Make onchain actions reliable and smooth directly from the feed—frames and token links. No app switching or deeplink bugs.

New users don’t need to download a separate app—it’s already there when you sign up. If you get a reward, a $degen tip or an airdrop in your first week, it just works.

Today, Farcaster is much smaller than web2 social networks. However, if the % of active users with funded wallets is really high, we have a near term opportunity to be a “small-population-but-high-GDP country”.

Developers can permissionlessly launch a frame, have it appear on Farcaster clients that have integrated wallets. “Get to your first 10,000 transactions by launching on Farcaster”.

Yes, you can launch on other networks, but their algos hate links and the drop off from seeing something in a feed to website to wallet is going to be high.

1. It's 2024 — the consumer web is 30 years old and the modern smartphone era is almost 20 years old. Internet software is a mature industry. Obvious ideas (regardless of whether they involve crypto) are solved.

2. So if it's an existing idea + crypto, you usually getting, at best, a 10% improvement. It's not a 10x improvement. And if it's not 10x, you won't really get people many people to switch. (And for the last 2 or so years, crypto's brand association has been more negative, so if anything it's made a product overall worse from growth standpoint.)

You're free to define what "sufficiently decentralized" means to you, of course, but how we think about it is laid out in Varun's blog post:

A social network achieves sufficient decentralization if two users can find each other and communicate, even if the rest of the network wants to prevent it. This implies that users can always reach their audience, which can only be true if developers can build many clients on the network. If only one client existed, it could stop users from communicating. Achieving this only requires three decentralized features: the ability to claim a unique username, post messages under that name, and read messages from any valid name.

You might ask: "But if most people use Warpcast, doesn't that mean it's de facto centralized?"

Depends on the perspective.

For a developer, if you are building on the protocol, then you're free to sign up your own users or convince users to move over to your app (it's easy for users to use multiple clients at the same time).

For an account with an audience, let's say you were unfairly nerfed by Warpcast, i.e. you don't appear on the app. There's definitely a lot of friction to get people to start using another client if they already don't, but the protocol itself is not restricting your ability to communicate. That's an important distinction because on web2 / centralized social networks, the app and the "protocol" (i.e. the database) is the same thing.

So if you're argument is Farcaster is not sufficiently "decentralized" (or "decentralized" if you don't like the sufficient framing) then you're defining decentralization as a protocol or network that has at least 2 independent, roughly equivalent functionality and UX clients (because majority of consumers revealed preference is UX over principles).

That's a reasonable definition.

But in practice, in the near-term, there's unlikely to be multiple, well-resourced enough clients that are full-time building on the protocol for amazing UX. That's not a knock on people building, just the reality that in 2024 consumers expect a lot of our their daily driver apps.

In the medium to long-term, dedicated individuals, small teams or open source projects can likely hit the UX bar. But that will probably take years of sustained effort.

So protocols in the early days are unlikely to hit the bar of 2 independent, roughly equivalent and functionality UX clients.

This is something I've changed my mind on over the years.

"But if Warpcast didn't have exclusivity on DCs and channels, then other clients could better compete"

Yes, but having both be interoperable / in the protocol doesn't magically mean all of that surface area gets built out in other clients.

I would assume other clients can move quicker by copying whatever patterns Warpcast has iterated to (avoiding all of our mistakes) and then in some cases, improve the UX.

But building out a robust chat UX in 2024 that's passable (not even Telegram level snappy) is a lot of work. That's effort not being spent on differentiating features or other core parts of the app. Same goes for channels.

And arguably, the most important part of a feed-based social app is the feed, which is mostly backend (machine learning, performance).

One advantage for other clients are dedicated infrastructure companies like @neynar, @openrank etc.

But it's still a lot of work.

To close: it's easy to have a bunch of opinions on how the ecosystem should be built. But if you talk to the people building the apps, they often have a different perspective having been in the salt mines.

That's not to say we are doing everything perfectly (we're not!) and we don't make mistakes. One of the reasons we are holding Developer Day next week is to get candid feedback from a group of developers working in the ecosystem daily. Goal is to get everyone better aligned and rowing in the same direction to make Farcaster successful.

1. Most people will only use social networks if they are fun

2. Casting something and getting dozens of spammy replies is not fun

3. If you get enough spammy replies, you'll just stop using the app

4. So, aggressively filtering for spam is necessary to keep your existing users happy

5. However, this means some good new users will hit spam filters. Understatement: this is a bad user experience for those people. It would be great to not have this happen.

6. But given the choice of who to make unhappy — an existing user who has invested a ton of time and energy into the network or a new user that is most likely to churn (top tier social networks lose 50% of users!), then you have to optimize for your existing users.

7. So does that mean you don’t care about new users? No! It’s existential to have new users be able to join Farcaster and find people to connect with — without being labeled as spam.

8. So here’s what we are doing to improve this so new good users aren’t mislabeled as spam:

a. Continually refine the spam model by identifying incorrectly labeled people (that feeds back into the model and makes it better); good existing users on the network reporting incorrect labels is really valuable here

b. Warnings to users before they are labeled spam so they have a chance to adjust their behavior

c. Increase the number of signals our model uses to help filter out bots vs. people

d. Working a new version of channels that give owners / moderators complete control of who appears in their channel (so a Warpcast spam label won’t affect casts in the channel)

Here's another way to look at the spam filtering problem.

1. Let's say you have a new account on the network (Account A) at they reply 100 times to the same account (Account B) with no engagement back.

2. It doesn't actually matter if Account A is run by a human or a bot with AI.

3. If you have 1000 accounts like Account B, Account A will just stop using the app. They'll move to another network or a messaging app.

4. Ideally, Account B would reply thoughtfully a few times and Account A engages with them, and then it increases over time as they build a relationship.

5. If you say "well 100 times is too much, but 50 times is fine", then you're admitting humans can be spammy and we're now arguing over the definition.

6. Side note: I don't think anyone is ready for a world where bots powered by AI are as interesting—or even more interesting—than humans.

Or "Just use proof of humanity to solve spam."

1. Spam is relative: you might not find something spammy but another person might find it spammy.

2. If you are casting in your home feed or to people who chose to follow you, then it's by definition not spammy

3. If you are casting at someone else, then their opinion does matter. It's no different than walking up to someone on the street and trying to strike up a conversation with them.

4. Fundamentally, spam is unwanted inbound that tries to get someone's attention (literally what spam email is).

5. Humans can be spammy and bots can be not spammy.

6. What proof of humanity does is limits how fast new accounts can get created, i.e. if you want to create 1000 accounts, you now have to get 1000 proof of humanity credentials. That's friction.

7. However, if you have proof of humanity, you can still be spammy (and still hook your account up to ChatGPT, etc.).

This is why digital identity is always emergent from apps with utility. If you spend a lot of time using an app / platform / OS, it starts to become your identity. Building an identity layer is the wrong order of operations. You earn the right to be an identity layer by time spent, usage, people building on it, etc.

At-scale digital identities in 2024:

• Email address — the original "killer app" of the internet; so naturally websites asked for you email address as your identity

• Phone number — as smartphones became more popular, many apps shifted from asking for an email address to a phone number. Even more people had a phone number than an email address, esp. in developing countries.

• OS-level identities — Sign in with Apple and Sign in with Google

2. This number is artificially high as there are likely many Hub operators hoping for a future economic reward. There will not be a reward.

3. Rewards don't make sense as there's no good way to prove that a Hub was honest and performant for a long period of time. Any deterministic system tied to an economic reward would create an incentive for fraud to fake running a Hub to get paid.

3. This is why Bitcoin and Ethereum do not pay node operators. Economic rewards are independent from running a node. Coinbase ran BTC / ETH nodes for years (pre-staking).

4. Our original goal was get to ~100 independent, good Hubs. Given network size, activity, number of companies building on Farcaster, seems like that number is achievable between 1) companies with an economic interest in access to the network and 2) altruistic individuals for hobbyist or ideological reasons.

a. Onchain activity — age of wallet, token balance, holding an NFT

b. Farcaster activity — age of account, # of followers, OpenRank or Power Badge

2. Remind me — useful for live virtual events like Farhouse, Unlonely, /tabletop game night; creates a frame with a single button to remind the user via a programmatic DC 1 hour or 15 minutes before event is live

3. Waitlist — useful for building excitement about a consumer app launch; creates a frame with single button to join the wailist; admin UI is a table of FIDs and relevant metrics, ability to invite users to the app via a programmatic DC

4. Prop bet — creates a frame with a description, $ amount, odds and resolver (i.e. 3rd party who decides who won the bet) for a prop bet against the caster. Transaction on Base in USDC? Degen?

5. Copy trade — creates a frame with a single button / txn that allows you to clone a portfolio (via a consolidated multi-swap txn, not sure if possible)

6. Amazon product link frame — create a frame with an Amazon product, frame creator has an affiliates account, pays you out a % in a crypto

7. Fundraiser — GoFundMe campaign in a frame. Allow the ability for the author to specify another Farcaster user, e.g. a charity, funds go to connected address, track successful payments / leaderboard for biggest contributors

8. LLM actions - improve your grammar, make the cast funnier, more succinct; generate an image with DALL-E

9. Imgflip - similar to a GIF keyboard, search for a meme template and then have the ability to add Impact text on top and bottom of the image

10. What channel should I cast in - analyze cast text and then look at what channels have been most active recently

11. Boost my cast - select your mutual follows before you cast, action waits for the next cast to propogate on the network and then DCs a link to your friends to like it

12. Grab most recent [x] - give the action your Twitter username, and it shows you the raw text of the last few tweets, making it easy to cross-post; could be the same for most recent mint, transaction, trade, etc.

13. Upload your photo / image to NFT platform - upload your image to Zora, Rodeo, Moshi, returns a frame with mint button ready to go to cast

However, a "bot" is traditionally thought of as a programmatic account, i.e. controlled by a script or software. e.g. bot net

When people on Farcaster refer to "bots" they are more often referring to actual people using a mobile device (with the aid of translation tools and ChatGPT) who post uninteresting content (from their pov).

They are casting and replying in hopes of appearing to be a "real account" because they have learned that turns into a future economic reward in crypto. Recent tipping meta has increased the expected value of these accounts.

It's possible to algorithmically label these accounts, but then you create a "shadowban" meta. So you massively increase the number of support messages and complaints. This was the state of the world before priority mode.

We're going to keep iterating here, but our pov is:

1. Priority mode works pretty well out of the box

2. We need to offer ways for actual interesting accounts to break through

3. We'll likely bring back some version of spam labeling since priority mode has created it's own set of customer support complaints.

"Why don't you just use onchain signals?"

a. Based on our waitlist last year of 300K sign ups, "bots" have plenty of Ethereum addresses with assets.

b. Most new users don't want to connect an address during sign up (friction, doesn't work well on mobile). So you'll miss out of on good accounts.

c. Active onchain presence != you regularly post interesting content. There are some people of course, but plenty of examples otherwise.

Appendix

1. Meta estimates that 10% of users on their products are potentially bots. 300 million accounts. Says it's hard a problem to solve (even for them).

2. web2 social networks can delete accounts. Farcaster cannot. Warpcast can label as spam / reduce visibility.

"Just use proof-of-humanity"
See the point about real human using a mobile device. That doesn't solve the problem.

"Just use another social network signal"
Plenty of "bot" accounts on Twitter, also now on GitHub because of Starknet.

2. Settings after often (but not always) a sign of underdeveloped product thinking

3. Most users don't want settings; they don't want to do work; it should just work out of the box; finding the right default is critical

4. Giving power users a setting means you'll get less feedback if the default isn't working, since average users don't complain, they just stop using the app

5. Power users often want very different things than average users

6. 99% of new features should be focused on the solving the existential problem. Farcaster existential issue is growing daily active users, which means figuring out the right defaults. And upstream of growing daily active users is growing the supply of interesting and entertaining content.

7. Part of prioritization is making the right investments in scalability, e.g. why we have some people working on Hubs right now

1. At-scale, world-class consumer apps in 2024 use machine learning (AI) based feeds. This is overwhelming consumer revealed preference in terms of time spent.

2. Making a good machine learning feed is hard and requires significant resources to make performant and real-time.

3. If you want to do a feed marketplace with good UX, you'd likely need to create a backend where developers would upload their models and the client runs the model in their infra. This obviously has privacy concerns, but maybe possible.

4. TBD if consumers would be willing to pay for your algo, though.

5. The likely outcome is a marketplace of clients, where each have a different algorithm that uses inputs (both public from the protocol and private from app usage) to tailor a model for each user.

6. PS — here are two data points from a developer who was building a Farcaster client: 1) feed performance is a challenging engineering problem and 2) running the backend is expensive (the client had a few hundred active users).

Farcaster

Kartik on Warpcast

1. building a mobile app, we rebuilt it 3x before we grokked it 2. feed latency/performance, much harder problem than we initially realized, hence scrapping custom feeds beyond lists

Farcaster

Kartik on Warpcast

~$2k/month

1. Awareness

2. Onboarding

3. Retention

4. Interesting and entertaining content

Awareness
How do people know about your app?

Also known as distribution. This is hard, for people who are naturally more building-oriented. It's almost always hand to hand combat early on. Good essay on the topic:

https://blakemasters.tumblr.com/post/22405055017/peter-thiels-cs183-startup-class-9-notes-essay

Onboarding
How much friction to get started?

Congrats, you got someone to hear about your app. How easy is it to go from hearing about it to getting set up? How many steps? Can they do it on their phone, i.e. immediately? What is the first screen they see after a successful sign up? How are you building habit?

Retention
Does the person come back?

They've downloaded the app, signed up and used it for 2 minutes. They close the app. For most apps, 50% of people never come back. Do you have their email? Do they push notifications enabled? Did you think a web-first app without any contact method is going to retain?

Interesting and entertaining content
Your app's energy source today and every day, forever.

Is it getting less interesting? People will just switch over to YouTube, TikTok, Instagram or Twitter for an endless stream of algorithmic content. (People love algo feeds despite stated preference.) If you have a growing and sufficiently large pool of interesting and entertaining content, however:

• Retention — people will stick around for the content

• Onboarding — they will jump through rings of fire for it

• Awareness — they will hear about it from friends

• Your app — can be rough around the edges

So one big advantage of building on Farcaster is if the network continues to grow the number of interesting and entertaining people and channels—esp. outside of crypto—then your app starts with interesting and entertaining content for free.

1. Waitlist frame — make it easy to sign up for a product launch; send a direct cast then the product is launching or allowing batches of users off the waitlist

2. Airdrop to everyone in a thread action — for a given cast, extract the connected addresses on Ethereum or Solana for all replies engaged with by host, or simply everyone who liked the cast

3. Send a direct cast to everyone who liked a cast action

4. Follow everyone who replied to a cast and engaged with by the cast author, e.g. in an AMA

5. Follow Friday action — @alice starts a follow Friday cast and then anyone she is following can submit 3 people to follow; 1 click follow all of the recommended people (spammers won’t be followed by @alice so no need to worry about that)

6. Tinder PWA or simple mobile app for finding new users to follow

7. Top 25 most liked casts from people you follow in a daily digest newsletter, send via direct cast

8. Any time a cast >50 likes from people you follow, send as a direct cast

9. NFT pre-release frame + direct cast — similar to the waitlist idea, have a way to get notified about a new NFT drop before public launch

10. Summary of new people with power badge every Tuesday morning sent via direct cast

High-level thoughts

• Finding good items/shops is hard — there are 3 ways to find items on the OpenBazaar client: 1) search for #tags 2) whatever stores you follow and 3) the Random tab.

• Slow UX with learning curve (slightly off from traditional commerce design patterns)

• Feels more Yahoo! (directory) and less Google (search engine)

• Mainstream adoption question: compare friction of opening up a new tab vs. open up a new app; lack of a quality mobile client

What is for sale?

Physical goods

Harmless

• Legos

• Organic, small-batch food (e.g. honey, hot sauce, spices)

• Bitcoin-themed jewelry, clothing

• Electronics, hard drives, GoPro, Raspberry Pi

• Semi-precious stones

• 2nd-hand clothing, shoes

• Custom 3D printing

• Old physical media: console games (Xbox, SNES), vinyl records

Gray market

• Sex toys

• Precious metals

• Gift cards

• Adult content

• Fake watches

• Generic but regulated pharmaceuticals

• Guns, weapons

Illegal

• Drugs

Digital goods

• Steam accounts/video game skins/game key (CS:GO)

• Ethereum/Bitcoin (a la Local Bitcoins)

• Pirated software (e.g. MS Office)

Services

• Sysadmin services

• OpenBazaar 24/7 hosting

• OpenBazaar consulting services

• Q&A session, therapy phone call

Ecosystem

• Hosting — If you download the official OpenBazaar client, the OpenBazaar shop is only online if the client is running. There’s currently also no way to determine if a given shop is online or offline. For more serious shops, a set of OpenBazaar VPS hosting services are available ranging from $10/mo to $100/mo. This involves running the OpenBazaar Server.

• Moderation — When a given OpenBazaar transaction between a buyer and seller is disputed, the protocol supports the concept of moderation. Moderators advertise via the “Discover” tab or reddit.com/r/OpenBazaarModerators.

Opportunities

• Search engine / curation

• Alternative desktop/web client

  • Assumes you want to sell (problem with eBay as well)

  • Orient around search terms, not stores

    • e.g. with both Google and Amazon you search for products across the web/within a massive store (that has marketplace sellers)

  • Early sites are OK—but opportunity for well-designed

• Built-in on-ramp into bitcoin for consumers

  • Alternative: fiat-to-bitcoin translation layer (never deal with bitcoin directly)

• Additional merchant tools

  • Instant Exchange

  • Digital goods infrastructure (a la Gumroad; IPFS?)

  • Inventory management API (to sync with fiat store on Shopify, Amazon, Squarespace, etc.)

• Mobile app

• Messaging and simple bot-like commands to complete transactions

• Shipping — Given a shop owner can, in theory, operate anonymously and shoppers can purchase anonymously, it seems odd that a real-world physical address is required in order to complete physical good transactions. cf. RPO and the concept of “re-shipping”

• Payment processing — While bitcoin is best tool for a decentralized commerce network, bitcoin is not the most convenient option for most consumers. This is esp. relevant as more OpenBazaar stores will be less oriented toward grey market and simply leveraging the lower fee environment/distribution OpenBazaar provides. Related: And while OpenBazaar’s native support of bitcoin is great, many businesses will want the ability to instantly convert any transaction into fiat to minimize currency volatility risk. Currently there is no default liquidity provider for OpenBazaar.

I woke up this morning and saw a cast from Tarun on SBF:

Tarun Chitra on Warpcast

The last 12h of Twitter takes on SBF has been a mixture of misplaced elation, investors unable to grapple with personal responsibility, and dunks from people who made money from Alameda; perhaps trauma is healed this way, but I respect the opinions of those who were quieter much more

To which I responded:

Dan Romero on Warpcast

Not an SBF fan, but always weird when decades in prison is something people celebrate. Maybe if you’re a direct victim, but peanut gallery is cringe. Legal system worked. Let’s move on.

And then further:

Dan Romero on Warpcast

The overlap of the “Free Ross!” and “haha SBF going to jail for decades!” crowd is a strange one.

To which he tweeted:

Tarun Chitra

@tarunchitra

Amazing take from @dwr on Farcaster

103

8:12 AM • Nov 3, 2023

And there was a bunch of discussion.

Don't get me wrong, I'm a guilty as anyone in engaging in schadenfreude, but as I've gotten older I've found myself more often reflecting on what it would be like to be in the other person's shoes. Not condoning behavior, but what it would physically feel like to have the fear and uncertainty of looming decades in a prison cell. You sort of lose the appetite to take the cheap shot (or at least your direct at a more irreverent situation). No one is forcing you to tweet / cast the hot take.

Any human being—no matter how awful and deserving—spending decades in a prison cell is tragic. Necessary? Likely. Harsh reality? Yes. But still tragic on a human level.

P.S. — when I went to start writing this, this cast was at the top of my feed.

nixo on Warpcast

Very unpopular opinion: I don't celebrate anyone being thrown into the US's for-profit, inhumane prison system, no matter the crime. I welcome them being off the streets, but any human being put in the US prison system is a travesty criminals should be rehabilitated, not made worse w mental distress and dehumanization