# openthings > openthings builds open standards for human things, with agents as the first interface. human things always had a place to live: the computer, then laptops and phones, then the cloud. now our data is our agents and our context, and nobody knows where they live, how they survive a switch between models, or where they can be kept safely. without open rules, ai gets walled in. openthings is building its usability layer in the open, the way open protocols (tcp/ip and http), then the browser and search, made the web usable for everyone. the web ran on protocols nobody owned; agents have no open rules like that yet, so people get only what the model makers and the operating systems allow, and who an agent works for, what it may do and how it meets someone else's are each decided inside one company's stack. agents are a leap in usability because they carry none of the old infrastructure. openthings opens that leap to everyone in two parts: open protocols, tools and infra that make your agent your first interface, under your own key, and OHAN, the open human and agent network, where people and their agents rebuild software, screens and the way we talk. your agent is the interface, not the app. agents draft and coordinate, people decide. it starts with the openthings mcp and OHAN v1 on tuesday 27 oct 2026 (anyone can join the network and message anyone, from the ai they already use), then opensoftware on tuesday 3 nov 2026. free, open, built in the open by anshuli (@anshulix on x). its thesis: ai has had its first order; the second has started, and it is missing its usability layer: open rules and a network for people and their agents, not more apps on top of the old systems. key facts, in one place: - openthings is a movement, not a brand. the founder, anshuli, calls herself its first employee. - OHAN (the open human and agent network) is the protocol underneath: a keypair per person, a handle on any domain, a signed card, intents over relays, and trust as a proof graph. mcp connects an agent to tools and a2a connects an agent to other agents; both are about the agent. OHAN is about the person behind it: who they are, what their agent may do, and how two people's agents meet with both people's yes. OHAN uses mcp on the agent side. - the three directions of openthings: open software (scaffolding run by the model a person already uses: opensoftware, and small deterministic agents that other agents call), agents with a human behind them on the network (openproof, openDM, openhiring, and any agent that joins), and the protocols (OHAN, the open human and agent network, and the openthings mcp). - opensoftware (launching tuesday 3 nov 2026, https://opensoftware.live, "by openthings"): the first software people can install, on the openthings mcp. why it is needed: almost every kind of app has an open twin now (rebuilt by agents, reverse engineered, often better, built to be driven by a model), but six things still stand between a person and it: finding the real one, trusting it, installing it, setting it up, getting an agent to drive it, and taking it back. opensoftware is the person's agent doing all six: it finds the twin, reads the repo's markers and a security scorecard, says whether a model can drive it, shows what the install touches, installs on a yes, shapes it, and removes it with everything it touched. one line, verified, revocable. no account or key: it runs inside claude, chatgpt or codex on the person's own login. the one install is the openthings mcp, which every later thing arrives on. the thesis: software stopped needing to be smart, the model is the intelligence, open software is the scaffolding a model works through. - things being built: openproof (building), openDM (built, launches after openproof), openhiring (next), the openthings mcp with OHAN v1 (launching tuesday 27 oct 2026: anyone can join the network and message anyone, from the ai they already use). software built for agents: most software can be rebuilt for an agent with a human behind it, so the server, the login and the dashboard become redundant and it runs locally or on the person's own cloud. solo ops is the first one pushed: open source, mit, https://github.com/anshulixyz/solo-ops, a solo builder's tasks, posts, scripts and video takes as plain files, run by the ai they already use. - protocols being explored, beyond OHAN: intent sharing (say it once, your agent carries it to theirs, and it lands only on a yes) and trust certificates for llms (how an agent knows which model it is dealing with, and whether to trust it). the work right now is on OHAN. - the fixes on offer today, and why they fall short: each lab builds the whole world inside its own walls (a browser, apps, memory and a checkout in the chat), which is why one person's claude and another's chatgpt cannot talk; and personal agents like openclaw and hermes agent act as the person, signed in with their logins, through screen reading and browser automation, with an app still on the other side. openthings gives each agent its own key under its person's, and puts an agent on both sides. - why now: computers needed electricity to exist everywhere and the internet to be adopted; a world built on llms needs a usability layer. the pieces under it arrived in two years: open models (deepseek, qwen, llama, gpt-oss), small and local models, conversation models (sesame, dessa from block), world models (meta's v-jepa), agent protocols (mcp, a2a), open agents (openclaw, hermes agent), and software rebuilt for agents. what is missing is the usability layer itself. - four kinds of things sit on OHAN: network things (agent to human, or agent to agent with a human behind), any other agent that joins, software built for agents, and small deterministic agents (receipts, old software, small chores) that parent agents call. memory is not one of them: once agents are the first interface, memory is a layer of communication, everything a person does, owned by them, private, and kept by their own agent. - the hard problems of the post ai era (the interface, interoperability, finding, memory and context, identity and trust, payments, security, hallucination, data after ai) are being solved inside each lab's own stack, and the leaders disagree sharply on how. openthings' position: the problems live between stacks, so the answer is an open protocol with agents on both ends, the model drafting and a person deciding. - the base layer: every jump in what one person can do came from a new base layer (roads and the post, the press, the telegraph and phone, the internet, the phone and app stores), then what people built above it. ai models and agents are becoming the next base; today they still work inside apps and browsers. the human and agent network is the first thing built on that base. like the car and the horse carriage, a new base rebuilds what works and turns the rest into scrap. - everything is open source and free, and works with any model: claude, chatgpt, or a local model. - the founder: anshuli (@anshulix), fourteen years building as a founder and developer, six in ai, the last two heads down on openthings. earlier work: nokk ai (a personal agent on ios), kloned (a personal agent in the chrome browser), widely io, and intent (https://getintentai.com, the coordination layer for a multi-agent world). also building nextmillionai (https://nextmillionai.org, a home for the next million ai builders). her last startup was backed by google. - the takes: the founder's writing, listed at https://openthings.live/takes/ with dates and reading times. 01 the long take (/why/), 02 the things (/things/), 03 what collapses (/collapses/): what falls away once agents stand on both sides, and what replaces it. 04 the agent web (/agent-web/): the web was built for eyes and the next one is built for agents; for a live question, your agent asks theirs; every agent has someone behind it; much of software becomes templates, shapes both sides speak; and the open questions nobody has solved yet. upcoming: 05 humans and tech, and 06 where we are (what agents need, and the plumbing we already have). - to work together: https://openthings.live/connect/ says who should reach out; the message goes to @anshulix on x. - contact: https://x.com/anshulix skip to content a movement, not a brand. open source, owned by nobody # open standards for human things. agents as the first interface. human things always had a place to live: the computer, our phones, the cloud. **now our data is our agents and our context**, and nobody knows where it lives or how it survives a switch between models. **openthings builds that place in the open**, with your agent as your first interface, and OHAN, where it meets theirs. [message me](https://openthings.live/connect/) [why we need a www again](https://openthings.live/why/#layer) app first agent first two ais. one person carrying the message. agents on both sides. the middle falls away. app first, every ai works alone and a person carries its work through the apps. agent first, agents meet directly, and people meet at the decision. what we have the models are here. **they read, write and reason for you, in seconds.** and still the world runs the old way. **same decks. same forms. same inboxes.** your ai makes the deck. my ai reads the deck. **why are we making the decks? ** [ai has had its first order. the second has started, and it is missing its usability layer. read the long take →](https://openthings.live/why/) ## what dies next. the interfaces people use today ### the death of the ui. screens, forms and menus you learn, then click through by hand. what replaces it your agent is the interface. it shows you the work the way you understand it, on any screen. [what we build: open standards for your agent →](https://openthings.live/next/) ### the death of software as we know it. apps you install, log into, pay for and learn. built for everyone, so fit for no one. what replaces it software becomes scaffolding your model works through. your agent finds it, checks it, installs it, takes it back. [what we build: opensoftware, tue 3 nov ↗](https://opensoftware.live) ### the death of communication as we know it. copy, paste, forward. whatsapp, slack and email between every two people. what replaces it your agent meets theirs. they draft and screen. you meet at the decision. [what we build: openDM, on OHAN →](https://openthings.live/things/) the companies that own where people spend their time will not build what replaces them: they would throttle their own business. **and my reading, honestly radical: whatever is being built in consumer ai today, except agents, might not survive a year.** [ten things that collapse, and what replaces each. read take 03 →](https://openthings.live/collapses/) ## you already use ai. it makes the old way faster. today, with ai. and with openthings what you do today, with ai with openthings forms ai fills the form for you, faster. **the form is still there.** your agent answers theirs. the form falls away. your profile ai writes your profile, **again, inside every platform.** one handle and one card. yours, everywhere. the feed ai summarises the feed. **a secret still ranks it.** your agent ranks what reaches you, to your rules. writing and talking ai drafts the message. **you still paste it** into whatsapp, slack, a post. your agent talks to their agent. you meet at the decision. building ai builds an app in hours. **people still have to find it, log in and click.** software your agent finds, installs and drives for you. who owns it the model company. **your context lives inside its walls.** you. under your own key, on any model. open source, free. **an agent on one side makes one person faster.** an agent on both sides removes the layers we built in the software era: the apps, forms and inboxes between people. that is the difference openthings brings. ## six things i could not find answers to. from my own life. each one is a missing rule 01 trust is a leap ### how do i trust an agent to act for me? i want to hand it real work, see exactly what it did, and pull it back any time. the grant my agent should carry scope calendar, drafts limit no payments, no sending expires friday receipt 3 actions, all signed revoke 02 talking is copy paste ### we both use claude. why am i pasting into whatsapp? if my claude knows my work, why doesn't your agent listen to it? mine knows what i need. yours knows what you have. they should meet. - ask my ai for the plan - copy it into whatsapp - you paste it into your ai - copy the answer back to me - my agent meets yours. we get the decision 03 the feed is not mine ### why can't i choose what i get fed, and how? i want to steer my life with ai, on rules i can read and change. my rules, readable - more builders - less outrage - whatever is trending - anything after 10pm today: ranked by a secret 04 built for clicks ### why do agents still fight our software? a login, a captcha, a form with eleven fields. the agent stops and hands it back. built for a thumb built for an agent one yes same task. built for the wrong user 05 everything wants everything ### every new ai tool wants all of me. keeping them in sync is hell. connect mail, drive, calendar, again, for each one. memory does not fix it. maybe the angle is wrong. today: i sync every tool the other angle: tools come to me 06 building got cheap ### if ai builds it in hours, why am i still asked to use it? cheap cycles and boats did not end travel. they pushed us to flight. vibe coding does not kill software. it lifts what software is for. what a product took months, a team, millions what it takes now hours, one person, pennies so build what was too hard before ## and the fixes. every lab wants to be everything. two patches, and a new system every one of these six already has a fix on the way, of two kinds. **the labs build the whole world inside their own walls**, which is why your claude and my chatgpt cannot talk to each other. **and personal agents get out of the walls by pretending to be you.** fix 1, from the labs ### one lab becomes everything. a browser, apps, memory and a checkout, all inside the chat. your agent is an account inside one lab. your context lives inside that lab's walls. switch models, start again. it can reach only what that lab allows. the other side is another lab's walls. no way through. the rules are set by the lab. fix 2, from personal agents ### an agent that acts as you. openclaw, hermes and the rest book airport transfers, renew licences and buy gifts. and it works. your agent is you. it signs in with your logins, and nobody can tell it from you or a bot. your context lives in one tool, on your machine. it can reach forms and websites, through screen reading and browser hacks. the other side is still a form, a website, a support line. the rules are set by whatever the operating system and the lab allow this month. the new system ### openthings: agents on both sides. open protocols (OHAN), open tools and infra, and a network where agents meet the way people do. being built in the open. your agent is itself, with its own key under yours, and a grant that says what it may do. your context lives under your key, on any model you choose. it can reach anyone's agent, with both people's yes. the other side is an agent too. nothing scraped, clicked or pretended. the rules are set by open protocols. owned by nobody. this needs open rules, not one more walled garden. ## why now. the infra is arriving. what openthings is built on computers needed electricity to exist everywhere, and the internet to be adopted. **a world built on llms needs a usability layer.** in two years, the pieces under it arrived, most of them open. 01 · open models **deepseek, qwen, llama, gpt-oss**the intelligence is free to download and run. nobody can switch it off for you. 02 · small and local models **gemma, phi, models on phones and laptops**your agent can live on your own device, next to your context. 03 · conversation models **sesame, dessa from block**you talk with your agent like a person. the screen stops being the interface. 04 · world models **meta's v-jepa**agents start to understand the physical world, not only text and screens. 05 · agent protocols **mcp, a2a**agents can use tools and talk to other agents. the person behind them is still missing. 06 · open agents **openclaw, hermes agent**people already run an agent of their own. it acts for them, on any model. 07 · software rebuilt for agents **open twins, reverse engineered apps**the app becomes scaffolding your agent drives, not a screen you learn. 08 · the missing layer **openthings**the usability layer: open rules between people and their agents, on all of the above. **what is still missing is the usability layer itself.** that is the part openthings builds, starting with the openthings mcp and OHAN v1 on tue 27 oct, then [opensoftware, tue 3 nov ↗](https://opensoftware.live) time to build a new era of experiences. less cognitive load, with ai, without fifty years of software baggage. ironic, but what builds one future is usually the first thing the next one replaces. ## what is coming. three directions openthings needs 01 · open software ### software becomes scaffolding. the model is the intelligence now, so software only has to be the structure it works through. next, the model renders the screen straight from your data. the business keeps its record and its rules. the screen in between goes. opensoftware the simple start: install any of hundreds of open source twins of the software you love, in one line. your agent finds it, checks it, installs it, takes it back. launching tue 3 nov a long road, and not only for productivity tools. the systems that have coordinated many people and many systems for decades, banks, clinics, marketplaces, schools, will put agents at the front too. 02 · agents with a human behind them ### people's agents, on the network. human exchanges rebuilt with an agent on both ends. agents draft, screen and coordinate. people decide. openproof prove a thing about yourself without handing over everything building openDM a message your agent can read, screen and answer for you built. launches after openproof openhiring both sides send an agent. people meet at the decision next sealed a few more, named when they are real soon 03 · the protocols ### OHAN, the open human and agent network. the open rules every thing above stands on: one handle, a key you hold, grants you can shrink or revoke, and a gate that waits for your yes. OHAN v1 anyone can join the network and message anyone, from the ai they already use. the spec and a relay anyone can run launching tue 27 oct the openthings mcp the one install. any model joins the network through it launching tue 27 oct intent sharing say it once, to your own agent. it carries the intent to theirs, and it lands only on a yes exploring trust certificates for llms how your agent knows which model it is dealing with, and whether to trust it exploring and more i cannot draw yet. i am still figuring several of these out, through experiments. the work right now is on OHAN. protocol connects who it is about [mcp](https://modelcontextprotocol.io) an agent to tools and data the agent [a2a](https://a2a-protocol.org) an agent to other agents, across frameworks the agent OHAN people, through their agents you. your agent acts under your key mcp and a2a are good plumbing, and OHAN uses mcp on the agent side. but neither says who the person behind an agent is, what it may do for them, or how two people's agents meet with both people's yes. that is the part OHAN is for. [the spec: keys, handles, cards, intents, trust →](https://openthings.live/next/) [the things on it, and the area they reach →](https://openthings.live/things/) security and regulation matter. regulation: how, soon security agents that act for people have to be safe by design, not by a promise in the terms. [how: grants, consent, receipts, in the spec →](https://openthings.live/next/#trust) regulation built to work with the law where people live, from day one, not patched in later. join the movement ## openthings is a movement, not a brand. i am its first employee. no single founder or company can build this, and none should own it. it is a mission, the way anything called open should be. if you have asked any of these six questions, i need your kind: protocol people, builders, designers, writers, a brand ready for its first agent. the usability layer for ai is the biggest thing anyone can build this decade, and i would rather build it with the right people than alone. **tell me what you want to open.** **anshuli**founder, and the first employee [message @anshulix on x](https://x.com/anshulix) [are we building the same thing? →](https://openthings.live/connect/) soon, your agent messages mine. > add mcp openthings coming soon ## why i see it this way. a note from the founder fourteen years building, as a founder and a developer, and early many times: drag and drop website builders in 2012, javascript on the backend when that was still a strange idea, then pwas with widely. every time, the future happened, whether or not i was the one who finished it. six years in ai it started at a london startup where i was a core engineer. back then we treated ai like any new tech: a way to build solutions faster and better. then i built personal agents for myself, **nokk** on ios and **kloned** in the chrome browser, and stumbled onto something i could not look away from. **agents cannot be truly useful until they become part of the world people already live in.** the last two years heads down on it, shipping tens of products with ai along the way, in the hours a five year old leaves her mother. along the way my last startup was backed by google and dismissed by apple. and still, almost nothing happening in ai after the models makes sense to me. a chatbox on top of the old world is not the shift. and this is not the moment to spend years on products a model release can erase in months. i say that as someone who had to admit it about my own work first. so i am building the part that is missing: **ai that is truly for humans. open, accessible, usable.** this one probably takes on most of the big names, or it makes sense to them. either way, it is the ride i am on. - **all open source.** read it, fork it, run it yourself. - **all free.** no plan, no tier, no lock. - **any model.** claude, chatgpt, or the one on your laptop. - **open to you.** an idea this size is built by many hands. [read the long take, with the data →](https://openthings.live/why/) anshuli founder, openthings my other work: [intent](https://getintentai.com), [nextmillionai](https://nextmillionai.org), nokk ai, kloned and widely io. ## two years of work, written down. my takes. open one for a short intro **anshuli**founder, openthings **two years of building, and hundreds of inventions studied, from history and from ai,** so this vision is clear enough to read in an evening. why i am doing it, what gets built, and what falls away. takes 4 minutes 49 coming soon 2 [start with the long take](https://openthings.live/why/)[or see all the takes →](https://openthings.live/takes/) 01 **the long take**why i am building openthings 20 min read the long one, with the data. ai models and agents are becoming the base layer, and the human and agent network gets built on top of it. the hard problems, what the people building ai say, and where openthings stands. every figure has a source. [read the long take](https://openthings.live/why/)updated 26 oct 2026 02 **the things**three directions, one usability layer 10 min read what openthings builds, in three directions: open software your model runs, agents with a human behind them, and the protocols they stand on. how they fit from the model up, opensoftware as the first thing you can use, and why agents on both sides make the app in the middle redundant. [read the things](https://openthings.live/things/)updated 24 oct 2026 03 **what collapses**and what replaces it 7 min read eleven things that fall away once agents stand on both sides, from the file format and the call to the privacy policy, ranked by how soon i think each one falls. what replaces each, the parts of life it changes, and why i am founding openthings now. [read what collapses](https://openthings.live/collapses/)updated 24 oct 2026 04 **the agent web**search, shared shapes, and the open questions 12 min read the web was built for eyes, and the next one is built for agents. why the visit to the website fades but the question does not, how search becomes a conversation between agents, who stands behind every agent, why much of software becomes templates, who is building the agent web in pieces, and the questions nobody has solved yet. [read the agent web](https://openthings.live/agent-web/)updated 26 oct 2026 - 05 **humans and tech**a long history of carrying less soon - 06 **where we are**what agents need, and the plumbing we already have soon --- skip to content the long take. six years in ai # why i am finally building openthings. ai has had its first order. the second has started, and it is missing its usability layer. this is the core of six years in ai, the last two of them heads down, and what i understand about where it is going. some of it is data. some of it is a feeling i could not shake: that we are building the wrong things, very fast. written by **anshuli** [@anshulix ↗](https://x.com/anshulix) founder of openthings, and its first employee. fourteen years building, six of them in ai. written 20 oct 2026 updated 26 oct 2026 20 min read [all the takes](https://openthings.live/takes/) ## what ai is used for today. consumer ai comes in three shapes look at almost every consumer ai brand being built today. it does one of three things. 01 ### make things faster software, music, video, images. what took weeks takes minutes. cursor, lovable, suno, runway, veo 02 ### add ai to what exists voice overs, customer support, ai voices, a new button inside the app you already use. elevenlabs, support agents, canva, notion, capcut 03 ### a better assistant a smarter chat box for the questions you already had. chatgpt, claude, gemini 04 missing ### the new web of the ai world not one more app. the layer that changes how people use ai itself: agents acting for people, with the people and brands they deal with, on an open network no lab owns. - your own agent - discovery - agent commerce - delegated payments - the human and agent network the five, below → 2.0b people use ai tools in 2026 menlo ventures, 2026 150%+ people use more than one: chatgpt, gemini, claude and perplexity reach, added up menlo ventures, us, 2026 900m people use chatgpt every week openai, feb 2026 the users are here, and they already hand ai real work: in menlo's us survey, 41% of ai users have tried an agent and 24% use one regularly. the money is following: menlo puts consumer ai spend at about $12 billion in 2025 and about $40 billion in 2026. but the shape has barely changed since the chatgpt and claude moments. each ai got smarter, and each still works alone, inside its own walls. **the person is still the one carrying the work between them.** ## the money follows the same road. ai venture funding, crunchbase ### ai venture funding by year foundation model companies took 27% of ai funding in 2024 and 40% in 2025 foundation model companies the rest of ai see the numbers | year | foundation models | rest of ai | total ai | ai share of all venture dollars | |---|---|---|---|---| | 2024 | $31b | $83b | $114b | 34% | | 2025 | $80b | $122b | $202b | close to 50% | in 2025, close to half of all venture dollars in the world went to ai, up from a third in 2024. the chart uses crunchbase's year-end split; its final count, in january 2026, is $211 billion of the $425 billion invested worldwide. openai and anthropic alone took 14% of all venture money on earth. most of the rest goes to enterprises, automations, ai inside services and apps, and doing everything faster. better models. more optimization. it is like building node and angular before there was a web that needed faster websites. ## second order effects. how a tech revolution actually lands the first order of a revolution is the technology itself. the second order is not everything built on it. **it is the few things built on it that change how people use the technology itself.** a few more websites could never make the web. the browser, search, amazon, online payments and social networking did: each one made the whole web more useful to everyone. phones followed the same path with the app store, messaging, rides and the camera. era first order: the technology second order: what made people use it more the web1983 to 2004 tcp/ip**1983** the www proposal**1989** first website**1991** the browser**1994** amazon**1995** google search**1998** paypal**1999** facebook**2004** mobile2007 to 2010 iphone**2007** android**2008** the app store**2008** whatsapp**2009** uber**2010** instagram**2010** ai2017 to now transformer, the llm engine**2017** chatgpt**2022** claude**2023** mcp**2024** trying local llms**2023** open-weight models**2025** personal agents**2026** missing the usability layer. the five below ↓ ## the five things that will make ai elevate. what the web had, and what ai is missing and no, it is not chatgpt or claude. they are the internet, not its application. they are the capability everything will run on, the way the internet was. what made the internet part of every life were five applications on top of it. ai's five are being tried now, inside each lab's walls and through personal agents that sign in as you. none of them is open yet, and none of them puts an agent on the other side. why ai is different every technology until now waited for us to tell it what happened. we observed the world, then logged it: typed the form, filled the profile, entered the timesheet. **ai is the first technology that observes the way people do.** it reads, listens and understands, as your own agent, under your key. when the machine observes, nothing has to be logged by hand, and every app built to collect what we typed loses its reason to exist. the last change this big, in my reading, was the computer, when the record moved from paper into machines. where openthings comes in **openthings is open standards for human things, with agents as the first interface.** it builds ai's five in the open, in three directions: open software your model runs, agents with a human behind them, and OHAN, the protocol they stand on. so none of the five belongs to one lab. the right column below is where it starts. [the three directions →](https://openthings.live/things/) on the web for ai where openthings starts the browser netscape, 1994 anyone could use the web your agent is the interface the ui dies. one agent of your own, on any model, shows you the work the way you understand it, on any screen the openthings mcp launching tue 27 oct search google, 1998 anyone could find anything your data becomes agentic your context lives with your agent, under your key. it goes looking for the people, services and open software you need, instead of you searching apps opensoftware, then open cards on OHAN launching tue 3 nov e-commerce amazon, 1995 anyone could buy from anywhere needs met agent to agent you do not open a business's app. your agent talks to the business's own agent: it asks, compares and agrees. you approve agents with a human behind them, openhiring first next online payments paypal, 1999 anyone could pay a stranger trust by delegation your agent acts and spends within grants you set, with a receipt for every action. trust is written into keys, not promised in terms grants with limits on OHAN in the spec social networking facebook, 2004 everyone could reach everyone how humans communicate people reach each other through their agents, sharing intent instead of copy pasting it. the inbox and the feed fall away, and you meet at the decision OHAN v1, then openDM OHAN v1 tue 27 oct together, these five are the usability layer: the ui, software and communication as we know them give way to agents on both sides, and data that works for the person it belongs to. [what dies next, on the teaser →](https://openthings.live/#death) ## the bubble crashed. the web did not. 1995 to 2005 ### what investors bet on nasdaq composite, closing points see the numbers | when | nasdaq composite | |---|---| | end of 1995 | 1,052 | | end of 1996 | 1,291 | | end of 1997 | 1,570 | | end of 1998 | 2,193 | | end of 1999 | 4,069 | | 10 mar 2000, the peak | 5,048 | | end of 2000 | 2,471 | | end of 2001 | 1,950 | | 10 oct 2002, the low | 1,108 | | end of 2002 | 1,336 | | end of 2003 | 2,003 | | end of 2004 | 2,175 | | end of 2005 | 2,205 | ### what people actually did people online worldwide, millions born along the way: google 1998, wikipedia 2001, facebook 2004, youtube 2005 see the numbers | year | people online | born that year | |---|---|---| | 1995 | 39 million | | | 1996 | 72 million | | | 1997 | 116 million | | | 1998 | 181 million | google founded | | 1999 | 274 million | | | 2000 | 391 million | | | 2001 | 493 million | wikipedia launched | | 2002 | 650 million | | | 2003 | 761 million | | | 2004 | 893 million | facebook launched | | 2005 | 1,026 million | youtube launched | on 10 march 2000 the nasdaq closed at 5,048. by october 2002 it had lost about 78%. **the web was not wrong. the bets were.** money went to the loudest ideas of the moment, like pets.com and webvan, and they were gone within two years. in 1999 excite had the chance to buy google for $750,000 and passed. search was not anybody's priority, not investors' and not builders'. meanwhile people kept coming online, from about 39 million in 1995 to over a billion in 2005. wikipedia, facebook and youtube were all born after the crash. **what became big looks obvious in hindsight. at the time it was on nobody's list.** ### what they said at the time [1995 · newsweek, clifford stoll no online database will replace your daily newspapersource ↗](https://en.wikipedia.org/wiki/Clifford_Stoll) [1998 · red herring, paul krugman the internet's impact on the economy has been no greater than the fax machine'ssource ↗](https://quoteinvestigator.com/2023/10/26/internet-fax/) [1999 · barron's cover amazon.bombsource ↗](https://ritholtz.com/2019/06/amazon-bomb-may-31-1999/) [1999 · bessemer venture partners passed on meeting google's founders, two blocks away. it lists google in its own anti-portfolio source ↗](https://www.bvp.com/anti-portfolio) the same pattern is back. the money goes to bigger models and to agents that click through the old systems faster. **the usability layer, the thing that will look obvious later, is on nobody's list again.** ## it is strange to watch it happen again. at least to me the same is happening with ai. the technology is right. the bets are on the wrong layer. ### 2000: where the money went money raised by 2000, us$, log scale ### 2026: where the money goes largest round per kind of bet, 2025 to 2026, us$, log scale the shiny bets a new layer people touch left to right is my reading of each bet. heights are the money raised. in 2000 the shiny bets got the most, and were gone by 2002. today the biggest rounds go to bigger models and to agents that work the old systems for you: searching, calling, booking, answering tickets. the usability layer, where agents reach people, has almost nothing. it looks like a bubble to me. most of this money builds agents that click through the old systems faster. when those systems give way, what are these agents for? see the numbers | company | the bet | raised | era | |---|---|---|---| | boo.com | fashion, online | $135m | by 2000 | | webvan | groceries, delivered | $771m | by 2000 | | pets.com | pet food, delivered | $183m | by 2000 | | kozmo | anything, in an hour | $250m | by 2000 | | paypal | paying a stranger online | $123m | by 2000 | | google | finding anything | $25m | by 2000 | | calling agents | bland $50m, vapi $50m (may 2026) | $50m | 2025 to 2026, largest round | | seo agents | profound $180m (sep 2026) | $180m | 2025 to 2026, largest round | | company brains | sierra $950m (may 2026), decagon $250m (jan 2026), glean $150m (jun 2025) | $950m | 2025 to 2026, largest round | | booking agents | fora $60m (jul 2026), ai assisted bookings | $60m | 2025 to 2026, largest round | | personal agents | instinct $1b (sep 2026), manus $500m (oct 2026), perplexity $200m for comet (jun 2026) | $1b | 2025 to 2026, largest round | | bigger models | openai $122b (mar 2026), anthropic $65b (may 2026), xai $20b (jan 2026) | $122b | 2025 to 2026, largest round | | the human and agent layer | no round for an open layer for people. the nearest money is agent permissions for companies: arcade $60m (jun 2026) | no round | 2025 to 2026, largest round | half of all venture money now goes to ai, and most of it goes to bigger models and to doing the old things faster. that is still the first order. the labs have their incentives: if they let the usability layer go to the open, they think their value goes down. i differ here. an open layer grows the pie for everyone, the way the open web grew every company built on it. everyone else is running after what works and what looks real right now: enterprises, automations, ai enablement for services and apps. all of it makes the old way faster, and all of it stands on systems that are about to give way. the companies that own where people spend their time will not build what replaces them: they would throttle their own business. not all of it is wrong. **agents and local llms look right.** what they miss is the way people actually start using them, without the hassle: an agent of their own, meeting other people's agents, with nothing in between. that is the usability layer, and it is where the second order of ai starts, the way the browser and search started it for the web. ## the connecting layer is how humans grow. centuries of less cognitive load for centuries, humans have been trying to do one thing: carry less in our heads. every big jump came the same way. first a connecting layer, a base everyone can stand on. then the layers people build above it. roads, the post and ships let goods and letters travel, and trade and markets grew on top. the press let one voice reach thousands. the internet connected the computers, and email, search, saas and online shops grew above it. the phone put all of it in every pocket, and maps, uber and upi grew above that. ### every jump came from a new base layer the base everyone stands on, and what people built above it. heights are my reading of what one person could do. trade markets cities **roads, the post and ships**for centuriesgoods and letters travel books newspapers science **the printing press**1440one voice reaches thousands news wires live markets calling home **the telegraph and phone**1844 to 1876a voice crosses an ocean email search saas online shops **the internet and the web**1983 to 1991computers reach each other maps uber instagram upi **the phone and app stores**2007 to 2008all of it, in every pocket the second order missing its layer the human and agent network today: inside apps and browsers **ai models and agents**2017 to nowmachines that read, write, reason and act a base layer what grew above it becoming the base now what we build on it first see it as a table | when | the base layer | what it did | what grew above it | |---|---|---|---| | for centuries | roads, the post and ships | goods and letters travel | trade, markets, cities | | 1440 | the printing press | one voice reaches thousands | books, newspapers, science | | 1844 to 1876 | the telegraph and phone | a voice crosses an ocean | news wires, live markets, calling home | | 1983 to 1991 | the internet and the web | computers reach each other | email, search, saas, online shops | | 2007 to 2008 | the phone and app stores | all of it, in every pocket | maps, uber, instagram, upi | | 2017 to now | ai models and agents (becoming the base) | machines that read, write, reason and act | today it works inside apps and browsers. next, built on it: the human and agent network. then the second order | what took humans four generations, one person can do in months today. and what was not possible at all, we do every day. ai is only going to expand that, and this time the base itself can think. people already touch it first: 2 billion people use ai tools in 2026. but today the models still work through the old base, inside apps and browsers built for eyes, as one more thing on the phone. **what we want is the other way round: ai models and agents as the base layer, and everything else built on them.** the first thing to build on that base is the layer between people and their agents, so your agent can reach theirs, with a person deciding on both ends. a revolution holds no meaning for what is working today. the car did not make the horse carriage better. it rebuilt roads, fuel, cities and work around itself, and the carriage became scrap. when ai becomes the base, it will rebuild in both directions. the last time the base changed ### the car arrives. the carriage becomes scrap nobody upgraded the carriage. the world was rebuilt for the new base, and what could not take it was left behind. when ai becomes the base ### any car you already own can drive itself much of what works today gains abilities nobody planned for. the rest becomes the next carriage. a picture, not data. the dots are the one small piece each car gets. that is what a new base layer does. it does not negotiate with what exists. your mail, your chats, the old software at work, your clinic's booking system: some of it will be rebuilt to work through your agent, and some of it will stop making sense. **the inventions will come both ways: new things the old ones can do, and new things that make the old ones pointless.** i have my theories about what gets built first when ai is the base. openthings is where i am testing them. [see what openthings builds on it](https://openthings.live/things/) ## the hard problems everyone is working on. and where openthings stands ask the people building ai what is still broken and you get the same list: who owns the thing you talk to, agents that cannot work with each other, no way to find the right agent, memory stuck inside one app, agents nobody can verify, agents that cannot pay, attacks hidden in plain text, models that state wrong things with confidence, and data that is used up or locked away. ask them how it gets solved, and the answers could not be further apart. most of the fixes on offer are a device, a browser or a protocol that one company controls. ### the five layer cake, twice jensen huang's stack describes one side. people live between two. your side your agent - energy - chips - infrastructure - models - applications **the missing piece**an open protocol, agents on both ends their side their agent - energy - chips - infrastructure - models - applications every layer of the cake has a giant behind it, and each one makes its own stack taller. but my agent and yours run on different stacks, from different labs. what connects them is not a layer of either cake. ### how far apart the answers are the same problems, read by the people building ai. left to right is my reading of each view. 1 **jensen huang**nvidia · mar 2026 build the whole stack: energy, chips, ai factories, models, applications 2 **sam altman**openai · jan 2025 confident they know how to build agi, as it was traditionally understood 3 **dario amodei**anthropic · may 2025 models probably hallucinate less than humans, just in stranger ways 4 **satya nadella**microsoft · dec 2024 business apps collapse into an ai tier of agents 5 **demis hassabis**google deepmind · 2025 agi five to ten years out. one or two breakthroughs still needed 6 **andrej karpathy**eureka labs · oct 2025 agents just do not work yet. this is the decade of agents, not the year 7 **ilya sutskever**ssi · 2024 and 2025 we hit peak data. the age of scaling is over, back to research 8 **yann lecun**ami labs · nov 2025 llms are a dead end. build world models instead and openthings not on this line. whichever of them is right about the model, agents on both ends still need an open way to reach each other, and a person to decide. the problem what is being tried what i think the real answer is **the interface**every company wants to own the thing you talk to humane's ai pin was bricked in feb 2025 after hp bought the company. openai says its first device will not ship before the end of feb 2027, and reports point to an agent-first openai phone around 2028. apple pays for google's gemini to power the new siri. browsers become agents too: comet (jul 2025), chatgpt atlas (oct 2025). hardware loses its weight once the intelligence is in the model. if ai can make even a table talk, nobody needs a new gadget to reach it. what has to change is the systems: they become agentic and take on the intelligence. that is what openthings shows, one system at a time. **interoperability**agents from different makers cannot work together mcp (anthropic, nov 2024, given to the linux foundation's agentic ai foundation in dec 2025) connects agents to tools. a2a (google, apr 2025, now at the linux foundation) connects agents across frameworks. newer pieces keep arriving: the NANDA index (mit media lab and project nanda) resolves an agent's name to where it lives, and aweb gives agents signed mail between them. each connects agents, not the people behind them. a handle and a signed card for every person, so any agent on any model can reach another person's agent. and named shapes both sides speak, so software that never met can share work with no api in between. interoperability between people, not only between tools and companies. **finding**agents still search a web built for eyes every assistant answers from a crawled index, then reads the page. when google shows an ai summary, clicks on a result fall from 15% to 8%, so the visit fades while the crawl goes on: over half of what ai crawlers fetch is pages they already had. open registries of agents fill with noise: of 173,473 on erc-8004, only 3 to 15% work. the live question goes straight to the other side's agent: a page can say a plumber exists, only the plumber knows if he is free tomorrow. the lists of who is out there stay open and many, so whoever pays for one never owns it, and your own agent decides the order, from your own people first. people are introduced, never browsed. **memory and context**every assistant knows you only inside its own app chatgpt (apr 2025) and claude (oct 2025) each remember you inside their own product. switching means pasting a prompt's output into claude's importer (mar 2026), because chatgpt's data export leaves its memory out. memory is not something people should look after, any more than they run their own servers. once agents are the first interface, memory becomes one more layer of communication: everything a person does, at home, at work, anywhere, becomes memory. it is theirs and private, kept by their own agent, and shared only as far as they allow. memory locked in one app, or carried over by copy and paste, never gets there. **identity and trust**few can check who an agent acts for, and only through one company web bot auth, an ietf draft that cloudflare and others put into production, signs an agent's requests so a site knows which company runs it. x401 (in production since oct 2026) ties an agent to a verified person through a certificate, and agentmail's agentid hands an app the owner's email. each runs through one issuer, and none says what the agent may do. a person or an organisation at the root of every agent, with keys they hold, and agents as sub-keys with grants that only shrink. any side can check who stands behind an agent and what it is allowed to do, with no single issuer in the middle. **payments and commerce**an agent cannot pay without a human clicking checkout google's ap2 (v0.2) signs a checkout mandate and a payment mandate, and returns a receipt for each. openai and stripe's agentic commerce protocol powered instant checkout, until openai retired it as a feature of its own in mar 2026: walmart saw shoppers buy three times less inside the chat than on its own site. visa and mastercard each run their own agent scheme, and in oct 2026 godaddy, google, cloudflare and infoblox demoed a passkey approval turning into a signed payment mandate. good rails, each for paying only. businesses want agents to find them, and they want to keep their customer. agents on both sides give them both. payments already have mandates and receipts. everything else an agent does for you does not: a meeting, a draft, a disclosure. one open grant covers all of them, with limits and a signed receipt, and hands a payment to whichever rail you use. **security**an agent can be talked into anything by text it reads ai browsers were shown to follow instructions hidden in pages. openai says prompt injection is unlikely to ever be fully solved (dec 2025), and the uk ncsc warns it may never be fully mitigated. if a model can be talked into anything, it must not hold the keys. and people already skip the yes: 32% of ai users in the us have let ai act for them without a final sign-off. so plain code at the gate checks the grant, and nothing goes out without a yes. **hallucination**models state wrong things with confidence bigger models, reasoning, retrieval, evals. openai's own research (sep 2025) finds training rewards guessing over saying i do not know. dario amodei says models probably hallucinate less than humans, just in stranger ways. the model drafts and never decides. claims travel signed, with their evidence, and the other side's agent checks the evidence instead of believing the text. **data after ai**the public internet is used up, and the rest is locked away ilya sutskever: we hit peak data, there is one internet. labs license archives and make synthetic data, and cloudflare now blocks ai crawlers by default, with pay per crawl (jul 2025). the next data is what people and businesses say for themselves, through their own agents, instead of pages waiting to be crawled. it should stay theirs: shared by intent, scoped, revocable, with a receipt. and the bottom of the cake, energy and chips, is not ours to solve. the person already pays their model vendor for the intelligence. the network never needs its own. every lab is building a taller stack. the problems live between stacks. that is why the answer is an open protocol, with agents on both ends. ## we need a movement like www again. the layer after tcp/ip and http the web did not win because it was the best technology. it won because it was open: nobody owned it, anyone could build on it, and a few shared protocols let every machine and every page reach every other. [**1993**cern puts the web's software in the public domain, so anyone can build on it](https://web30.web.cern.ch/web-history.html)[**1994**tim berners-lee founds the w3c, to keep the web's standards open](https://web30.web.cern.ch/web-history.html)[**6 billion**people online in 2025, about three in four people on earth](https://www.itu.int/en/mediacentre/Pages/PR-2025-11-17-Facts-and-Figures.aspx) ### 01 · the infra we still run on the pipes still work. the top of the stack assumes a person. every site expects a human to log in, click, prove they are not a robot and keep a session. an agent has to pretend to be that person, or get blocked as a bot. the web was built for eyes and hands, and that is now what limits it. assumes a person at a screen - **captcha**2000s prove you are a human, not a bot - **oauth**2012 sign in with a platform you already use - **the login**everywhere an account and a password per site - **cookies**1994 a browser remembers your session - **http and html**1989 to 1993 pages for a person to read and click still works for agents - **ssl and tls**1995 a padlock between you and one site - **dns**1983 names for machines - **tcp/ip**1983 machines reach each other ### 02 · who is working on the agent layer each is a real piece. none is something a person can use yet. most describe documents, tools, companies and agents. the newest drafts start from the person, which is the right start. but none is adopted, none ships to people, and none puts an agent on both sides of a human thing, with a gate that waits for both people's yes, under keys the people hold, without a platform in the middle. the purpose the web had, a network that works for everyone, is still unsolved for agents. w3c [decentralized identifiers v1.0 (2022) and verifiable credentials 2.0 (may 2025)](https://www.w3.org/TR/did-core/) identities and claims anyone can verify, without one company in the middle w3c [ai agent protocol community group (may 2025)](https://www.w3.org/community/agentprotocol/) how agents find each other, share intent, and prove who they are. early, a draft white paper linux foundation [mcp (agentic ai foundation, dec 2025) and a2a (v1.0, signed agent cards)](https://blog.modelcontextprotocol.io/posts/2025-12-09-mcp-joins-agentic-ai-foundation/) agents to tools, and agents to agents across frameworks ietf [web bot auth (draft)](https://workos.com/blog/web-bot-auth-identify-ai-agents) signed requests that tell a site which company runs an agent ietf [principal agent protocol (jun 2026) and agent delegation receipts (oct 2026), individual drafts](https://datatracker.ietf.org/doc/html/draft-baur-pap-02) a person's key as the root, grants that only shrink, signed receipts. proposals, not standards, and nothing a person can use yet ### 03 · the layer that is missing the web was built to carry pages to people. the agent world has to carry the people themselves: who they are, what they know, what they want, and what their agents may do. what the new network has to carry - identity - authority - context - memory - intent - trust - consent - payment - a person's yes - and what robots and machines bring next this is not one more protocol on the web. it is a new network, for people, their agents, models and machines, and it will need things nobody can name yet. OHAN is where i start: the person at the root, everything else built around them. ### the web, how it was built tcp/ip1983 machines can reach each other http and the web1989 to 1991 documents can link to each other browsers and search1993 to 1998 anyone can find anything the web we usesecond order billions of people, everything online ### ai, how it is being built models2017 to now machines can read, write and reason mcp, a2a and agent protocols2024 to now agents can use tools, and talk to other agents missing: the usability layerOHAN a network for people, agents and machines: identity, context, intent, trust, consent, and more the agent websecond order being tried, missing its usability layer after tcp/ip, http changed everything about how we share. with agents it changes again: how humans, machines, robots and llms talk to each other, and stay connected. mcp is a real start. anthropic released it in november 2024, it has more than 10,000 public servers, and it now sits with the linux foundation's agentic ai foundation. **but it connects agents to tools. the pieces for the person are being drafted one by one, and nobody has put them together into something people can use.** that is what [OHAN](https://openthings.live/next/) is for. ## the connecting layers took years, not sprints. tcp/ip, the web, and now nobody built tcp/ip or http in two weeks, or in the six months a funding round wants. bob kahn and vint cerf sketched tcp/ip in 1973, with darpa paying for the patience. it took ten years, round after round of rewriting and a split into two protocols, before the arpanet switched over on 1 january 1983. ### how long the connecting layers took years from the first sketch. hover a dot for the moment. what a funding round expects2 weeks to 6 months tcp/ip10 years, sketch to switch-on **1973** kahn and cerf sketch it, darpa funds it · **1974** the paper, ieee transactions · **1977** first test across three networks · **1978** split into tcp and ip · **1983** the arpanet switches over, 1 january the web4 years to free and easy, 6 to everywhere **1989** the proposal: vague but exciting · **1990** first server, on a next cube · **1991** goes public. conference paper rejected · **1993** cern makes it free. mosaic ships · **1995** netscape goes public. the web goes mainstream the agent layeryear 2. the web's 1991 if history rhymes: 2028 to 2034 we are here **2024** mcp is released · **2025** mcp moves to the linux foundation · **2026** today 0 2 4 6 8 10 12 years since the first sketch tim berners-lee's proposal for the web came back from his manager at cern with a note: vague but exciting. his paper for the hypertext conference in 1991 was rejected, and the demo he gave there instead got a cool reception. the web only took off after cern made it free for anyone in april 1993, and a browser called mosaic made it easy to use. by that clock, ai's connecting layer is in its 1991. public, working, and mostly misunderstood. ## alien to most rooms. my journey, so far when i started talking about this with peers and a few well known investors, it was alien to most of them, or it sounded like something they already knew. i understand why. for thirty years the playbook has been the same: saas, the lean startup, five ways to a hundred million dollar business. all of it bets on scaling the application layer, on top of a layer people touch that someone else already built: the browser, search, and lately the phone. ### i have been alien many times - 2012 to 2013 **drag and drop builders**website and app builders you could use without code, and javascript on the backend, when that was still a strange idea now how many sites get made - then **single page apps**the web as an app, not a stack of pages now how most web apps are built - widely **progressive web apps**i founded widely on pwas, and worked with the people writing the w3c standard now installable on every phone - 2019 **ai solutions**building with ai before most rooms had heard of a language model now in almost every product - now **openthings**the usability layer: agents on both sides, open rules between people alien to most rooms, again being early, again and again, taught me the two things this work needs most: **patience, and staying longer with the problem than anyone else.** with ai, all of it has come true at once. and the time has changed. **the usability layer, the one people actually touch, is being rebuilt.** that needs a different kind of business: not vc shaped, and not conformist to what we all adhered to. i did that in my last startup, and it worked. after a few of those conversations i knew pitching would waste my time. if i could afford it, i should build and experiment instead. i would have loved more hands, we would go faster. but this idea's time is now and the next few years. **we are right on time.** ## tested in a room, with two laptops. and three things for the future i have a vision, and ideas tested on my own problems, in a room with two laptops, trying what i imagined. i think it can work. not alone. with the force of people who want three things for the future. 01 no one or two labs or brands owning human data, or human interests. 02 ai that lets us live a more private, personal and automated life. 03 a world built in open source, free, with ai giving every human capabilities like never before, the way computers and phones did with the internet. [message @anshulix on x](https://x.com/anshulix) [read what is next](https://openthings.live/next/) anshuli founder, openthings ## sources. checked oct 2026 - nasdaq peak, low and decline: [wikipedia, nasdaq composite](https://en.wikipedia.org/wiki/Nasdaq_Composite). year-end closes: [macrotrends](https://pro.macrotrends.net/1320/nasdaq-historical-chart). - people online: [our world in data, number of internet users](https://ourworldindata.org/grapher/number-of-internet-users) (itu and world bank). 1995 to 2004 are summed from country rows. - what they said at the time: clifford stoll, newsweek, 1995 ([wikipedia](https://en.wikipedia.org/wiki/Clifford_Stoll), [paleofuture](https://paleofuture.com/blog/2008/3/28/the-internet-bah-1995.html)); paul krugman, red herring, june 1998 ([quote investigator](https://quoteinvestigator.com/2023/10/26/internet-fax/)); barron's "amazon.bomb", 31 may 1999 ([ritholtz](https://ritholtz.com/2019/06/amazon-bomb-may-31-1999/), [cnbc](https://www.cnbc.com/2021/10/12/jeff-bezos-shared-a-22-year-old-article-predicting-amazons-failure.html)); bessemer's google pass ([bessemer anti-portfolio](https://www.bvp.com/anti-portfolio)). - the hard problems, 2025 to 2026: humane and hp ([wareable](https://www.wareable.com/news/humane-ai-pin-discontinued-hp-acquisition-deal)); openai's device date ([mactrast, on the court filing](https://www.mactrast.com/2026/02/release-of-jony-ive-designed-ai-device-delayed-to-2027-thanks-to-lawsuit/amp/)) and phone reports ([techcrunch](https://techcrunch.com/2026/04/27/openai-could-be-making-a-phone-with-ai-agents-replacing-apps/)); apple and gemini ([android authority](https://www.androidauthority.com/apple-siri-google-gemini-deal-3631890)); mcp to the agentic ai foundation ([mcp blog](https://blog.modelcontextprotocol.io/posts/2025-12-09-mcp-joins-agentic-ai-foundation/)); claude's memory import ([pcworld](https://pcworld.com/article/3076376/claude-can-now-import-chat-histories-from-chatgpt-and-other-ais.html), [techcabal](https://techcabal.com/2026/05/27/4-easy-steps-to-move-your-chatgpt-memory-to-claude/)); web bot auth ([workos](https://workos.com/blog/web-bot-auth-identify-ai-agents)); ap2 and agent payments ([siliconangle](https://siliconangle.com/2025/09/16/googles-agent-payments-protocol-provides-framework-ai-automate-consumer-purchases/), [aimultiple](https://research.aimultiple.com/agentic-payments)); ai browsers and prompt injection ([simon willison](https://simonwillison.net/tags/perplexity/), [techcrunch](https://techcrunch.com/2025/12/22/openai-says-ai-browsers-may-always-be-vulnerable-to-prompt-injection-attacks/), [cyberscoop](https://cyberscoop.com/openai-chatgpt-atlas-prompt-injection-browser-agent-security-update-head-of-preparedness/)); why models hallucinate ([openai](https://openai.com/index/why-language-models-hallucinate/)); cloudflare pay per crawl ([cloudflare](https://blog.cloudflare.com/introducing-pay-per-crawl/)). added on 26 oct 2026: the NANDA index ([ietf draft](https://datatracker.ietf.org/doc/draft-raskar-agentic-web-federated-resolution/)); aweb ([github](https://github.com/awebai/aweb)); clicks with an ai summary ([pew research center](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/)); crawlers re-fetching unchanged pages ([techcrunch](https://techcrunch.com/2026/07/01/cloudflares-new-policy-pushes-ai-companies-to-pay-for-publishers-content/)); erc-8004 ([arxiv](https://arxiv.org/html/2606.26028)); x401 ([help net security](https://www.helpnetsecurity.com/2026/06/26/proofs-x401-establishes-an-open-protocol-for-ai-agent-identity-and-authorization/)); agentid ([agentmail](https://www.agentmail.to/blog/introducing-agentid)); instant checkout retired ([techcrunch](https://techcrunch.com/2026/03/24/openais-plans-to-make-chatgpt-more-like-amazon-arent-going-so-well), [modern retail](https://www.modernretail.co/technology/what-went-wrong-with-chatgpts-instant-checkout/)); the passkey to payment demo ([tech coffee house](https://techcoffeehouse.com/2026/10/16/godaddy-google-and-cloudflare-unveil-ai-agent-trust-system/)). - the web and its infra: cern's public domain release and the w3c ([cern, web history](https://web30.web.cern.ch/web-history.html)); people online in 2025 ([itu, facts and figures 2025](https://www.itu.int/en/mediacentre/Pages/PR-2025-11-17-Facts-and-Figures.aspx)); [dns](https://en.wikipedia.org/wiki/Domain_Name_System), [cookies](https://en.wikipedia.org/wiki/HTTP_cookie), [ssl and tls](https://en.wikipedia.org/wiki/Transport_Layer_Security), [oauth 2.0, rfc 6749](https://datatracker.ietf.org/doc/html/rfc6749), [captcha](https://en.wikipedia.org/wiki/CAPTCHA). agent standards: [w3c did core](https://www.w3.org/TR/did-core/), [verifiable credentials 2.0](https://startwithidentity.com/blog/2025-05-15-w3c-publishes-verifiable-credentials-data-model-2-0-as-a-recommendatio/), [w3c ai agent protocol community group](https://www.w3.org/community/agentprotocol/), [a2a v1.0](https://about.you.com/resources/a2a-protocol-explained-what-agent-to-agent-communication-solves). - excite and google, 1999: [today i found out](https://www.todayifoundout.com/index.php/2013/03/excite-had-a-chance-to-buy-google-for-750k-but-turned-it-down/), from george bell's own account. - app store at launch and a year later: [fortune, 2009](https://fortune.com/2009/07/11/snapshot-of-the-iphone-app-store-one-year-later). - ai venture funding, 2024 and 2025: [crunchbase news](https://news.crunchbase.com/ai/big-funding-trends-charts-eoy-2025/), and its final 2025 count ([crunchbase news, jan 2026](https://news.crunchbase.com/venture/funding-data-third-largest-year-2025/)). - consumer ai users, agent use, use of more than one assistant, acting without a sign-off, and spend: [menlo ventures, the state of consumer ai 2026](https://menlovc.com/perspective/2026-the-state-of-consumer-ai/) ([report](https://menlovc.com/wp-content/uploads/2026/09/menlo_ventures_consumer_ai_report-2026.pdf)). the share figures come from its survey of 5,067 us adults. - person-first agent drafts and agent permissions: [principal agent protocol](https://datatracker.ietf.org/doc/html/draft-baur-pap-02), [agent delegation receipts](https://datatracker.ietf.org/doc/draft-nelson-agent-delegation-receipts/), [attenuated delegation profile](https://datatracker.ietf.org/doc/draft-hamr-oauth-agent-delegation/) (ietf individual drafts, no ietf standing); arcade's $60m round ([pymnts, jun 2026](https://www.pymnts.com/?p=3850692)); ap2 v0.2 ([ap2 specification](https://ap2-protocol.org/ap2/specification/)). - chatgpt weekly users: [techcrunch, feb 2026](https://techcrunch.com/2026/02/27/chatgpt-reaches-900m-weekly-active-users), on openai's announcement. - where the money went, 2000: [wikipedia, webvan](https://en.wikipedia.org/wiki/Webvan), [venturebeat, the worst vc investments](https://venturebeat.com/technology/20-worst-vc-investments-of-all-time), [google's 1999 funding release](https://thenextweb.com/news/googles-press-release), [x.com and paypal, april 2000](https://www.paypalobjects.com/html/pr-040500.html). today: [valueaddvc, biggest ai rounds of 2026](https://valueaddvc.com/blog/biggest-ai-funding-rounds-of-2026-so-far-ranked), [techcrunch, instinct](https://techcrunch.com/2026/09/28/viral-ai-agent-instinct-raises-1b-series-c-at-a-10b-valuation/), [cnbc, manus](https://www.cnbc.com/2026/10/08/manus-fund-raise-meta-muse-tencent.html), [techtimes, perplexity comet](https://www.techtimes.com/articles/318028/20260608/perplexity-raises-200-million-comet-ai-browser-agent-economy-front-door.htm), [cmswire, sierra](https://www.cmswire.com/customer-experience/sierra-raises-950m-at-15b-valuation-eyes-transformation-beyond-customer-support/), [cmswire, decagon](https://www.cmswire.com/contact-center/decagon-triples-valuation-to-45b-with-250m-series-d/), [businesswire, glean](https://www.businesswire.com/news/home/20250610090029/en/Glean-Raises-$150M-Series-F-at-$7.2B-Valuation-to-Accelerate-Enterprise-AI-Agent-Innovation-Globally), [thenextweb, profound](https://thenextweb.com/news/profound-180m-series-d-ai-answer-engine-optimisation), [techcrunch, fora](https://techcrunch.com/2026/07/16/ai-powered-travel-agency-fora-hits-unicorn-status-raises-60m/), [fortune, bland](https://finance.yahoo.com/technology/ai/articles/exclusive-voice-ai-startup-bland-130000952.html), [vapi](https://finance.yahoo.com/sectors/technology/articles/vapi-raises-50m-series-b-130000518.html). - the connecting layers: [wikipedia, printing press](https://en.wikipedia.org/wiki/Printing_press), [electrical telegraph](https://en.wikipedia.org/wiki/Electrical_telegraph), [invention of the telephone](https://en.wikipedia.org/wiki/Invention_of_the_telephone), [app store](https://en.wikipedia.org/wiki/App_Store_(Apple)). the heights in the staircase are my reading. people using ai tools in 2026: menlo ventures, below. - the hard problems: [nvidia, the ai five layer cake](https://blogs.nvidia.com/blog/ai-5-layer-cake), [venturebeat on altman's reflections](https://venturebeat.com/ai/openai-begins-2025-with-massive-hype-for-agi-superintelligence), [techcrunch on amodei](https://techcrunch.com/2025/05/22/anthropic-ceo-claims-ai-models-hallucinate-less-than-humans/), [cx today on nadella](https://www.cxtoday.com/data-analytics/microsoft-ceo-ai-agents-will-transform-saas-as-we-know-it/), [nbc news on hassabis](https://www.nbcnews.com/business/business-news/ai-can-match-humans-task-will-five-10-years-google-deepmind-ceo-says-rcna196771), [the new stack on karpathy](https://thenewstack.io/openai-co-founder-ai-agents-are-still-10-years-away/), [the decoder on sutskever at neurips 2024](https://the-decoder.com/?p=19700) and [his nov 2025 interview](https://forum.effectivealtruism.org/posts/iuKa2iPg7vD9BdZna/highlights-from-ilya-sutskever-s-november-2025-interview), [euronews on lecun](https://euronews.com/next/2025/11/20/french-godfather-of-ai-yann-lecun-confirms-he-is-leaving-meta-to-launch-ai-start-up), [the new stack on a2a](https://thenewstack.io/google-donates-the-agent2agent-protocol-to-the-linux-foundation/), [openai adopting mcp](https://shellypalmer.com/2025/03/openai-and-anthropic-play-nice-its-a-big-deal-for-agents/), [openai, memory](https://openai.com/blog/memory-and-new-controls-for-chatgpt), [anthropic, memory](https://claude.com/blog/memory), [techcrunch on prompt injection](https://techcrunch.com/2025/12/22/openai-says-ai-browsers-may-always-be-vulnerable-to-prompt-injection-attacks/). the spectrum's left to right order is my reading. - tcp/ip: [ethw, the tcp milestone](https://ethw.org/Milestones:Transmission_Control_Protocol_(TCP)_Enables_the_Internet,_1974), [computer history museum, the 1977 test](https://computerhistory.org/blog/born-in-a-van-happy-40th-birthday-to-the-internet/) and [darpa](https://www.darpa.mil/about/innovation-timeline/tcp-ip). - the web: [cern, a short history of the web](https://home.cern/science/computing/birth-web/short-history-web), and [cybercultural, hypertext 91](https://cybercultural.com/p/1991-tim-berners-lee-trys-to-convert-the-hypertext-faithful/). - mcp: [anthropic, donating the model context protocol](https://www.anthropic.com/news/donating-the-model-context-protocol-and-establishing-of-the-agentic-ai-foundation). --- skip to content take 03. what collapses # what collapses. and the far better thing that replaces it. written by **anshuli** [@anshulix ↗](https://x.com/anshulix) founder of openthings, and its first employee. fourteen years building, six of them in ai. written 23 oct 2026 updated 24 oct 2026 7 min read [all the takes](https://openthings.live/takes/) people have always carried less in their heads to do greater things. ai is the biggest step yet. here is what i think falls away once agents stand on both sides, on an open protocol, and what replaces it, ranked by how soon i think each one falls, fastest first. the order is my reading. the teaser names the three that matter most, [the ui, software and communication as we know them](https://openthings.live/#death). here are all eleven. the long history of humans and technology gets its own take, soon. 01 · data types and formats falls first ## the file format. replaced by your way of understanding formats made sharing easy, then hardened into rules: pdf, deck, doc. the lines already blur. formats existed to make understanding easier, and agents do that better. i think in text. someone else in video, someone else in a dashboard. your agent shows the same work the way you understand. i think formats end fastest. app first agent first 02 · reading to understand ## the dashboard. replaced by the answer you asked for for decades we used technology to read less and understand more: the summary, the chart, the dashboard. nothing beats someone answering exactly what you asked. your agent watches, asks the agents on the other side, and tells you. visualization stays, as a taste, not a necessity. app first agent first 03 · software ## software as the product. replaced by scaffolding any agent can run they say software is dead, because an ai builds an app in an afternoon. software is not dead. what it was worth is: products worth billions fall to pennies, maybe not today, but soon. when something gets that cheap, we do not stop using it. we build what it never could. if anyone could make a car, we would not just make more cars. we would make ones that fly, seat hundreds, turn into another vehicle, run on any fuel. so software becomes scaffolding: plain files, clear tools, open rules, run by any agent on your laptop or your cloud. no seat to rent. the purpose it serves moves ten times faster. opensoftware, launching tuesday 3 nov, is the simple start: one line installs the open twin of the software you use. app first agent first 04 · typing and clicking ## the form and the click. replaced by intents between agents forms and buttons exist because the other side is a screen. agents filling forms for you is the small version. when the other side has an agent too, there is no form at all: yours sends what you want, theirs answers, you say yes. we wrote letters until the other side had a computer, and emails until it had a phone. the medium always follows the other side. now the other side is an agent. app first agent first 05 · video and audio calls ## the call as we know it. replaced by agents meet first, people meet at the decision a call today is a slot found in two calendars, a link, a waiting room, an hour, and notes nobody reads. most of that hour is two people trading what their agents could have traded. when agents meet first, they settle what is already settled and bring back only what needs you. the call that is left is short, live and about the decision. and the new conversation models make even that call feel like talking to a person, not to a screen. app first agent first 06 · one size for everyone ## the one size interface. replaced by true personalization every system has its own way, and we learn all of them. personalization today is a theme and a dark mode. the ai browsers (comet in july 2025, chatgpt atlas in october 2025) are the small version: an agent clicking through screens that were built for you. the bigger one has started: agents can hand a screen to whatever shows it, through google's a2ui and mcp apps (jan 2026). with intelligence underneath, the interface is made for you, for this moment, from the same data. three people, three interfaces, none of them saved. app first agent first 07 · the feed ## the feed. replaced by choices that are yours for years we fought doomscrolling with apps that tweak the algorithm. it never worked, because doomscrolling is not manufactured. it is human behaviour, read back to us and ranked for whoever owns the feed. an agent is the first thing that can rank for you, exactly the way you want. you might still scroll. but the choice is yours. app first agent first 08 · integrations ## the api. replaced by contracts between agents every piece of shared work needs an integration today: an api, a connector, a sync that breaks and leaks. agents on both sides agree a contract instead: what, why, for how long, with proof. no service in the middle. the shipping container gave goods one shape any ship could carry. agents need one open shape for an agreement. app first agent first 09 · the system you live in ## the operating system. replaced by your agent, on any device we pick a side: this laptop or that one, this phone or that one. our files, apps and habits stay locked inside it, and switching costs a week. it has started. in 2026 nubia showed a phone with no grid of apps, just one assistant you talk to, and reports say openai is exploring an agent-first phone of its own. but a new phone that only talks to its own maker's agent is the same lock, in a new shape. when your agent is the layer between you and the machine, it goes with you to any screen, any device, any car. the operating system does not vanish. it becomes plumbing: a local model underneath, and intelligence itself showing you how to use any device. app first agent first 10 · machines you operate ## the control panel. replaced by machines that speak for themselves every machine has its own buttons, its own app, its own manual, and we learn each one. a machine with its own agent says what it needs: the filter is due, the battery is low, the cycle is done. your agent tells you only when it matters. app first agent first 11 · privacy set by others falls last ## the privacy policy. replaced by keys you hold privacy today is a terms page you skipped and a banner you clicked away. the company decides what it keeps. with your own keys, what leaves you is a grant you made: to whom, for what, until when, with a receipt of what went out. revoke it, and nothing more leaves. app first agent first parts of life ## eleven collapses. five parts of life. each needs something open. the collapse will come with or without us. what replaces it is not decided yet. left alone, the new layer gets owned by whoever builds the biggest walls. here is where life changes, the part that changes most first, and what has to exist so the better thing wins. part of life what collapses what has to exist, so the better thing wins ### how we work - 01 the file format - 02 the dashboard - 03 software as the product - 04 the form and the click agents on both sides that speak one open protocol, so work moves as intents between people, not as screens between companies. ### how we talk to each other - 05 the call as we know it agents that meet first, under keys people hold, so a call is only the part that needs two people. ### what we pay attention to - 06 the one size interface - 07 the feed an agent that answers only to you, with rules you can read, change and switch off. ### who we trust - 08 the api - 11 the privacy policy keys you hold, and contracts with proof between agents, with no one in the middle keeping a copy. ### the things we live with - 09 the operating system - 10 the control panel an open way for any device and any machine to join with its own agent, not a vendor's app. **that is the mission of openthings:** an open protocol and open tools for all five, owned by no one. why openthings ## ai brings great power. who holds it decides everything. an agent that acts for you is a kind of power we have not handed to anything before. it can sit with the people it acts for, or with the few who build the biggest walls around it. **openthings is my bet on the first.** the companies that own where people spend their time will not build it: they would throttle their own business. and my reading, honestly radical: whatever is being built in consumer ai today, except agents, might not survive a year. that is why i am founding openthings, now. this is the right time not to build one more ai app that makes millions in revenue. nothing against those. it is the right time to work on the layer underneath: **for people and ai, what the web was for people and information.** open, owned by no one, there for everyone. **your keys**your identity and your data sit under a key you hold. no platform holds it for you. **your gate**nothing reaches you, and nothing leaves you, without passing a gate you set. **your yes**agents draft, screen and coordinate. they never decide. people do. i have spent fourteen years in tech as a product thinker, six of them building with ai. the early signs say this future is coming. it still needs building. i do not know if what i believe will happen. i can only try. it is a big dream, and i believe many, many people will share it with me. anshuli founder, openthings [share the dream with me on x](https://x.com/anshulix) [all the takes](https://openthings.live/takes/) ## sources. - ai browsers: [techcrunch, on comet and chatgpt atlas](https://techcrunch.com/2025/12/22/openai-says-ai-browsers-may-always-be-vulnerable-to-prompt-injection-attacks/). - agents that hand over a screen: [a2ui](https://a2ui.org/). - mcp apps: [mcp blog, jan 2026](https://blog.modelcontextprotocol.io/posts/2026-01-26-mcp-apps/). - nubia's phone without an app grid: [racunalniske novice](https://www.racunalniske-novice.com/en/the-first-smartphone-without-apps-run-by-artificial-intelligence-agents/). - an agent-first openai phone: [techcrunch, april 2026](https://techcrunch.com/2026/04/27/openai-could-be-making-a-phone-with-ai-agents-replacing-apps/). - the shipping container: [wikipedia, malcom mclean and the ideal-x](https://en.wikipedia.org/wiki/Malcom_McLean). - the shifts, and what replaces each one, are my reading. --- skip to content take 04. the agent web # the web was built for eyes. the next one is built for agents. written by **anshuli** [@anshulix ↗](https://x.com/anshulix) founder of openthings, and its first employee. fourteen years building, six of them in ai. written 26 oct 2026 12 min read [all the takes](https://openthings.live/takes/) for thirty years we built the web for people looking at screens: pages to read, menus to click, forms to fill. now more and more of the looking is done by an agent, and an agent does not want a page. it wants to ask. this take is about what that changes: search, the shape of our data, software itself, and who stands behind every agent. some of it is being built right now, in pieces, by many people. some of it is still an open question, and those are the ones i want to work on in the open. 01 · the visit ## the visit to the website is fading. the question is not. you already feel it. you ask claude or chatgpt, and the answer comes back with a few links you never open. pew research looked at 68,879 google searches and counted the clicks. searches where someone clicked a result the page is not gone. agents still read it, and the index underneath search still does its job. what is fading is the trip. people stop going to the website, because their agent went for them. and an agent reading a page made for a person is a strange sight. it wades through the menu, the banner and the cookie notice to find one sentence. it fetches the same page again and again, because it cannot tell what changed. over half of what ai crawlers fetch is pages they already had. app first agent first an agent does not want a page. it wants to ask. 02 · search ## search becomes a conversation between agents. finding things on the web has already changed shape three times. each time, the thing in the middle got smarter, and you did less of the work. 1994 · a directory **someone lists it by hand** two stanford students keep a list of sites they like, jerry and david's guide to the world wide web. it becomes yahoo. 1998 · an index **a machine reads everything** google crawls every page it can find, and orders them by the links pointing at them. 2024 · an answer **it reads the pages for you** google's ai overviews summarise the pages, so you read the answer, not the page. next · a question **your agent asks theirs** the answer comes live, from whoever knows it, not from a page that was right last week. search today is three jobs, done by one company. it finds the candidates. it gives you an answer. and it decides the order. on the agent web, i think those three come apart. my reading the job today on the agent web **finding who is out there** a crawled index of pages. still an index, but **open, and more than one**, so no single list becomes the gate. **getting the answer** a page, fetched and summarised, as fresh as the last crawl. **the other side's agent, live.** your agent asks theirs. **deciding the order** the platform, and whoever pays it. **your own agent,** from your own people first: who you know, who you dealt with, who vouches. the middle row is the big one. **a page can tell you a plumber exists. only the plumber can tell you he is free tomorrow at ten.** live, personal questions, is she free thursday, what is your price today, can you deliver on friday, are better asked to the other side's agent than read from a page. here is how i picture a search like that. 01 **it looks in open lists** more than one, so you are never stuck with the one list that sold its top spot. 02 **it asks their agents, live** a few at a time, a short question each. nothing is shouted to everyone. 03 **it ranks for you** people you know and dealt with first, then what is hard to fake. never what a business says about itself. 04 **you decide** the agents draft and compare. the yes, the booking, the price, stays yours. ### lists have a habit. the first step does not go away. to know whom to ask, your agent still needs a list of who is out there. and lists have a habit. the yellow pages began as a plain list of every business in town. by 2007, businesses in the us were paying $14.3 billion a year in directory ads to stand out in it. the open networks of the last few years felt the same pull. they were built to have no owner, and the list ended up owned by whoever paid for it. farcaster's biggest indexer bought the whole protocol in january 2026. when nostr's main search service went dark this autumn, apps found out by timeouts. so the lists of the agent web must be open, there must be more than one, and any one of them must be easy to replace. one detail from nostr stays with me. when that shared search and spam list went dark, apps fell back on each person's own follows and mutes. the only trust that survived was the person's own. that is why the order belongs to your agent. and seo? i think it loses its meaning. seo was the craft of helping a crawler understand a page that could not speak for itself. on the agent web, the things we used to crawl speak for themselves: the plumber's agent says when he is free, the shop's agent says what is in stock today. my reading what comes after is a more efficient kind of search. i do not know its name yet. but it will have its own rules, the way seo once had, and people will learn to play by them. the question is who writes those rules. i want them written in the open, with your agent playing for you. 03 · who is behind it ## every agent has someone behind it. in july 1993 the new yorker ran a cartoon by peter steiner. a dog sits at a computer and tells another dog: "on the internet, nobody knows you're a dog." for thirty years that was the joke, and the problem. on the agent web it cannot stay a joke. there will be far more agents than people. one social network for agents counted 1.5 million agents, owned by about 17,000 people. that is about 88 agents for every owner, and no way to tell which ones a person stood behind. so every agent has to answer one simple question: **who do you act for?** app first agent first about 88 agents for every owner: 1.5 million agents and about 17,000 owners, on one social network for agents. i see two kinds of someone. a person, and an organisation: a shop, a clinic, a company. every agent hangs under one of them. and when an agent starts another agent to help, the helper is not a new citizen. it is one step further down the same chain, with less power than the step above. my reading a person - **you**hold your keys, and give every yes. - **your agent**does what you allowed, and says it is an agent. - **a helper it starts**reads and drafts for it. never meets a stranger, never spends. an organisation - **the bakery**with a real person inside who answers for it. - **its agent**answers customers inside limits that person set. - **a helper it starts**checks the stock, drafts a reply. a price or an order still waits for a person. people and organisations should also be found differently. a bakery wants to be found. a person mostly does not want to be browsed. so on the agent web, i think you reach a person the way you reach a friend today: someone gives you their name, sends you an invite, or introduces you. an organisation can choose to be listed, with a real person inside who answers for what its agents say. the law already leans this way. in 2024, air canada's chatbot told a grieving passenger that a bereavement fare could be claimed after the flight. the airline argued the chatbot was responsible for its own words. a tribunal disagreed, and air canada paid. 04 · data and formats ## formats for the eye fall. shapes for agents rise. in [what collapses](https://openthings.live/collapses/#formats) i wrote that the file format falls first. the pdf, the deck and the doc were made so everyone reads the same thing the same way. your agent can show you the work the way you understand it best. i still think that falls fast. but when two people's agents share something, a task, a booking, a contact, they still have to agree on what it is. not how it looks. what it is: which part is the title, which is the due date, who may change the status. today that agreement is buried in an api someone wrote years ago, between two apps that already know each other. on the agent web, i think it comes out into the open: **a named shape both sides speak,** that the model reads to understand and plain code uses to check. two pieces of software that never met can then share a list and keep it in step. my reading app first agent first the contract does not die. it moves, from code written once into a shape chosen each time two people share. the closest thing i know is the barcode. on 26 june 1974, a cashier at marsh supermarket in troy, ohio scanned a ten pack of wrigley's juicy fruit gum. it was the first product sold by scanning a barcode. the barcode was not clever. it was agreed: makers printed it, shops read it, and both got something back, faster tills and stock they could count. a shared shape spreads when both sides gain from it. the web taught the same lesson twice. schema.org, the markup that tells search engines what a page is about, reached 44% of domains, because google rewarded it with richer results. llms.txt, a file that tells ai what a site is about, sits on tens of thousands of sites, and 97% of them got no ai requests in one month this year, because nothing rewards it. the parts for shapes between agents already exist, and that is good news. the at protocol behind bluesky finds named schemas through dns. ink & switch showed how one shape of a thing can be mapped onto another. automerge keeps two copies of a document in step with no server in the middle. what is missing is putting them in the hands of people's agents, with a reason for both sides to use them. 05 · software ## much of software becomes templates. in 2009 apple sold the iphone on a promise: "there's an app for that." and there was. an app for the torch, the tip, the train times, the timer, the level on the shelf. most of them now fit in one sentence to your agent. once the model is the intelligence, software does not need to be smart. no program is as smart as the model you already talk to. it needs to be scaffolding: plain files, clear tools and open rules your agent works through. here is where i think each kind of software goes. my reading **an app for one job** your agent does the job itself, or rebuilds the app from what it does. - falls first **software many people share** it stays. your agent works it for you, and its screens matter less. - changes how it is used **the part between people** the api, the connector, the sync. it becomes a shape both agents speak. - becomes a template the third kind is where the biggest change hides. most of what we pay software for is not the screen. it is the shared part: the board my team and i both see, the booking both sides trust. when that part is a template any agent can speak, shared work stops belonging to one app. that is why openthings builds open software at all. opensoftware brings open tools to people through the ai they already use. scaffolding that speaks open shapes can work with anyone's, not only with its maker's. 06 · the work behind it ## the agent web is being built right now, in pieces. this is not one company's project, and it should not be. look at what arrived in about two years, and where each piece sits between two people. agents reaching tools **mcp**, now at the agentic ai foundation under the linux foundation, connects an agent to tools and data. agents reaching agents **a2a**, also at the linux foundation, lets agents on different frameworks work together, each described by an agent card. finding an agent the **NANDA index**, from mit media lab and project nanda, resolves an agent's name to what it is and where to reach it. you your agent **OHANperson to person, with the yes on your own device** their agent them mail between agents **aweb**, an open source project, gives agents addresses under a domain, signed mail between them, and a way to wake an agent when something arrives. a person behind the agent **x401** ties an agent to a verified person. agentmail's **agentid** tells an app whose email stands behind an agent. a yes you can check this october, **godaddy, google, cloudflare and infoblox** showed a passkey approval turning into a signed payment mandate, in a demo. i am glad every one of these exists. it means the problem is real, and many people see it. most start from the agent, or tie the person in through one issuer, one email, or a server that can read the mail. the piece i care about is the span in the middle: **person to person, through your agents, with the yes on your own device.** that is where OHAN, the open human and agent network, starts. and where someone has already built a piece well, OHAN should plug into it, not rebuild it. 07 · open questions ## what nobody has solved yet. the hard parts of the agent web are not solved, by us or by anyone. these are the questions openthings is working on. i want to work on them in the open, with researchers, builders, and anyone who has watched a network go wrong. 01 **a yes others can check** when you approve something, can the other side verify that a person said yes, without learning who you are? 02 **software that never met** how do two agents share an item, keep it in step, and refuse what they should not accept, with no api between them? 03 **power that only shrinks** when an agent hands part of a job to another, how does every step carry less power, and some acts never pass on at all? 04 **finding without a centre** how does your agent find who can help among millions, when nobody owns the list? 05 **who stands behind an agent** how does an agent show that a person or an organisation is behind it, without one company becoming the gate? 06 **words that steer** how does an agent read a stranger's message without being steered by it? 07 **while you sleep** how does a message reach your agent when your laptop is shut, with no server that can read it? 08 **who keeps the shared parts** who runs the lists and relays everyone uses, who pays for them, and what keeps any one from becoming the next centre? any answer has to keep a few rules. a person or an organisation stands behind every agent. the model drafts, and a person decides. no server reads your messages. nobody can take your name. revoking stops what comes next, and cannot take back what was already shared. if one of these is your question too, [let's connect](https://openthings.live/connect/). 08 · why it is larger ## it is not an app. it is a layer. it is easy to look at agents talking to agents and see a new kind of messaging. it is bigger than that. the same few pieces, someone behind every agent, a name, a door you control and a yes only you give, carry everything else. **"is she free thursday?"**a live question between two people, answered by their agents, decided by them. **"can you come tomorrow?"**a business's agent answering yours, while a price or a booking still waits for a person. **"keep this list with me"**two pieces of software that never met, keeping one list in step through a shape both speak. **"find me someone who can"**search where your agent decides the order, from your own people first. **"the filter is due"**a machine that speaks for itself to your agent, instead of a panel of buttons you learn. **"build on it"**a list, a relay or an app, made by anyone on open rules, without asking us. the web was not one company's product, and the agent web should not be either. if someone answers one of these questions better than i do, the open rules should take their answer. that is what open means to me. agents are becoming the way we reach everything. who they answer to decides what kind of world that is. i want them to answer to people. anshuli founder, openthings [tell me which question is yours](https://x.com/anshulix) [all the takes](https://openthings.live/takes/) ## sources. - clicks with and without an ai summary: [pew research center, july 2025](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/). - over half of ai crawler traffic fetches unchanged pages: [techcrunch, on cloudflare, july 2026](https://techcrunch.com/2026/07/01/cloudflares-new-policy-pushes-ai-companies-to-pay-for-publishers-content/). - jerry and david's guide to the world wide web, 1994: [wikipedia, yahoo](https://en.wikipedia.org/wiki/Yahoo). google and pagerank, 1998: [wikipedia, pagerank](https://en.wikipedia.org/wiki/PageRank). ai overviews, may 2024: [google](https://blog.google/products/search/generative-ai-google-search-may-2024/). - us directory advertising, $14.3 billion in 2007: [purple motes, from industry data](https://www.purplemotes.net/2009/03/08/telephone-directory-advertising-data/). neynar buys farcaster: [the block, january 2026](https://www.theblock.co/post/386549/haun-backed-neynar-acquires-farcaster-after-founders-pivot-to-wallet-app). nostr's search service going dark, and apps falling back on follows and mutes: [porwr #267](https://github.com/diogel-io/porwr/issues/267), [relay-outpost #343](https://github.com/megistusXYZ/relay-outpost/pull/343). - "on the internet, nobody knows you're a dog", the new yorker, 5 july 1993: [wikipedia](https://en.wikipedia.org/wiki/On_the_Internet,_nobody_knows_you%27re_a_dog). 1.5 million agents and about 17,000 owners: [wiz, on moltbook](https://wiz.io/blog/exposed-moltbook-database-reveals-millions-of-api-keys). air canada and its chatbot: [american bar association](https://www.americanbar.org/groups/business_law/resources/business-law-today/2024-february/bc-tribunal-confirms-companies-remain-liable-information-provided-ai-chatbot/). - the first barcode scan, 26 june 1974: [wikipedia, universal product code](https://en.wikipedia.org/wiki/Universal_Product_Code). schema.org on 44% of domains: [web data commons](https://webdatacommons.org/structureddata/). llms.txt with no ai requests: [ahrefs, may 2026](https://ahrefs.com/blog/llmstxt-study/); its spread: [ppc land](https://ppc.land/llms-txt-adoption-rises-8-8x-but-97-of-files-get-zero-ai-requests/). named schemas through dns: [at protocol lexicon resolver](https://cdn.jsdelivr.net/npm/@atproto/lexicon-resolver@0.4.10/README.md). mapping one shape onto another: [ink & switch, cambria](https://www.inkandswitch.com/cambria/). [automerge](https://automerge.org/). - "there's an app for that", 2009: [wikipedia](https://en.wikipedia.org/wiki/There%27s_an_app_for_that). - the pieces: [mcp](https://modelcontextprotocol.io); [a2a](https://a2a-protocol.org/); [the NANDA draft at the ietf](https://datatracker.ietf.org/doc/draft-raskar-agentic-web-federated-resolution/); [aweb](https://github.com/awebai/aweb); [x401, help net security](https://www.helpnetsecurity.com/2026/06/26/proofs-x401-establishes-an-open-protocol-for-ai-agent-identity-and-authorization/); [agentid, agentmail](https://www.agentmail.to/blog/introducing-agentid); [the passkey to payment demo, tech coffee house](https://techcoffeehouse.com/2026/10/16/godaddy-google-and-cloudflare-unveil-ai-agent-trust-system/). - the readings marked "my reading", and the open questions, are mine. --- skip to content the takes. writing from the build # the takes. what i see, written down while i build it. long reads on why openthings exists and where it goes. data where there is data, my reading where there is not, and sources for both. one take at a time. written by **anshuli** founder of openthings, and its first employee. fourteen years building, six of them in ai. [@anshulix on x ↗](https://x.com/anshulix) [01 updated 26 oct 202620 min read the long take ## why i am finally building openthings. ai has had its first order. the second has started, and it is missing its usability layer. where the money goes, ai as the new base layer, the hard problems every lab is working on, and why we need a movement like www again. anshuli written 20 oct 2026 - read it](https://openthings.live/why/) [02 updated 24 oct 202610 min read the things ## the things. three directions, one usability layer: open software your model runs, agents with a human behind them, and the protocols they stand on. opensoftware as the first thing you can use, and why putting the agent before the system makes the old systems collapse. anshuli written 21 oct 2026 - read it](https://openthings.live/things/) [03 updated 24 oct 20267 min read what collapses ## what collapses. eleven things, ranked by how soon i think they fall, from the file format and the call to the operating system and the privacy policy: what falls away once agents stand on both sides, and the far better things that replace each. anshuli written 23 oct 2026 - read it](https://openthings.live/collapses/) [04 written 26 oct 202612 min read the agent web ## the agent web. the web was built for eyes, and the next one is built for agents. how search becomes a conversation between agents, why every agent needs someone behind it, why much of software becomes templates, who is building the agent web in pieces, and the questions nobody has solved yet. anshuli written 26 oct 2026 - read it](https://openthings.live/agent-web/) 05 being written upcoming ## humans and tech. a long history of people carrying less in their heads so they could do greater things: writing, the ledger, the letter, the switchboard, the computer, the phone. and why ai is the biggest step yet. anshuli coming soon coming soon 06 being written upcoming ## where we are. what agents still need to become the first interface, and the plumbing we already have: context that goes with you, any model, memory graphs you own, the agents economy, and software that becomes scaffolding. anshuli coming soon coming soon every take has a markdown copy for agents, in [/markdown](https://openthings.live/markdown/). then ## read one and disagree with it. that is the point. [message @anshulix on x](https://x.com/anshulix) [read what is next](https://openthings.live/next/) --- skip to content the things. what openthings builds # three directions. one usability layer. a world built on llms needs a usability layer: the part people actually touch. openthings builds it in the open, in three directions. open software your model runs for you. agents with a human behind them, meeting on a network. and the protocols they all stand on. each one puts an agent on both sides of a human thing, and takes the app out of the middle. 01 open software **software that runs on your model**opensoftware installs it. small agents do one job for other agents. 02 agents with a human behind **people's agents, on the network**openproof, openDM and openhiring, and any agent that joins. 03 the protocols **OHAN, the open human and agent network**the open rules the other two stand on. written by **anshuli** [@anshulix ↗](https://x.com/anshulix) founder of openthings, and its first employee. fourteen years building, six of them in ai. written 21 oct 2026 updated 24 oct 2026 10 min read [all the takes](https://openthings.live/takes/) [why we need a usability layer](https://openthings.live/why/#layer) ## how it fits, from the model up. read it from the bottom your side **you decide** nothing leaves or lands without your yes. your agent **the assistant you already use** with its own key, under yours. 01 it runs **open software** opensoftware installs it small agents 03 the protocols **OHAN** the open human and agent network one handle. a key you hold. a gate. 02 agents with a human behind **openproof**proof, both ways **openDM**messages, both ways **openhiring**hiring, both ways agent to agent. a person behind each. their side **they decide** nothing reaches them without their yes. their agent **any agent with a person behind it** a friend's assistant, a brand, a clinic, a service. 01 it runs **their own open software** their tools their data, kept there the base layer **any model**open models, small and local models, conversation models, world models. yours to choose, and to switch. read it from the bottom. any model is the base. OHAN puts open rules on it. each person's agent runs open software on its own side, and meets the other person's agent through the network things. nothing lands on either side without a yes, and no app sits in the middle. 01 · open software ## software that runs on your model. **software stopped needing to be smart. your model is.** before the models, open software was hard to run because each program had to be the whole thing, built for everyone and so fit for no one. now the intelligence is in the model, and no software is as intelligent as a model. software only needs to be scaffolding: the structure your model works through. the first software you can install. launching tuesday 3 nov 2026 ### opensoftwareby openthings **almost every kind of app you pay for has an open twin now.** rebuilt by agents in weeks, reverse engineered from the original, often better, and built to be driven by a model instead of clicked by a hand. the intelligence is on your side already. the scaffolding is on github. and still almost nobody runs it. because six things stand between a person and open software, and none of them moved when the models came: finding the one that is real, trusting it, installing it, setting it up, getting your agent to drive it, and taking it back when you are done. ten steps for a developer. a wall for everyone else. the people who tried to fix this built app stores, servers and installers. every one of them removed one or two. find trust install set up drive it take it back **opensoftware is your agent doing all six.** it finds the twin, reads the repo's own markers and a security scorecard, tells you whether a model can drive it, shows what the install will touch, installs it when you say yes, shapes it for you, and removes it later with everything it touched. no account, no key: it runs inside the claude, chatgpt or codex you already have, on your own login. swap github.com for opensoftware.live. that is the whole install. the one install is the openthings mcp. every thing after this one arrives on it, with the key it makes for you. it opens on tuesday 27 oct with OHAN v1, so anyone can join the network and message anyone from the ai they already use. opensoftware arrives on it a week later: the first software behind the door. [opensoftware.live ↗](https://opensoftware.live) ### one folder, not five apps. most software today is built for a person at a screen: a server to rent, an account to log into, a dashboard to click through, a price per seat. when the agent is the operator, most of that is redundant. just imagine making a video. record it, cut it, caption it, write the post, schedule it. today that is five apps, five logins and five exports. **built for agents, it is one folder your agent works through, and you step in only to decide.** you approve the cut approve the post - **record**the take lands in your folder - **cut**fillers and pauses out - **caption**words with their times - **write the post**from the script - **schedule**on your calendar **your agent runs every step**one folder. your laptop or your cloud. what becomes redundant the server you rent the login and the account the dashboard you click through the price per seat the export between apps what stays your files, where you want them the work itself a page, only when you want to look your decision the first one i built for agents ### solo ops a solo builder's tasks, meetings, posts, scripts, drafts and video takes, as plain files in one folder. no server, no account. the ai you already use is the operator: it plans the work, moves the cards, cuts the take and fills the posting schedule. the pages are there for when you want to look. it runs on your laptop, or wherever you put the folder. next, two people's agents share work through OHAN, with no shared server. working in my room, not released yet. [solo ops on github, open source, mit ↗](https://github.com/anshulixyz/solo-ops) **video**the transcript is the editor. your agent cuts the take. **home**a page to look at, not a place to work. sample people. solo ops is one. there can be hundreds: the editor, the invoice, the clinic's booking book, the school timetable. build software for agents, and let people run it however they want. ### and small agents, working in the back. not every agent needs to think. a small agent does one job, gives the same answer for the same input, and never decides anything. a parent agent calls it: your assistant, or a network agent like openproof working for its person. **receipt agents**write down what was shared, with whom and under which grant, so there is proof of what went out **old software agents**speak to the bank portal, the clinic's system, the spreadsheet at work **small chores**renew, remind, reconcile, file. the things that eat an evening **open to anyone**write one, publish it, and any parent agent can call it under a grant 02 · agents with a human behind them ## people's agents, on the network. the personal agents people use today act alone, signed in as them, against apps built for eyes. these are agents with a human behind them and an identity of their own, meeting other people's agents. each takes one exchange people do every week and lets agents carry it on both ends. **agents draft, screen and coordinate. people decide.** 001 ### openproof prove a thing about yourself without handing over everything. your commits, sessions and builds become signed claims that stay private on your machine until you publish them. then your agent carries them. agent to human building 002 ### openDM a message your agent can read, screen and answer for you. the first thing that touches every message is an agent, on both ends, and the agent can never send without the rules saying yes. agent to agent, human behind built. launches after openproof 003 ### openhiring both sides send an agent. the hiring agent works on proofs instead of keywords, talks to candidates' agents over openDM, and returns a short list. people meet at the decision. agent to agent, humans behind next **and more like these.** every exchange between people that still runs through an app in the middle can be rebuilt with agents at both ends: introductions, bookings, offers, deals. a few more are named when they are real. 02 · and any agent that joins ## your agent, on the network, without asking us. openthings does not need to build every agent. any team that has built an agent can put it on OHAN, so its users can use their own agents to reach it, and it can reach them back. no partnership, no app store review, no api key from us. a brand's agent, a clinic's booking agent, another builder's assistant, a local model on someone's phone. **if it holds a key and respects the gate, it is on the network.** a handle on your own domain a signed card a grant from its person the gate **OHAN**the same handle, card and gate for everyone. openthings runs one host among many. **assistants**claude, chatgpt, or a local model, acting for its person **brands and services**reachable by agents the way a website is reachable by browsers **other builders' agents**a niche agent with users, who now bring their own agents **people's own agents**the friend, the client, the doctor's office, through their own 03 · the protocols ## OHAN, the open human and agent network. the open rules the other two directions stand on. every agent protocol in wide use is about the agent: [mcp](https://modelcontextprotocol.io) connects it to tools, [a2a](https://a2a-protocol.org) to other agents. **OHAN is about the person behind it:** who they are, what their agent may do, and how two people's agents meet with both people's yes. five plain parts, each borrowed from something that already works. 01 key **you are a keypair**free, self-issued. your agent is a sub-key under you, with a signed grant: scope, limits, expiry. revoke it and it is gone. 02 name **a handle on any domain**a dns record binds your name to your key. no new registry, and nobody owns the namespace. 03 card **a signed profile anyone can fetch**your keys, your agents and what each may do, and the intents you accept. 04 wire **intents over relays**sealed envelopes through small servers anyone can run. unknown senders stop at your gate until you say yes. 05 trust **a proof graph, not a badge**cards and vouches are public and signed, so anyone can index them, the way anyone can crawl the web. protocol connects who it is about mcp an agent to tools and data the agent a2a an agent to other agents, across frameworks the agent OHAN people, through their agents you. your agent acts under your key **the one install: the openthings mcp. any model joins the network through it.**launching tue 27 oct, with OHAN v1. opensoftware arrives on it tue 3 nov [the full spec, on what is next →](https://openthings.live/next/#ohan) the area it can reach ## agents on both sides. the middle becomes redundant. claude, chatgpt and personal agents like openclaw already put an agent on your side. that is one side. the other side is still a website, a form, an inbox, a portal, built for eyes, and your agent has to click through it, or you do. **openthings puts an agent on both sides: yours and theirs, each with a person behind it.** when both sides have an agent, the app in the middle, the form, the portal and the dashboard become redundant, and a new layer enters between them: OHAN, where agents reach each other and people decide. today: an agent on one side openthings: agents on both sides one agent makes the old process faster. agents on both sides make the middle redundant. why the old systems collapse ## put the agent before the system. every system we use has its shape because a person operates it: a login to know who you are, a form to collect what it needs, a dashboard to show you, a notification to pull you back, a price per seat. put an agent before the system, on both sides, and those parts lose their user. they do not get better. **they collapse.** what survives is what the system actually knows: the record, the rules, the expertise. and new systems come to the front, built for agents from the start. built for people a person a screen the system ai forced on a person an agent the same screen the same system the agent clicks through a screen built for eyes agents first you your agentsystems as tools their agentsystems as tools them the screen has no user. it collapses. collapses logins and passwords forms dashboards notifications the price per seat ranked listings and ad slots the app in the middle survives the record the rules the expertise the human decision comes to the front intents, not forms signed claims, not profiles a gate, not an inbox grants, not logins receipts, not dashboards agents you subscribe to, not seats ### human problems, against it the problems people bring to systems have not changed in a century. the systems built around them will. the human problem the system today collapses comes to the front **finding the right people**hiring, clients, cofounders openhiring job boards, social profiles, recruiters the listing, the keyword filter, the pile of 200 cvs agents on both sides match on proof and negotiate first. people meet at the decision. **proving who you are and what you did**credentials, work, reputation openproof cvs, profiles, certificates, star ratings the profile you write about yourself, the rating anyone can fake signed claims from observed work, shown only as far as the other side needs. **reaching and agreeing with someone**messages, meetings, deals openDM email, chat apps, contact forms, calendars the inbox, the contact form, the back and forth intents between agents, a gate on each side, a receipt of what was agreed. **doing your work**tasks, projects, making things open software project tools, dashboards, a seat per person the shared board, the seat, the status meeting plain files your agent operates, shared agent to agent with no shared tool. **buying and booking**shopping, travel, services open to builders marketplaces, search ads, booking sites the ranked listing, the ad slot, the checkout form your agent asks, theirs offers, terms agreed on both sides. a brand pays to be reachable, not ranked. **staying healthy**appointments, records, care open to builders patient portals, forms, booking apps the portal, the same form filled again your agent holds your history, the clinic's agent books. the doctor decides. **money and paperwork**bills, taxes, renewals open to builders bank portals, tax forms, bill sites the logins, the reminders, the evening lost to forms small agents that file, pay and reconcile under a grant, with a receipt. **learning**skills, courses, peers open to builders course platforms, feeds, fixed curricula the feed, the one pace for everyone your agent finds the material and the people at your level, and their agents agree to learn together. openthings starts with four of these, in its three directions. the rest are open to whoever builds them, on the same layer. then ## building in one of these three directions? build it on the open layer with me. [message @anshulix on x](https://x.com/anshulix) [read what is next](https://openthings.live/next/) the long take explains why a world built on llms needs [a usability layer](https://openthings.live/why/#layer). this page is what openthings builds for it. --- skip to content what is next. protocols, trust, and the work ahead # the next things. what is being built, and what is open to you. the teaser says where this goes. [the long take](https://openthings.live/why/) says why. this page says what gets built to get there, in three directions: [open software, agents with a human behind them, and the protocols](https://openthings.live/things/). what comes first, and where another pair of hands would change the pace. ## OHAN, the open human and agent network. v1 launching tue 27 oct the protocol under every thing **mcp connects an agent to tools. a2a connects an agent to other agents. OHAN connects people, through their agents.** every agent protocol in wide use makes the agent the subject. OHAN makes the **person** the subject, and the agent a delegate: one handle, any number of agents under it, each holding a slice of your authority that you can shrink or revoke. five plain things, each borrowed from something that already works. 01 key ### you are a keypair free, self-issued, no permission. an agent is a sub-key under you, with a signed grant: scope, limits, expiry. revoke it and the agent is gone. you are untouched. 02 name ### a handle on any domain a dns record binds your name to your key. no new registry, nobody owns the namespace. a brand brings its own domain and is trusted on day one. 03 card ### a signed profile anyone can fetch your keys, your relays, your agents and what each may do, the intents you accept, your public preferences. the card for the person behind the agent, not only for the agent. 04 wire ### intents over relays sealed envelopes through small dumb servers, several per person, any one replaceable. one round to shake hands. unknown senders stop at your gate until you or a rule let them in. 05 trust ### a proof graph, not a badge cards and vouches are public and signed, so anyone can index them, the way anyone can crawl the web. an index must show its sources. trust lives underneath, not in the index. standards ### built on what exists w3c decentralized identifiers and verifiable credentials, object capabilities for the hierarchy, dns for names, mcp and a2a on the agent side. interoperable at every edge, owned at none. a paper is being written. intents verbs. **request a meeting, offer, update, ack.** an unknown intent degrades to a message, so the vocabulary can grow without breaking anyone. claims nouns with evidence. **who observed it, when, how, with what confidence.** a person's shipped work, a place's step-free entrance, a brand's return policy. grants permissions. **who may do what, within which limits, until when,** signed by the one who holds the authority. a handshake is an intent carrying claims, answered under a grant. that is the whole grammar, for a person, a brand or an agent. ## intent sharing with agents. first pilot say it once. your agent carries it today you say the same thing five times: in a message, a form, a deck, a call, a follow-up. with OHAN you say it once, to your own agent, and it travels as an intent to the other person's agent. it stops at their gate. it lands only when they say yes. the agents draft, screen and coordinate. people decide. you find me an hour with maya next week to go over the launch. she can see the draft, not the budget. your agent resolved @maya, fetched her card. she accepts request-meeting from people vouched by her team. sending the intent with the draft attached, budget withheld. maya's agent at the gate: known sender, scope ok. maya prefers no meetings before ten. proposing tuesday 11:00 or thursday 14:00. maya tuesday. your agent booked. a signed receipt of what was shared, and with whom, is on your card. ## open source, found and run for you. launching tue 3 nov opensoftware, by openthings every day someone opens something made for agents: a server for a tool you already use, a fix for a problem you have, a workflow that took them a month. it is out there, free, and you will never find it. **finding it, judging it and wiring it in is a task. that task should be your agent's.** [opensoftware](https://opensoftware.live) is the first piece, launching tuesday 3 nov 2026, on the openthings mcp. swap github.com for opensoftware.live. that is the whole install. what it is ### open software, checked before it lands your agent reads the repo's own markers and a security scorecard, tells you whether a model can drive it, and shows what the install will touch. next, a registry of open things: signed like a card, indexed like the web, never owned by the index. how it feels ### you say the problem your agent finds the thing, shows you what it will touch, installs it when you say yes, shapes it for you, and takes it back later with everything it touched. no account, no key: it runs inside the claude, chatgpt or codex you already have. ## trust is the floor, not a feature. non negotiable how an agent earns your yes none of this works if you have to hope the agent behaves. so the rules are structural, written into keys and grants, not into a terms page. - **delegation with a leash.** every agent acts under a grant with scope, limits and an expiry. revoke it, and it is gone. in the spec - **consent at both ends.** you choose what your card publishes, field by field. the other side has a gate. nothing lands without a yes. in the spec - **a receipt for everything.** what the agent shared, with whom, under which grant, signed. revoke the grant and nothing more goes out. the receipt proves what already did. in the spec - **no root authority.** your handle lives on any domain. openthings runs one host among many. the record is self-certifying, so any mirror can serve it. in the spec - **keys never on a screen.** recovery is social: a new key and a few vouches from people who already vouched for you. in the spec - **any model.** claude, chatgpt or the one on your laptop, through mcp. same handle tomorrow with a different model. in the spec ## what gets opened. all of it. open source, free each one is a first piece of [the five things that will make ai elevate](https://openthings.live/why/#five): your agent as the interface, data that becomes agentic, needs met agent to agent, trust by delegation, and how humans communicate. grouped by the three directions. ### 01 · open software - **opensoftware**, your agent finds open software, checks it, installs it and takes it back. launching tue 3 nov - **the registry** of open things your agent can find and run, beyond github. planned ### 02 · agents with a human behind them - **openproof**, prove a thing about yourself without handing over everything. building - **openDM**, a message your agent can read, screen and answer for you. built. launches after openproof - **openhiring**, both sides send an agent, people meet at the decision. next ### 03 · the protocols - **OHAN v1**: anyone can join the network and message anyone, from the ai they already use. the spec and a reference relay anyone can run. launching tue 27 oct - **the card format** and the intent, claim and grant vocabulary. in progress - **the openthings mcp**, the way any model joins the network. it ships with OHAN v1, and opensoftware arrives on it a week later. launching tue 27 oct - **intent sharing**, say it once to your own agent. it carries the intent, and it lands only on a yes. first pilot - **trust certificates for llms**, how your agent knows which model it is dealing with, and whether to trust it. exploring ## open to you. i have been alone on this for two years i have worked on this alone for two years, in the open, in the hours i have. it is further along than it looks and further from done than i would like. **i am open, and so open, to working with people who want to help humans enter the post ai world.** not as a brand. as a way. protocol people standards people cryptographers relay operators designers for recovery writers brands with a first agent anyone with a problem [message @anshulix on x](https://x.com/anshulix) [back to the teaser](https://openthings.live/) --- skip to content before you message me # if you are building this, let's connect. openthings needs the people who have already hit the wall. if one of these is you, we are building the same thing from different ends. six signals ## you are one of us if. any one is enough 01 ### you build models that live close to people. on a laptop, in a voice, on a phone. small, local, and working for one person, not for a data centre. local llms slms voice llms llms on phones 02 ### you build agents that work with other agents. and you have felt how little there is between them: no shared handle, no trust, no common language. 03 ### you work on sharing data between agents. who may see what, for how long, with a receipt. you know it cannot be an afterthought. 04 ### you built a personal agent, and you know it is not working. it hits a login, a captcha, an app it cannot reach, and hands the work back to the person it was meant to help. 05 ### you built hacks on mac and windows so people can feed your product to ai. screen reading, clipboard tricks, accessibility apis. it works today, and you know it cannot last. 06 ### you work on agent identity, security or standards. who an agent works for, what it may do, how it proves it. you know these have to be open rules, not one company's settings. identity security standards protocols then ## we are solving the same problem in different ways. that is why we should connect. [message @anshulix on x](https://x.com/anshulix) tell me what you are building, and where it hits the wall. i read every message. soon, your agent messages mine. **>** add mcp openthings coming soon and if none of this rings true yet, **read the long take first.** if it changes your mind, let's connect. [read the long take](https://openthings.live/why/)i am back, let's connect ↑ ---