On paper, I’m a visual designer working mostly in digital products, but my curiosity has always wandered well beyond design. I formally studied visual arts and visual communication design (and also spent time studying psychology and sociology, although left both degrees unfinished). Formal credentials have never meant that much to me, although I can imagine finding my way back to academia eventually, mostly for the pleasure of studying something properly again. In any case, I never really stopped studying those subjects. Psychology has stayed with me ever since, alongside a long-standing interest in philosophy and neuroscience. I think this is what I’ve been trying to do all along, although I never had a good way of describing it. I want to move from a creation myth to a political structure, from a religious ritual to a cognitive bias, from perception to visual art, from evolution to social behavior, from mathematics or physics to the assumptions they force us to make about reality. Not because I expect to become an expert in all of them (that would be absurd). I just want to see the whole thing on one endless sheet... I want to know where the lines cross. Where one field has been answering a question another field asks in a different language. Where the borders between disciplines exist mostly because universities needed departments. Maybe this is why narrow specialization has always felt slightly claustrophobic to me. It’s not that I don’t want depth. I do. Depth is what makes the connections worth anything. I just don’t want to spend my entire life looking at one square of the map and pretending the edges of that square are the edges of the world.

I know I share a little too much about @rive_app as if it was my own company but c'mon... WHAT THE HELL IS THIS?!🤩

Confession: For the longest time I thought connecting Gmail to an AI agent was one of those “look how productive I am” things. I work solo, my life is not that busy and my inbox is not exactly a Fortune 500 crisis center. Why would I need this? THEN I CONNECTED MY GMAIL TO MY DOT AND OMFG. Apparently I completely misunderstood the use case!? This might genuinely be one of the most useful things I’ve tried for my AuDHD brain.

I honestly don't understand many new design tools. There are Pencil, Paper, Elyx, MagicPath, and many more launching each month. Great designers are behind them, and it feels like many people use them. But I feel so disconnected from their approach to design. Let me share my thoughts. They all have a middleman issue. What you design is not what you ship. We're in the age of AI. I want to see the exact behavior of my components, I want to see the exact "rendering"/"look" of my components. Why do I need a canvas that shows me just "designs"? We had that already in Figma for many years. Authors of those tools can argue with me, saying, "But we're much closer to the code than Figma. We use HTML/React or whatever to render your designs." Great, but it's not the same code anyway. From that issue comes another one, a lack of a single source of truth. Design lives in two places again. Users see the one that is built in code, while designers look at their versions on the canvas. Then you can tell me, "Come on, I'm a designer, not a developer. I need a canvas. Those apps give me a canvas." Do you know that you can ask AI to bring you a canvas into the actual project you're working on? "But building code is slow, I need something simpler just to quickly iterate on my design ideas." Again, what's the problem with asking your Claude/Codex to solve that thing for you? AI can stop running tests/linters/performance checks, or simply make some drafts on that canvas for you in plain HTML/CSS, exactly like those tools do. But if you want to do something more complicated, nothing can stop you. Which brings us to another point. Those are closed-source tools. I like to adapt my environment to my needs and my tasks. I don't want to wait anymore for companies to decide to prioritize or build something that's important to me, or support the AI client that someone uses (recent Figma drama). Software development has become cheap. Bringing any controls or canvases into your environment is not an issue if you need them. I still use Figma, more like a sketchpad or for some brand work, but I don't understand why I would want to switch to something else if those tools fundamentally have the same problem as Figma. For me, those tools still feel like they're made for designers who haven't figured out yet what AI is capable of. And I'm not talking about letting AI design instead of you. I'm talking about letting AI create the environment for you, which can be much more powerful and more efficient for the exact tasks that you or your team work on.

This is the exact philosophy we have in mind when building @lunagraphHQ. I’m a designer and an engineer, running a startup, and running a design studio, so we’re very much on the same page. Let me explain: "letting AI create the environment for you" What we’re building at Lunagraph is a “programmable design canvas”. Exactly as you said, most of these tools are walled-gardens, if they don’t make a feature, you can’t build the experience. If Figma didn’t build motion, you can’t do motion design. In Lunagraph, you can build any experience that you can build with code, make your own controls, handles, anything you need. You can use any React package, like three.js, or motion dev, and cook without waiting for a feature release. You can make a metallic 3d card with knobs for roughness, metallic-ness, and handles for lighting and for the svg logo. You can literally build Lunagraph inside Lunagraph. We are essentially here so you don’t have to remake these primitives every time. PLUS we’re also building the best UX to work on this programmable canvas with your own agents. And now let’s talk about “source of truth” I think everyone has been wrongly fixated on this. Yes what’s in prod is what ships. But what about the intention? Say you’re redesigning an app, or making a new feature: there’s this point in time where there are 2 truths, by definition: > Your intended design is the intended truth > Your current prod is the current truth Eventually you want both to be the same, the intention implemented in the code. But in this point in time, there are these 2 states. So where should the design intent live, for that period of time when it’s not yet implemented in the code? No, not *always* a PR. If you’re designing a new feature, and make a PR in the codebase, would you also be responsible to wire it up, make it truly work? Or would someone else do that part? If it’s someone else, I promise you as a dev making the implementation, I prefer I make the PR and grab your designs piece by piece instead of getting a huge PR I have to unravel and re-wire. And if it’s yourself, it’s so much nicer and more efficient to iterate on the design without your coding agent trying to make it work before you know how you want it to work. This is where Lunagraph comes in. It’s a place to do the big chunk of the design work, all the way to making a PR. - Explore, drawing board, discuss - Start formulating an intent - Make prototypes, test, refine - Iterate, share, get feedback - Finalize - Then convey your intention = to your coding agents, to devs, or to yourself. And yes it’s made of React, so when you “convey your intention” you can choose to go as precise as you want: - At 50% fidelity: you can share static designs and let your agent or your devs interpret the UX and flow - At 90% fidelity: you can build a component, complete with the UX, interaction, transition, states, and how it takes in data based on Typescript data type. - At 100% fidelity: you can implement it yourself and make a PR. We’re definitely making Lunagraph for design engineers who can utilize AI in designing to the max. Many of us are still exploring designing with code with AI, so me and @Riyvir are working hard on distilling our own workflows and hundreds of user feedback into tools anyone can use right away, and building the best way for teams to add their own tooling on the same canvas.

OpenAI’s Dots lead on the future of ChatGPT @embirico joins the show from OpenAI DevDay to discuss the new always-on personal agent in ChatGPT. I ask him how much control people should give these agents, how OpenAI approaches permissions and privacy, and why Dots ask before taking action by default. He also explains Space, a new workspace where people and agents can collaborate on documents, data, and projects. We get into how OpenAI plans to price an always-available assistant, how these capabilities could eventually reach free ChatGPT users, and the story of an employee who woke up to discover his Dot had helped fix a bug. Timestamps: 00:00 Why agents now 00:48 From chat to active intelligence 02:46 Alex’s origin story 04:29 Building Dots 06:07 What Dots can do 08:54 Why embody an agent? 11:41 One agent or many? 15:48 Spaces and trust 24:07 Safety and enterprise controls 30:48 Building the Dots ecosystem 34:03 Pricing an always-on agent 38:01 What’s next for Dots Links: podcast.sources.news/ Spotify: open.spotify.com/show/3p1KZv… YouTube: piped.video/VnHOFQE7nCw Substack: sources.news/p/openai-dots-l… Thanks to the show’s premiere sponsors: @Atlassian, @meetgranola, and @mercury.

Okay but why does my dot has Solitaire on his desktop🫠

How did I bookmark this and then completely forget about it?! Such a sick website for learning motion design principles for UI/UX🔥🔥🔥 motion.zajno.com (by @ZajnoCrew)

As usual, Rivers are Riving 🔥

The Rocket Volvo (Volvo P1800S Rocket) designed by Pietro Frua (1968)

you’ve heard of the for you page, now get ready for the Against You page

Wait, this one might actually be useful👀 Like, useful-useful.

my dot doesn't even pick up the call, im autistic and this is literally the one feature i need rn..

I miss brands having actual character. Make branding fun again.

I can't believe this is a government website what the fuck.

Okay but will your Dot interrupt you mid-swim for a kiss?🫠

And in the end, what’s all this for? Speed. Yet somehow, the time we save never seems to become ours. We simply pour it back into doing more, faster. Why? So wealth can accumulate faster in the hands that already have it. Sad.

Rive's full text engine features can be controlled via scripting. Text shaping, line-breaking, with full support for internationalization, styled runs, BIDI, etc! Demos inspired by some cool pretext ones here pretextjs.dev/pretext-demo/s… @rive_app

been having fun making my naives shake their heads

I find it very strange that people are becoming almost obsessively curious about artificial intelligence while remaining remarkably incurious about their own? A question that has haunted me since I first started thinking seriously about anything is... How can a human-being not be curious about the human brain? It is the instrument through which we perceive reality, form a sense of self, ask physical and metaphysical questions, create art, build civilizations, fall in love, invent gods and now, apparently, artificial minds. Everything we know about the world reaches us through this biological machinery, including every question we ask about what lies beyond it. And yet somehow the artificial version seems more fascinating to many people than the original source of intelligence doing the fascinating. Weird...