Assume it was 1-click and 100% automated to deploy, run and update a Hub. How much per month would you be willing to pay to run one? No direct economic incentive other than you know you're supporting Farcaster decentralization and you have your own copy of the network. OK, if you're answer is you wouldn't run one.
rn i would pay even 30$/month if it’s powering some projects of mine
$20/30 per month if I knew it was going back to the protocol itself One click to deploy would be huge, espc. if there’s continued support for data replication w/ Postgres
@neynar is doing exactly this for $30/mo, felt pretty fair
it's already costing me quite a bit to run it, as I'm also running an Ethereum node so that I'm 100% permissionless. If I were a SMB I'm still not sure of the proposition beyond the technical coolness of it all. 💜 the project, but I am not sure I see the path around the behemoths @daivak and I were just discussing
Might be more compelling to me if doing this also let you “save” your own casts forever since otherwise they will be eventually deleted. Also wondering about potential “pihole” like benefits (ad blocking etc) from running your own hub.
Depends on the power cost to run but I can see anywhere between $25-$40 a month.
I wouldn't pay to run one. No need for a copy and I'm sure there are others who would benefit from having one and would pay for it.
1/2 i would pay 0 ethereum already showed that people don’t run nodes if not incentives by staking (EF is trying to fix it by paying incentives out of their own pocket https://esp.ethereum.foundation/run-a-node-grants) BUT…
Wouldn't mind paying $20-$50 if my cryptoarts sell on the regular. Happy to support the protocol if it remains decentralized.
tough without economic incentive imo would be nice if my annual farcaster subscription is bundled in, gets me a badge / premium features, etc.
Currently paying about 60/mth. Purely for support and learning at this stage.
1-click, 100% automated. Probably somewhere between $5-10 a month
I wouldn't pay to run a Farcaster hub if I could use the Warpcast client for free
$10-15 with the current monthly income I have . I may shell out more when i start earning more. What's the system requirements btw. Does it have to mandatorily run 24*7. Wouldn't mind running one on my work machine if it's feasible 🙃
As a dev, the ~$30 average answer seems reasonable as a non-dev, $0. I think the easiest way to increase operator count is to boostrap support from people who already run ETH nodes by having a DappNode package like I mentioned in Boston
People saying $30 is bonkers to me with no incentive. I'd pay around $5/month? I care about a lot of projects, so paying double digits per project would quickly get hella expensive. If they were $30/month, I'd see if there was a way to sponsor one.
would do it for the tech, not sure about the money part though, $1/day ($30/month) is good number what i would like it that i'm able to understand important things that hubs do, how they store data, and like any other useful resources so i don't end up running a (malicious) software by accident
i already index all of Farcaster in elasticsearch for @alertcaster using the APIs. running a hub adds cost and complexity that's unnecessary today
I'd pay 50-100 a year if this came rolled into a personal archival/rewind service a la Jumbo + Pocket + Google Photos.
Already running one. The requirements on the spec are very low, so I just run it on servers that still have spare resources without spending a dime 🥳
$44/month if I got a special NFT each month
Immediate reaction: no After giving some thought: I would if there is some form of social advantage: 1) list of supporters; 2) impact on my casts visibility; 3) early access to new features; 4) coffee with you once a year :-))
$100 per month seems reasonable for a crypto dev who‘s prototyping a new app. But make it easy to cancel and use “virtual cards” like privacy.com. $10 per month for a student.
Yes if I get a vanity badge (like signal)
Running an IPFS node on a Pi Ubuntu server, so low power machine. Have lots of space and bandwidth for more. If install of the Hub would be as easy as installing Kubo on a Pi would do it immediately. Would pay € 5/month to run one.
Depends. If there is a credible alternative to the MM-run hub, either bc: 1. A non Warpcast client succeeds , or 2. People start using Warpcast to follow other hubs …then if do it for $50 a month to help censorship resistance. If in practice a non-MM-approved cast appears nowhere, I wouldn’t run my own hub.
For dev stuff it might be worth it for some niceties If I get to have preference for my app’s queries then it could be worth $100s per month
I think helping decentralization is mostly stated preference.. considering warpcast api, I think $0-50/mo is reasonable and probably wouldn’t pay much more without an economic incentive
15-20$/m , supposing I also use that node data directly for an app or dev. No economic incentive for running a node is ok, just saying I'd need another incentive than just noble support of decentralisation.
500-800 INR/mo seems reasonable 👀 ($6-9/mo - sounds low but in local currency i feel it's fairly rated)
Would max out around $10 as a user with no incentive
You would have to bring the value down deeper than support Farcaster for me. Why? What does me supporting a hub do for others (vs a company)? Economic value does not equal all value. But I don’t hear intrinsic value yet in that pitch.
To improve decentralization, it should be the opposite, right 🧐? I think there should be incentives to users running a hub.. Or am I missing something here?
Perhaps some degree of Hubs support could run through a 501c3 structure, creating economic incentives through tax deductibility
$10 no brainer $30, fine but I have to cancel Netflix first $50, I'll want an NFT or something commemorative... often $100+ moar stuff and some kinda financial incentive (cost saving as dev, revenue slice, or some other unlocked monetization options)
$5 with badge to indicate support like Signal does as others have mentioned
I've hosted a bunch of free services over the years (usually instances of whatever OSS project I'm working on, like ssh.chat), I've converged to a $10/mo budget limit per for-fun project. If I were building something on Farcaster (hope to someday!) that'd probably be my limit for hosting infra long-term.
30 USD if I’m building on Farcaster 2 USD else
I’m confused Aren’t we supposed to get paid to run a hub?
I think having a visible flex is a good option for most. Ideally some kind of NFT that displays the status and then warpcaster can pull that info in.
$10 and a sort of incentive to run one. This will boost no of hubs on the network.
As a consumer, I wouldn’t pay. If I was building on it, I would consider it. Up to $30/mo, which is the most I typically pay for any side project.
my barebones EC2 Hub + Postgres napkin math: - GCS 1 TB data and 50,000~ users. - storage is ~$23/mo for 1 TB. - data egress costs ~$12/mo for 100 GB transfer. - operations cost ~$5/mo for 10M Class A and 50M Class B ops (write/read). total monthly cost is ~$40, and yearly cost ~$480
As an enthusiast for decentralized social media, $25 to $30. I've not yet had a spark of inspiration of what I do as a developer with a node.
Most ppl who are willing to pay are probably all US/male/dev or a similar profile . To get beyond this group i'd consider creating a prestige community of ppl running hubs, with unique perks, early access to projects building on FC, and probably...keep some hopium of speculation about incentives...
Im more or less a regular user but I'd pay $10 a month if I could fork it and have a federated instance catered more for my needs wrt custom feeds and moderation practices.
really want to help projects to grow, but my financial conditions don't let me
I wouldn’t mind running one on spare computer at cost, but can’t guarantee uptime In place of economic incentive, premium features or a special badge would be cool
Worth noting this type of hosting from home is (annoyingly) a breach of many ISPs consumer contracts. Not that it really gets enforced.
A hub doesn’t increase decentralisation. It adds one node to the network as it’s one point of control to be ordered to stop providing the service. Decentralisation happens through dispersion of devices across the geography of the network. It’s why the internet was designed for geographical dispersion post a nuke.
Prob wouldnt run one tbh, Would be more inclined to do it tho with some incentive (even if not financial)
I would pay like $10 if at least I’m getting a nft or some kind of badge on my profile showing that I’m supporting Farcaster
$10-$20. $300 to run it at home on my own hw, where I can use it over my VPN, etc.
As a dev I would pay $0 because I can do it for free. Actually I will start running one, on same hardware as my ethereum node
What kind of resource usage would it require? Like an ipfs node I can run on my desktop or could I run it on an old laptop?
Given, we're global. And to effectively meet purchase power parity. We could aim for the price of one Starbucks coffee at that respective demography.
I hate to say that I pay more than many have mentioned in the comments. In fact, I prefer my practical commitment as the best proof of that.
If (when) we had an app that ran on Farcaster, spending $100/month to run a Hub would be a no-brainer.
Probably around $20? On a Linode node 🤷♂️ if I had a non-commercial side project built on top of Farcaster, I’d probably pay more (up to $60?).
If I could run it from home on a Raspberry Pi / Intel NUC / homelab server I already have running 24/7: I would consider it. If I have to pay a monthly fee to a VPS: quite unlikely. And that’s from someone who’s run Bitcoin and Ethereum (pre-PoS) nodes for years without economical incentives.
$0 because Warpcast's API is really good. On a similar note: Stacks/Gaia also had its own "hub" (data layer). Didn't go as planned, might be an interesting discussion to read: https://forum.stacks.org/t/rfc-moving-gaia-data-to-home-computers/14706/6
what's the expected near/medium term hdd space requirement?
measured against other “subscriptions” $5, no more than $10
Zero. I have no particular reason to, and don’t have anything I’d be particularly upset at losing if some Evil Attacker erased my cast history on public hubs or something.
35 usd/month is reasonable as a dev but for non-devs that may seem a bit much 10 usd/month for students? 👉 👈 nfts for hub runners
Still learning so would not be able to help it. But might consider in future when I have some extra avenues of income.
For services I appreciate but are not essential (and with no additional functionality, I patron around $3/mo a month without blinking an eye.
Would be willing to do a coop kinda thing with others and kick in $30/ mo.
$10/month max. That will require me to be very active on the network for it to go beyond 6 months. There should be a path for shared hubs, where small clusters of users share/rely on one hub and that way cost can be amortized.
As a dev, anywhere between 30-50$/month As a user/non-dev, upto 10$/month if I get some extra benefits: maybe a NFT, part of a community who run hubs etc
For Love of Farcaster and whole of Web3 ethos, degen, cyberpunk vibes, I do it voluntary, free. 💕
$10, $20 is expensive but I'll do it anyways. Higher than that no
Cliche, but Charge to run LLMs on the protocol? W/f caster has a dedicated community of builders. Smort people to learn from. Channels should make it easier to train a model on a given topic. Sell the data. Here, take mine 🤷♂️
I wouldn't pay as I'm running my own Hub. It costs pennies and it makes me feel good. For a dedicated Hub I'd host it myself alongside the infra for my app.
I wouldn’t pay for it but am going to run a hub on my NUC soon
realistically, you should expect people to do as long as the price is around $10-$20 (which is what most people already pay for a Twitter subscription)
What is the resource reqirement? How much CPU, mem, storage, and internet bandwidth would it put on my DappNode, for example?
What was missing from eth nodes was an ability to index, so as to make querying the state of contracts much more powerful. Farcaster has the same problem. Compiling cast threads and reactions is really what developers need. An extensible indexer so people can make a Reddit style app would be amazing…
Must each Hub have a copy of the whole network (don't know in detail how it works)? Then it would make sense to have different tiers based on the percentage of the network you're storing. 10$ is nothing in some parts of the world, but a lot someplace else
Probably $0 if there was no incentive at all. Around $20-$30 if it gave me a non-economic perk or some kind of voting power in Farcaster roadmap
$10-20 for sure. In reality I would likely put it on a server I run anyway ($40/month) because I self-host, and self-hosting is something I do spend money on
Ideal situation would be I could run it on a NUC at home but with just a one-click setup that rarely requires updates manually, and would be happy to pay the electric/bandwidth for that. If trying to economically abstract, maybe $20/mo?
I’d be open to run one, but it would definitely help to have some kind of direct/indirect incentive as running a node/hub, in my experience, almost always ends up getting more involved than what people are originally led to believe. For the culture? At most $10-20 and maybe a bunch of soft benefits like badges… etc
I tend to think it's the wrong model to expect devs to run infra (and pay for the privilege) out of the kindness of their hearts.
What’s the ongoing cost for doing maintenance and updates?
An incentive could be a different or preferential access to Farcaster network, doesn't have to be an economic one. IMO I don't think the absolute standard user would pay for a hub just for the 'fun' of it without any goodie at the end.
As a user I'll pay $10/mo if the social network is high-quality, no bots, interesting conversations. Then if you pool these payments together and distribute to node operators, and I can see a graph that shows "Decentralisation Factor" (aws, gcp, etc), I'll be continuing to pay to keep the network up an decentralized.