ckt

AUIs are how agents become the users, we become the overseers, and websites disappear — making the internet headless. GenUI + Agent + AIPs = AUI Here's the AUI-driven architecture shift happening now:

Thank you sir! Will def check it out. A couple of my favorites are: @TheRewatchables (with Bill Simmons, Chris Ryan, and Sean Fennessey), and Scriptnotes (with Craig Mazin and John August). Great ways to (re-)experience films from two different, but equally informed & impassioned perspectives.

. @ckttam here's the podcast on films and TVs I mentioned to you yesterday, @filmspotting . It's my favorite podcast on films. Adam and Josh always deliver a great two-man show. Their chemistry is enviable.

Thanks to @questflow & @CoinbaseDev for hosting an amazing event! It was great sharing what we're working on at @substrates_ai with an enthusiastic, knowledgeable crowd!! s/o to @programmer @Hassan_NY @Bobbxu for great demos and MC duties as well!

Excited to share some of our latest @substrates_ai demos and developments with the x402 community at @questflow's SG event — swing by if you're around!

@BackseatVC Absolutely; been working on this for a while now as well, happy to chat. Will reach out.

Stripe is building a new prototype in the stablecoins and agentic payments space If you're thinking about how to charge agents for API requests, tool call, page crawl, etc, I'd love to chat through some early design feedback Please feel free to DM or email us at stablecoins@stripe.com

While this is a neat framework, it seems a touch too simplistic — in particular, I think it's more important NOT to open unnecessary loops to begin with, AND to ignore worthless ones. Too many people spend their whole days (and lives) opening and closing meaningless loops. It's also important to learn how to deal with having important but long-term loops continuously open, while still making consistent progress toward closing them eventually. And how to decompose large loops into smaller loops that you can close more easily. So while the point about open loops draining your energy rings true, the solution is more about prioritization (open fewer loops, ignore worthless ones) and management (compartmentalize long-term loops, decompose larger ones) than about just closing every open loop in your life.

@Viggy_117 Would love to try it out!

Does anyone want EigenAI access? I have some API grants to give out. If you can use the OpenAI API you can use the EigenAI API. The difference is we can help you guarantee a system prompt and model behavior to your users so they don't get rugged. DM or Comment if you are interested.

Create facts. Not theories. Theories come *after* facts. To build the future, you must discover, generate, and create new facts. Not just explain facts with theories. Take action. Let them analyze what you've done after. Create facts. Build the future.

Haven't tried @strawberry yet, but conceptually it seems a step closer to achieving full AUI by being able to co-opt Sheets as a generative interface for this type of task result. I suspect you still have to specify that yourself, but hopefully their agent will soon be able to choose the output interface on its own based on context. One thing I'd note is that their tagline is the "self-driving browser" (which is great), but it still seems to be the type of self-driving browser that an agent drives with robotic arms — pointing and clicking at GUIs, albeit automated. The car itself is still designed for humans. Again, it's likely the right transitional form, but hopefully they can leverage more agent-native interaction protocols like MCP and A2A to make their browser truly agentic.

@0xDesigner For sure. They want to become (and own) the gateway to the internet. They already have the generative capabilities for a "generative UI". Now they want to own the act of browsing, as well. Agentic user interfaces incoming.

oh my god. chatgpt atlas isn't about computer use. starting a search from the URL opens a chat and native search results. they're trying to takeover search.

@boringmarketer @gregisenberg Giving agents hands to use GUIs designed for humans is like giving self-driving cars robotic arms to steer. Agentic Interaction Protocols (AIPs) like MCP and Google A2A are the future. More like brain interfaces for agents. Websites are obsolete; the future is headless.

@gregisenberg hands need buttons and tools vs content heavy sites - structured api documentation - helpful q&a content - technically optimized sites big shift coming to what a "good" website looks like to an agent

MY TAKEAWAY FROM TODAY'S OPENAI AI BROWSER LAUNCH: the internet just got hands. the average person won’t google, click, compare, or fill out forms within the next 24 months. they’ll just say “book my trip”, “find me a job”, “launch my store” and the agent will do the 20 steps behind the scenes. that means whole industries... travel, e-commerce, real estate, insurance, education are about to get rebuilt around outcomes instead of pages. you won’t “go” to Expedia, you’ll just get the trip. if you’re a founder, this is the moment to think in verbs. don’t build platforms people visit. build agents that finish what people start. that’s where the next $100B companies come from. the web is shifting from human browsing to agent doing. internet hands.

@gregisenberg Absolutely agree that Agentic User Interfaces (AUIs) are taking over. Today they have hands, but in future they'll have interaction protocols more akin to brain interfaces for agents. And then the internet will go truly headless. (And our anthropomorphic metaphors may run dry)

@OpenAI Quickly moving towards a future with Agentic User Interfaces (AUIs), where websites are obsolete, agents are the users, and the internet is headless.