public keys as identities.
I think public key identities are a step in the right direction, but there’s still a gap between that and what the ultimate solution is going to wind up being.
We need to have some layer of indirection between user identities and public keys so that users can do things like rotate keys, have multiple keys, and recover their identities.
I don’t know what the right solution to that is; I think it’s an open problem and probably one of the most important ones to solve. Keybase probably came closest to a good solution, but it wasn’t decentralized.
https://github.com/bitcoin/bips/blob/master/bip-0032.mediawi...
Easy rotation and recovery of individual keys, but you do have to protect your master seed.
Nostr also supports user verification through DNS hostnames.
https://github.com/nostr-protocol/nips/blob/master/05.md
Both seem to use a “signed rotation” approach. Algorand keeps your public key stable while adding metadata that your spend key has changed and links the two. Atproto similarly uses the recovery key to sign a rotation op which can regenerate your signing key, additionally readjusting the tree to preattack state (by setting prev of the rotation to the last precompromise state).
This seems like an improvement of some kind, but still leaves gaps for lost keys. Keybase style approach, or multisig social recovery may also help.
UCAN also seems interesting, JWT with extra steps and attenuation. But orthogonal to this issue for the most part.
And even if you're ok with the master key, the only way to solve this without centralized providers is with blockchains. A blockchain for rotating keys doesn't make sense.
But I do want to know if you're ok with a master key and subkeys that can be rotated.
There are certainly solutions, but I don’t know what the best solution is, hence why I called it an open problem.
An example solution would be something like having your identity be a hash of your initial public keyset, making each key have a set expiration date, adding new keys by signing them with one of the existing keys, and then storing all of the rotation operations in a transparency log.
“the only way to solve this without centralized providers is with blockchains”
That’s not true; you probably want a transparency log, but that doesn’t require blockchains.
Multiple keys: nothing to change. Works like that now.
Recover your identity: Well, if you want a well-known identity use NIP-05/NIP-35 and just change your .well-known/nostr.json file to point to your new identity, the one that hasn't been stolen. Hopefully nostr clients of your followers will respect that (who knows what programmers actually will do).
I think these problems are easier than you think they are.
One of the neat things about nostr is that while it has already been used to build a decentralized Twitter like social network, the protocol could also be used to build encrypted P2P chat, traditional discussion forum, alerting/push style notifications, and numerous other applications.
https://abovethelaw.com/2022/11/hey-elon-let-me-help-you-spe... is a pretty decent rundown of a mix of these things; it is specifically pointed at Elon Musk's decision to buy Twitter and make it a haven for "free speech" but it is a glimpse at what is in the future for anyone setting up a "free speech" platform.
My experience as someone who has been running a Mastodon server since 2017 is that while "we are all for FREE SPEECH, we only block what the government ABSOLUTELY requires us to block!" sounds noble, in practice nodes of the Fediverse that say this become havens for people who are only there to be assholes to other people, and any sane admin will sigh and block the whole server, because it's just going to be a continual source of rude nasty bullshit.
I like that nostr abstracts this problem away from the relays. Relays only focus on storing data and handling subscriptions. They can choose to censor and/or curate content if need be, but it's not their concern.
It's up to the client to come up with a solution, and that client can be a platform or a protocol of its own.
edit it also feels really great to work on that problem from the application layer. I can come up with a solution that isn't confined to the parent protocol.
> I don't know, but I imagine it has to do with the fact that people making social networks are either companies wanting to make money or P2P activists who want to make a thing completely without servers.
Except it has been done. In fact, that’s literally what KaZaA was with its “superpeers”. And what they realized was that by making a semi-decentralized system, they just introduced the weaknesses of both systems (slow downloads via peer-latency and network limits + easy censorship by killing relays/nodes). In addition, this is exactly how IRC works, despite the fact that it’s mostly used with a few nodes these days.
I’m not against semi-decentralized systems. They’re great and help deal with some scalability problems; but they don’t solve for the number one issue most people moving to decentralized are seeking (anonymity, privacy and free speech), so it’s not fair to compare it to platforms/protocols that do offer those features.
And I disagree very much that IRC is "semi-decentralized". IRC is completely centralized, it is just chat rooms on a server. You have to register on each server and each server has full control over its rooms and users.
More seriously, this is the second time I've seen someone on here characterize IRC in this (very wrong) way in the last day. Where is this coming from?
IRC networks are made up of servers that relay (hence Internet Relay Chat) with each other. You connect to one server and you can communicate both with people local to that server and people on other servers that are part of the same network (including ones that server is not directly connected to). Channels prefixed with # are shared across all servers in the network, while channels starting with & are local to that server (though rarely used).
See https://drewdevault.com/2021/07/03/How-does-IRC-federate.htm...
But as for peer latency issues or easy censorship by killing nodes, I don't see it. Nostr has fan-out, but not as much as RSS does and I don't expect superrelays.
I also don't follow you on the issue of anonymity or privacy. The guy who started it fiatjaf is anonymous. We don't know who he is. And you can be too. Just make up a key, create and sign an event, and push it into whatever relay takes your fancy... through Tor if that's your thing.
Misinformation/disinformation is a thing and is not considered healthy for discourse. It also seems to be willingly embraced when it serves one's biases. Pity.
Plus, at the time of the censorship, facebook and twitter claimed that the misinformation was that the laptop did not belong to Mr Biden. At this point, every news outlet of repute has admitted the laptop belongs to Hunter.
The stupid laptop being a Russian forgery was misinformation/disinformation that affected an election. The NYT and WaPo have faced up to it by now, but middle-aged extreme partisans will post "cope" like they're millennials from 2015.
Or "butthurt."
If you didn't allow social media to get you invested in these manufactured pseudo-political pseudo-events, you'd be free to realize that Trump and Biden are both obvious idiots and crooks; that people's gender identities, biochemical makeup, and surgical history are obviously completely irrelevant to complete strangers; and that many other things that you waste your time constructing your tribal identity around are obviously complete bullshit that someone artificially planted into the discourse for their own benefit. That someone couldn't care less about your best interests as a person, anon291.
The problem isn't that Twitter and Facebook censor one side of those debates and promote the other. The problem is that people are using them in the first place! That way, social media is the thing that forces those debates to exist in the first place - by making you feel like you have a stake in events that are completely remote from your life. Whose agenda is being promoted is tangential to the fact that these unelected, unaccountable corporate bodies have entrenched themselves in a position where they can basically replace our individual world-models with this sort of outrageous nonsense. That's what's fucking democracy up, and in the most elegant way, too: your right to make free choices between alternatives is preserved, but only inconsequential choices are ever presented.
On the Internet, nobody knows you're a dog; on corporate social media, nobody cares that you aren't a dog.
At least in the UK, this became a matter of renewed public interest for two reasons: firstly, a proposed reform of the law to remove all gatekeeping from the process of changing one's 'legal sex', and secondly, an appalling case of several imprisoned women being sexually assaulted by a man (Karen White) who had been incarcerated alongside them due to having a 'legal sex' of female.
It was left-wing feminist groups, who organised largely offline to begin with, that reignited this debate. This wasn't some artefact of social media raging, it was a grassroots effort to halt and reverse a change in the law due to its clearly negative effects on women.
This is also a debate that has been ongoing for decades, long before social media websites even existed. Janice Raymond wrote what turned out to be a quite prophetic book on this topic in the 1970s, for instance. Renée Richards was stirring controversy in women's tennis at around the same time. Much of what you'll hear on this topic these days has already been covered by radical feminists for many years prior.
By bringing up the Karen White example, aren't you basically saying it would somehow be less appaling if it wasn't a trans person perpetrating the assault? Because I thought this sort of thing was abhorrent regardless of the salient details of a particular case?
Social media only makes it easier to focus on "which cage should we use for transgender people", and so much harder to ask ourselves "why are we putting people in cages". Or, as per the other example, "did Hunter Biden really lose his incriminating laptop?" vs "why are we letting ourselves be governed by people with familial ties to criminals?"
Social media and the polarizing meaningless debates that it enables serve the purpose of precluding people from focusing on the latter kind of question. (Which is already hard enough as it is, because it involves actual thinking.) If the public conversation is retreading ground that was already covered in the 1970s like you say, doesn't that mean that our society is regressing? Shouldn't be worrying first and foremost about that, since that's where we'd find the root cause of all the more specific issues?
How many people feel good for having the correct in-group opinions, while their contribution to e.g. the trans rights debate only goes as far as canceling JK Rowling, or, conversely, going to a Jordan Peterson talk? How many people have even heard of the actual examples you mention, as compared to the number of people who only know "uhh, so there's a debate on the Internet about some abstract hot button issue, and I'm required to pick a side in order to participate in society"?
A couple years from now the topics may be completely different, they'll just find another scapegoat or another thorny bioethical edge case, but the medium of debate will still be largely the same ol' Internet, and the AIs will only have become more effective at sowing discord.
(No, the president’s son didn’t actually leave a laptop at the repair shop of a blind man in a different state and then not come back for it. And while some emails were verified through DKIM, the one you specifically mentioned wasn’t.)
There are solid laws protecting copyright everywhere yet it is stilm trivially easy to find copyrighted content available for free online. Laws dont mean anything unless they are or can be enforced.
P2P services OTOH work on a decentralized and pull model. Users share and only subscribe to the content they're interested in. Censorship is distributed, and it's a problem for people who don't wish to see specific content. It's the way the internet works, and the existing approach of removing sensitive content applies to P2P services as well. Since there are no advertisers to appease, it's not an existential problem.
The question was, to wit, >How do you keep CP off it?
Your answer was: >I don't, just don't look at the CP.
The problem really in question here is:
You've just created a new distribution method for this type of thing and punted the consequences for someone else to deal with.
(Which is totally cool imo, but newsflash, expect to be the subject of a hit piece some time in the forseeable future. It shouldn't take long; either for actual criminals to set up on it, or for LE to do it to "snare unsophisticated actors").
Welcome to the Internet, where we can't have/make nice things anymore.
A new protocol that makes this content more accessible isn't an issue with the protocol, but with society and how we decide to deal with it. If CP is found to be served by nginx over HTTP, is that a problem with nginx or HTTP?
If anything, centralized services only make the problem more difficult to address, since they're expected to serve the demands of governments, companies and law enforcement agencies worldwide, while somehow being the arbiter of free speech. Those are impossible goals to reach, and go against the original design of the internet.
This discussion is as old as P2P protocols. I'm sure that if Nostr became a popular way to share copyrighted content, governments would try to fight it, just as they've done before. But at the same time, censorship is not something a protocol should care about, and just like BitTorrent thrives today, with enough interest, Nostr would also find a way to persist.
So, the OP has just given people a whole new way of making themselves liable for illegal content, just by running the P2P software in question.
In most cases, the police aren't interested in hauling a service provider into jail, they want to haul in the person who put that content there. But this does make a nice little lever that the police can use against any service provider.
Commonly filtered things (account block lists, post flag lists, filter rules, etc.) could be shared via the same system — indeed there could even be competing versions and everyone could follow their preferred filter source.
Users would also likely run statistical and machine learning based spam and content filters locally (perhaps on a personal relay/server of some sort, or an account on a shared one) configured to their preferences.
I would expect the infrastructure running such a network to be in the same position as Signal, who do not know the content of messages and can't censor them, leaving individual clients to figure out blocking etc. (albeit the client side options as well as ways to share configurations etc. would need to be much more advanced for a social network or similar than for a messaing app).
A lot of p2p protocols cheat with relays, it is really hard to traverse nats otherwise.
Nostr can be used for peer discovery to bootstrap a direct p2p connection.
You could also use a client/relay hybrid application, similar to other p2p networks. That would be fun to build. :-)
Relays are important for two reasons: peer discovery and communicating when one of the parties is offline. Same as with other p2p networks.
Edit: i RTFA. Sounds like relays can be run by anyone. That sounds p2p enough to me.
OTOH websockets are hard outside the browser :(
There is tremendous value in a much simpler protocol, especially if it can deal with the identity migration issues that Mastodon has faced since day one.
From what I've read, it sounds like the interoperability issues come from not implementing AP the same way as Mastodon rather than anything specifically protocol-related. Hence, e.g., Pleroma going out of its way to be Mastodon compatible as a design goal and GoToSocial having lots of issues because it hasn't yet gone that route[1]
[1] e.g. they have an issue where some client apps don't work with GTS because GTS doesn't advertise a version number sufficiently high enough - because the apps are looking for "v > 2.10 because that's when Mastodon introduced feature Z".
Here are the ones gotosocial have issues with: https://github.com/superseriousbusiness/gotosocial/projects/...
Not to say I don't believe you about Mastodon, I am sure they have their quirks as well.
I'm not piling on nostr here, that's an issue with ActivityPub as well as most decentralized platforms/protocols?
I wrote a simple directory you can use to find users and channels here: https://nostr-fzf.netlify.app
Of course this doesn't work for people wishing to remain anonymous.
Though at this point I probably wouldn't use that either.
Maybe this has been cleaned up in the interim, but at the time a bunch of apps for it already existed so it was probably to late to go back.
There are a number of web servers that host content (either for free or for money) [called relays].
Clients download recent posts
Identity is based on public key, allowing users more control and the ability to easily change relays.
So is RSS + pubkey based identity the right way to think about this?
Relays have to figure out how not to get smashed with too much data. I predict they will require an account/login at some point. But you can post to multiple relays and drop relays that don't serve you well at any time.
Initially restrained from posting it, since it's possibly rather grumpy, and the authors seem to have fun with the protocol, but here it is.
The protocol doesnt want users to run their own servers, but there also aren't really any incentives to run relays, so what happens when it gets big enough that running a relay is expensive or non-trivial? I feel like it would just fall back to users running their own servers like in mastadon.
It feels nice when a protocol has its core devs on the same wavelength as application devs. That's the feeling I get with nostr.
It's not shilling. It's recommending. Shilling is a bad thing. It's a simple thing.
Can you elaborate on this point? It would seem that meshing relays would've facilitated the dispersal of updates.
Nothing in the protocol specifies relay-to-relay communication, but nothing stops them either.
But I wouldn't be surprised if he was discouraged from continuing this work because ultimately people like ActivityPub because server admins can be little dictators that censor and ban at will.
Mastodon, Pleroma et al. are not really ActivityPub implementations. Each of the fediverse clients implement some random 10% of AP and just try to get compatible with each other using dirty hacks.
ActivityPub isnt the only distributed thingy to learn that lesson. There are a reason the various ultra-free-speech hosted on the dark web, things never take off.
From the sounds of it, you have the protocol but not the policy. Which is by itself huge that they are separate, but now the clients (?) needs flexible policy, no? Otherwise its just going to turn into a billion people all talking in the same room, or your going to have a ton of tiny rooms with no activity. The discoverability of interesting rooms will be difficult. Its sorta the IRC problem in a nutshell (or discord/etc). Balancing the noise, vs the quiet is the difficult part (AFAIK).