@kiptobuilding a better future with you @masknetwork @lens and @orb

happy weekend fam.here are some of my private unfiltered bouncing thoughts. grab a coffee or glass of wine or beer before you start. this got way too long, so i plan to shoot this out monthly from now on and also we plan to double down on comms anyways. just dm or tag me if anything anytimewe're in another phase of change. we've been shipping 24/7 since 2023, but back in march i saw how fast ai was actually moving, and the past few weeks have been another jump. the claude 5 family, gpt 5.6, local models have become new meta. and the interfaces are moving as fast as the models now. my favorites are herdr and clicky, but none of it looks like what we were using six months ago. even cursor has agent ui.future is blurry but bright. we're still in the c64 era of this. it works, it just doesn't look anything like what it's going to look like, and nobody has written the killer app yet.catch is that being right about something doesn't make it cheap. spacex ipo is the clean example. like it or not so is $eth. once everyone agrees a thing is inevitable, you're not buying the thing anymore, you're buying the expectation, and the expectation is expensive with multiples baked in. that isn't a reason to be bearish though. sometimes the easy things take forever and the far away dreams arrive early. decentralized onchain social is no different. Bulls always win in the long run because innovation rarely goes backwards, see S&P 500 over any twenty year window, but ofc the trajectory absolutely gets corrected on the way so timing matters.those of us still here are here for a specific reason, and it isn't reach or feeds. I believe it's that you own the thing you post and it survives the app you posted it from and good might come from posting it, now or in future. and that's what gets us super interested and it's what we build for.which means going back to first principles on nearly everything, because almost every pattern social media taught us assumes the opposite.so: what have we been doing this summer besides the world cup stuff, where is lens going, and where is onchain social going as a whole and what’s next.no more silosdays of rebuilding the same project on another chain are over, at least for now and always for me.for me the promise of crypto and open protocols has always been that A builds something B can use, or build on top of. but love it or hate it people got greedy and started gatekeeping their own property. been there done that. so like it or not crypto is still a walled garden that markets itself as openness.uniswap v4 is the honest example, and i love uniswap but the core of it is licensed so nobody can fork it until june 2027, and then it opens up. so it isn't closed forever, it's closed on a timer. but a timer is still a moat, and everyone here is trying to build one. lens had this problem too. so do the data and bridge providers.everybody wants to get paid, and that's fair. it's also the inherent problem of open protocols. what is the thing that makes your open source project better than the fork? there's a reason ai and pharma are closed. if there's no way to earn back the capital deployed, who invests besides philanthropy.the answer i keep landing on is interoperability. i'm still optimistic. ai centralizes at the very top, where training a model costs more than most countries gdp. but inference does the opposite and levels the whole field fast and also extrapolates to other things: why pay for software when it's now trivial to build the thing yourself. ngl but we’re nowadays building our own software as well as being able to cut costs dramatically.**where social has been stuck
**
four real problems. no point pretending otherwise.the public post is dead. people still spend the time, they just spend it in group chats and dms instead of broadcasting to everyone. what's still interesting to me: ownership, privacy, friends, realtime, connections. not reach.ui is dying. ai is killing the interface and rewriting the ux underneath it. will there be apps other than messaging apps in five years? if there are, you probably assembled your own setup for them. i get all my news on imessage from sources i picked myself. hackernews, crypto, celeb gossip, hollywood breakups, uk royalty drama, daily. it's fun, and no app built it for me.both algo AND following is broken. orb's foryou is good but algo content is so 2020 and we're all here for following, BUT following only works if you're in the app 24/7. miss a day and you missed the best post and the best comment. low volume is saving us right now. that doesn't scale.overselling. the reason i started building web3 social in 2023 still hasn't fully arrived. content still gets lost. apps still build walled gardens. users still aren't in charge of their data. it isn't fully onchain meaning enforceable trustlessly, and you can't rebuild it yourself if somebody drops it. privacy barely exists and anonymity even less.what is actually great on orb: one click makes a profile, and on a clean phone you can create a pseudonymous profile that leaves no trace on our side. that's real and still so rare.yet it does get harder the moment you want to fund that profile and start tipping. that works today, and the lens bridge makes it a lot easier.why things take so longpart of it is that this is frontier work in several directions at once. gas is far cheaper than it was but still blocks things. building something complete is the harder blocker. there's always a compromise.obfuscation is a good example. vitalik wrote about it recently and called it the final boss of cryptography. we know how to build it. the versions we know how to build would take longer to run than the universe has been around. so it isn't unsolved exactly. it's just unusable, which for our purposes is the same thing.
verifiability is the trust layer
the thing that did change is formal verification, and what changed it is ai. that just means proving mathematically that a piece of code does exactly what you say it does, no more and no less. it used to be far too expensive to bother with unless you had a research budget. now it isn't.that matters more than it sounds, because trust is the actual bottleneck in both of the things i care about and think we need to double down.nobody gets scammed by a good argument. you get scammed by a dust account that looks real enough for three seconds. or you trust that the platform you use is safe or returns you the correct data. every anti spam system we have is a guess about who somebody probably is, and guesses get gamed the moment there's money on the other side. if the account can prove what it is instead, the guess goes away and most of the problem goes with it.then ai, which has the same hole but newer. buying a prompt, a flow or an agent today is pure vibes. would you even consider buying one? you're trusting a screenshot from someone calling themselves a prompt engineer. there's no way to check the thing does what the listing says before you pay, and nothing you can do after. that isn't a moral failure of the people selling. the tool to check just doesn't exist yet.that's the tool. if i can hand you proven guarantees instead of promises, you get the real limits out of the box. no more overselling, from us or from anybody else. you can show exactly what's possible and under exactly what constraints. trust stops being a thing you extend on faith and becomes a thing you check. eg orb has device wallet. what does that do? what are the limits we can guarantee it does without you having to go through the source code and even worse is still is private…So we talked a lot about everything else than social BUT that's the future of social, and it's also the future of working with ai.the walkaway testwhy build it the hard way?because current social media doesn't have what it takes to survive bots and infinite everything else. the rules have to be baked in at a level that stays enforceable after the app dies. that's the walkaway test and we want to apply it to everything.this is what first principles actually costs. you don't get to inherit the shortcuts, so a lot of it takes longer and looks slower from outside.grove is our storage layer: your content lives offchain (teaser soon maybe onchain), and right now the permissions over it live on chain. we're introducing some ideas for where grove might go next, such as. you host your own data, and at minimum you can detect when your copy isn't complete and do something about it. and what about that eu ai act's article 50 mandate requiring machine-readable labeling and watermarking for ai generated media.a side note that fits here. when lens migrated from polygon to lens chain, collects didn't come along. lens said at the time that it decided not to migrate collects and actions, keeping their value on polygon, and that the new system lets people turn these features back on as needed. people read that as something the app or protocol took from them. what changed since is that a port like that is now something a person can build in an afternoon with ai, and last year it wasn't.that's the actual shift. with ai and a bit of money, you can take ownership back yourself. posting from an app that forces every reply to be paid? say no thank you, use your own tooling, edit your own posts. it's easy now, if you want it.**the canonical bridge is live
**
lens went all in early on speed and forgot that people need a way to move money in and out without asking anyone. that's fixed. the canonical routes (ETH, USDC, GHO) are now at lens.xyz/bridge.privacyencryption lands in august. the first thing it unlocks is private groups on lens, where the group is onchain but what you say in it isn't. pay to unlock works the same way. orb will also a pay to reveal incognito mode on top of it. we invite all the apps to build on top of this really easy and cool encryption logic.there's more to come, and it ships over the following weeks.what social looks like in 3-5 yearsit could be pseudonymous-first hybrid (human-ai) decentralized social but if we're not building it, who is?right now the SOTA is bluesky but first a correction i see repeated everywhere, including in my own thinking. bluesky isn't peer to peer. it's federated, which means a set of independent servers passing things between each other, with your data sitting on whichever one you signed up with. bluesky's lead protocol engineer wrote a whole post on why they ruled peer to peer out. nostr isn't peer to peer either, it's apps publishing to independent relays. neither is what people think it is, and they aren't the same thing as each other.the distinction matters because federation doesn't solve the problem people think it solves. you're still trusting whoever runs your server, and there's no shared source of truth, just a lot of copies that mostly agree. you either trust your neighbours or you spend time you don't have reaching consensus. when the truth is onchain, all of that gets much easier. same reason onchain payments beat p2p payments. social gets there too, with lens primitives.second, the shape is changing. block shipped buzz on july 21, an open source workspace built on nostr and aimed at slack and github. the interesting part isn't the workspace. it's that the ai agents in it have their own identity and sign their own work, so you can always tell what a person did and what an agent did. that’s super interesting direction. we might have something similar coming very soon.the future of social isn’t what it is today. maybe the future of social is hybrid. maybe parts of it are single player. you, your data, your agent, and the network only when you actually want the network. that sounds strange right now. i don't think it will in three years. but at the same time we need to solve to challenge for loneliness.**orb, rebuilt
**
orb is about to see a lot of experiments. we've rebuilt the app. the ui looks the same, the logic underneath is completely different. The app will be split to two things: smth we call “core” and “apps”. we’re still working on the notion and everything is fluid. maybe one day dark mode in your own private orb is realistic. i believe and am not alone but we need to put more things onchain and also focus more on fast extraction using just an RPC. what that might look like: you can set money aside for your own fees and nobody else can sensor you, no app can hold your posting hostage. you should be able to talk to the chain directly and skip servers entirely. don't like what's in your feed or you cannot do the action you want to do? build your own feed, or use somebody else's, or build one for other people and sell it. none of that is possible when the feed is a black box with knobs on like on X.**growth might not be the meta
**
onchain social has a lot to offer and that's why we started here. but building social is hard. you need enough people that it's worth showing up, and enough money to keep going.the old meta was growth. get every user to bring in more than one more user, hold that, you win. in 2026 that's either very hard or the new users are mostly bots or people farming incentives, which brings you no closer to a product anyone wants.the world cup showed both sides clearly. firefly and orb both saw the farmers grinding the leaderboard. however we also saw a lot of users who made real money on the prediction market trading competition, because they were right about something. to me that second group is the one that matters.so here's my bet: the best new users are the ones who make it. who either get rich, have fun, learn something, make a real contact or a real relationship. we're doubling down on them. orb is adding loops for trading and prediction and making things work together. we all can win here.. we're also pushing on user owned marketplaces and irl through orb. so selling things and meeting people. the options are wide open, as long as we can guarantee it's safe to sign up, get money in, invite a friend and everything after that actually works.get that right and orb is the obvious place for you and for your ai agent to spend time.what's coming to orbwe built a lot of the building blocks early. replacing reactions with stickers was one of them, and stickers are coming back in a new way. groupchats and stories were briefly live during worldcup and both return soon. we've tested video chat and that's coming back too.the most interesting thing for me personally will be that i don’t need other app to buy tokens or do preds. orb only mode.everything in orb will work like everything in lens. ready to use, and ready for other apps to build on. dms are openand if you're building something cool with ai, say an online shop for creators, this is where the verifiability part pays off. you sell the prompt or the flow on lens, and the buyer gets a checkable guarantee about what they're buying instead of your word for it. formal verification and encryption doing the work together. Sounds interesting.
none of this is the full list. there's more in the pipeline than fits in a post, and a lot of what we're doing right now isn't new features at all. it's making what already exists harder to break, and nailing down exactly what it promises. that part is deliberate. a bear market is when you build the boring things properly, so that when the next era starts you're shipping instead of rewriting. that's where we are.how we actually go forwardduring the world cup we got invaluable normie feedback and we're working through it. if things don't work 100%, or the user doesn't understand what the app is for, there's no second session. like it or not that's big part of the game.if we're going to grow, we go where the user already is instead of building in a silo. grants mostly fund the proposal rather than the product, so they don't work but hackathons can. investment can. and if crypto is going to make it, it has to move as a community in one direction.that goes past crypto. the world is on the verge of a change most people can feel and nobody can time. ai is rewriting what software costs to make while the world hasn’t been this unstable in a long time. social is getting rebuilt whether or not we're the ones doing it. p2p founders are getting arrest warrants like TG’s durov. these moments need good ideas. so build with us, build next to us, or build something that just talks to what we made. all three work and we'll help with all three. buy ai, do your thing.finallyall of the above is mine and mine only. some of it will ship, some of it might not. and none of it is especially original, great minds think alike, so if you got to the same place from a different direction then you're probably not wrong either.what it all comes back to is first principles. not what a social app is supposed to look like, but what has to be true for any of this to be worth doing. you own what you make. the rules outlive the app. trust is checkable instead of extended on faith. that's the whole list. everything else is negotiable and most of it still is.so the more useful question isn't what we're building. it's what you're trying to build and what's standing in the way. lens will be the flagship for decentralized social, which means it isn't ours to steer. tell us what would make your thing work better and we'll build toward that. together and for the people, or there's no point to any of it.and then the bigger one. how much are you willing to bet on a decentralized future that's actually good for humanity, in a world this uncertain and this full of gatekeeping?pussy riot were one of the first on lens. my dream is to welcome them and groups like them back one day, to do their own thing, on their own terms.

@0xjavisoftware developer, ai engineer, looking for a team to join

new on LoE:🐮 more animals to hunt!
(i think the cow is really cute, and near lakes will be some geese🪿)🌉 now villagers can build bridges to cross lakesthank u for the requests @elonmusk-x

@0xjavisoftware developer, ai engineer, looking for a team to join

🔨 new character on Lens of Empires 🔨we use to have "elondrunk" ( @elonmusk-x as a peaky blinder, throwing whiskey bottles)but now it's a nordic god of thunder, with a magical hammer that throws ⚡️⚡️s, named Thunderbeard!he always has his eyes closed 😑🧔🏼‍♂️ he can attack without even looking 🫣🎯🗼⚡️ the best thing of Thunderbeard is his range: 7 tiles! towers has 6-tile range, so Thunderbeard is perfect for destroying towers with his magical hammer!go play with it at https://empires.lensie.xyz !!

@artemproduct/design @orb

test

@datartistArtist

Trust me, guys... Orb is building.You have no idea. 😅