Posts with #Interface
#Interface
Public posts on X that use #Interface.
Related hashtags
Not sure which desktop operating system is best for you? We help you choose by comparing the top options based on app compatibility, ease of use, hardware availability, interface design, gaming support, and other key factors.0011K
Your AI agents now have their own little screen on your desk. Meet Harness Pro by Autonomous. A dedicated desk display that lets you monitor and control AI coding agents running across multiple computers. The interesting part is that it uses an ESP32-P4 rather than a conventional mini PC. It has a 3.95-inch, 720 × 720 touchscreen, three microphones and a speaker, all inside a compact CNC-machined aluminium enclosure. It connects over USB-C to a computer running the Harness app. You can switch between Claude Code, Codex, Cursor and other agents, see what they're doing and tap the screen to give them voice instructions. Even if you close the app or the device, your agents continue working on their machines. What I particularly like is the hardware design. The ESP32-P4 handles a dedicated interface for something that normally lives inside a developer's terminal. Building physical interfaces for AI agents is worth to check out for.157262.1K
@ reachartwork on tumblr & bsky@reachartworkCamouflage has no effect on those that see in black and white #interface #interfacefanart
429176
- @OpenAITwo OpenAI team members worked with GPT-6 Astra to design a prototype device for talking to Codex in under a week, as a side project. Astra designed the enclosure, circuit board, interface, and software. Getting it manufactured still required several rounds of review and iteration.564567759.4K
Samsung TV, but with an Apple TV-style interface!🔥 An open-source launcher for Tizen OS that replaces Samsung's default home screen with a clean tvOS-inspired design. It brings a glass dock, smooth app navigation, Control Center, 4K aerial screensavers, Continue Watching and even an optimizer to remove ads and unnecessary services. One catch: it requires a rooted Samsung TV with Vetro. GitHub link in comments.
Aman @mdaman010Native TV: A 60 FPS Launcher for LG webOS TVs An open-source native launcher built with C++ and OpenGL ES for rooted LG TVs. - Smooth, Google TV-inspired interface - Custom app and HDMI shortcuts - Home Assistant integration - OLED controls and monitoring - Ad and telemetry blocking - Phone and desktop dashboard No browser, JavaScript, cloud account or data collection. Tested on LG B3 (webOS 10.3.2). Other models are unverified. GitHub link in the comments.471008.3K- Native TV: A 60 FPS Launcher for LG webOS TVs An open-source native launcher built with C++ and OpenGL ES for rooted LG TVs. - Smooth, Google TV-inspired interface - Custom app and HDMI shortcuts - Home Assistant integration - OLED controls and monitoring - Ad and telemetry blocking - Phone and desktop dashboard No browser, JavaScript, cloud account or data collection. Tested on LG B3 (webOS 10.3.2). Other models are unverified. GitHub link in the comments.
71117652.1K 
i've helped startups raise millions through design decisions, skills & taste. I've decided to now curate all of my work over the past year at various startups into a skill (tactile-design) it teaches your agent how and when to add tactility, animations, how to use layouts and whitespace. it creates a premium and curated interface feel.10630710.5K
MOTION DESIGN WITH OPUS 5.5 Everyone on X is making videos with Opus 5.5 lately. But the real skill isn't writing a clever prompt. It's knowing how to brief the model, use references, and build a workflow that consistently produces good results. Here's a mini-guide to the tools and techniques I use. 1. FIND VIDEO REFERENCES AND PROMPTS Skillry 418 videos with prompts. You can compare the original with the remake and see what the prompt actually produces. It also has 359 ready-made video skills for specific formats, from product launch videos to logo reveals and title sequences. awesome-opus-5-5-prompts A collection of prompts focused on 3D, gaming, and social media videos. You don't need to copy them 1:1. Use them as visual references and starting points for your own concepts. 2. TOOLS FOR CREATING VIDEOS motion-graphics-skills A motion design skill that can turn a website, screenshots, or your own clips into a finished MP4. What it does: Develops 3 concepts and lets you choose a visual direction Reads your brand's interface, logo, colors, and fonts from a website Creates animations with springs, camera movement, motion blur, and 60 FPS rendering Edits your clips to the beat Generates original music for each video Reviews and improves its own output From my experience, this skill can make your videos significantly more dynamic and interesting. lemo-opuscar A collection of cinematic styles with demo videos generated using Opus 5.5. Useful for exploring the visual styles you might have seen in Reels or TikTok, including mini-series formats. HyperFrames A framework for creating videos with HTML. I initially assumed Opus used HyperFrames by default, but apparently it doesn't. Worth checking out if you want to build HTML-based videos. 3. HOW TO GET BETTER RESULTS The most important part is the brief and reference. If you just give Opus a random prompt, you might get a good result, but the odds are much lower. Here's what I recommend: Give it a reference video and a detailed brief. You can use another AI to help prepare the brief. Ask it to break the reference down into individual components and assets before recreating it. Ask for 3–5 rendered frames and have the model score the hook, readability, motion, composition, and visual bugs from 1 to 10. Keep reminding it about the original references, especially during revisions. The first iteration might have distorted logos, wrong colors, or elements that don't match your brief. Don't just accept it. Point out what's wrong and make it fix those details. The reference is what keeps the model on track. And the brief is what turns that reference into the video you actually want.Fokki @0x_fokkiSomeone needs to stop Claude Opus 5.5. It turned one picture into a robot octopus I can take apart piece by piece. i just said "make the model" and what came back floored me. the build: > the picture set the look, every petal, lens and link is three.js code > opus rendered one still next to the picture and fixed the gaps for 3 rounds before it touched the film > the first pass looked like a toy, so it went back and packed 202 links and 480 wheels into 8 arms > every arm is one draw call, posed every frame by walking its chain of bends > something opens every 2-3 seconds: the lens unscrews, the mantle hinges, the pump rises out spinning > at 0:16 the whole thing blows apart on a 330° turntable and reassembles before the turn ends 25 seconds, 2,616 lines of code and one octopus that leaves the frame on its own. the skill, the prompt and the live sim 👇103226
Someone needs to stop Claude Opus 5.5. It turned one picture into a robot octopus I can take apart piece by piece. i just said "make the model" and what came back floored me. the build: > the picture set the look, every petal, lens and link is three.js code > opus rendered one still next to the picture and fixed the gaps for 3 rounds before it touched the film > the first pass looked like a toy, so it went back and packed 202 links and 480 wheels into 8 arms > every arm is one draw call, posed every frame by walking its chain of bends > something opens every 2-3 seconds: the lens unscrews, the mantle hinges, the pump rises out spinning > at 0:16 the whole thing blows apart on a 330° turntable and reassembles before the turn ends 25 seconds, 2,616 lines of code and one octopus that leaves the frame on its own. the skill, the prompt and the live sim 👇1746536.8K664K
Harness engineering is the most critical and accessible shift in AI development right now researchers audited the raw production source of 11 major coding agents (Claude Code, Codex, Gemini CLI, OpenHands, Aider, Hermes, OpenCode, Pi) the central equation: agent = model + harness the model provides raw intelligence. the harness is the execution machinery that turns predictions into working software. every single production system in the study converges on 7 foundational subsystems: → agent loop: the execution engine. evaluate, trigger tools, inspect output, repeat until resolved → LLM integration: wire transports, strict prompt-cache boundaries, and runtime routing → tools & actions: the physical interface. AST-based file edits, workspace operations, shell access → memory & context: history rationing, persistent context, and threshold compaction → safety & permissions: policy as data, interactive approval gates, and OS-level sandboxing → orchestration: hierarchical sub-agents that partition heavy workloads across tasks → extensibility: lifecycle event hooks, skills, and plugins that preserve core stability what the 4-million-line audit revealed: > zero agentic frameworks: not a single production codebase uses LangChain, LangGraph, or AutoGen. every team hand-rolled native async loops in Python, TypeScript, or Rust > zero vector RAG: no one uses vector databases to search code. retrieval is 100% deterministic via ripgrep, tree-sitter AST, and discoverable Markdown context > skills overtook MCP: file-based SKILL.md bundles now lead MCP in real-world adoption (9 of 11 vs 8 of 11) for engineers building agents: the paper extracts 18 battle-tested design patterns and a 90-line minimal Python harness that implements 10 of them out of the box. 90 lines of clean code. that is the entire barrier to entry. save this, full paper breakdown below ↓damkin @damkina7172291.3K
May 2007: Apple had sold 100 million iPods, the Zune not yet 1 million. Bill Gates says the Zune team loved Apple’s “gigantic market.” Steve Jobs sat beside him with an unreleased iPhone. In the video below Gates tells it himself: "In a few cases like the Zune, that if you go over to that group, they think of Apple as a competitor." He says they're going to "come in and contribute something." Asked what he carries, Jobs answers a MacBook Pro "and an iPhone," a month before launch. Apple announced its 100 millionth iPod on April 9, 2007, after selling 10.55 million from January to March. The Zune went on sale in November 2006, and Microsoft expected its first million by the end of June. By the 2008 holiday quarter Zune revenue was down 54% on a year earlier. In one 2010 quarter Apple sold 14.1 million iPhones and 9.05 million iPods, the iPod down 11% in a year. In October 2011 Microsoft stopped making Zune players. Apple sold about 116,000 iPods a day in early 2007. The Zune's first million equals about 9 days of that. Zune revenue: about $185 million to $85 million in that holiday quarter. The case for the Zune: its interface grew into Metro, the design Microsoft carried into Windows Phone, Windows 8 and the Xbox. 100 million iPods when Gates called it a gigantic market. Four years later the Zune was gone and the phone Jobs carried that day outsold the iPod. When a rival builds a gigantic market, do you fight for a share of it, or build what replaces it?Tema Ultra @cryp_cryp2013958
I'm taking calls today from medical professionals who want to test the Interface app (my startup) for offering #Coronavirus advice by video. Price set to zero. My understanding is that practicing across state lines is approved during the crisis. #Interface #WhenHub1486401
these UI component libraries can save your hours of design and development. there are libraries filled with ready-to-use components, animations, and interactive elements you can drop straight into your projects. here are 9 websites worth bookmarking: 1. Board UI beautiful UI components for building modern digital products. → 2. a collection of ready-to-use React components you can discover, customize, and add to your projects. → 3. Aceternity UI animated react and tailwind components for creating visually impressive websites. → 4. Origin UI free, open-source UI components built for modern React applications. → 5. Transitions a collection of UI transitions and animations to make interfaces feel smoother and more interactive. → 6. Reverse UI distinctive UI components for developers who want something beyond the usual generic interface. → 7. BeUI animated react components you can use to add motion and personality to your interfaces. → 8. Kinetics 153 spring-powered animations that bring movement and responsive interactions to your UI. → 9. Astryx a collection of UI components and building blocks for creating modern interfaces. → the real advantage? if you’re building websites, dashboards, SaaS products, or modern web apps, stop creating every component from scratch. find a component you like, adapt it to your design system, and focus your time on building the actual product. save this list for your next project. what other UI component libraries should be on here?58573K
UI/UX Designers, did you know you can design directly on top of your real codebase without going back and forth between Figma and your developers? I uploaded a dashboard codebase from GitHub into Modeinspect, and it instantly turned the existing code into an editable visual interface. Then I tweaked the layout, spacing, and overall look without going back to Figma or waiting on a developer. Honestly, this is how product design should work.754120.8K
ithub.@ithubdirectoryUX/UI Agent Skills UX/UI Agent Skills provides design tokens, component workflows, accessibility checks and interface review guidance for AI coding agents.00238
Intelligence needs a user interface. At @designerfund, we're backing @trajectorylabs alongside Sequoia and NVIDIA to build it. We believe continual learning is one of the most interesting design challenges in AI. We've spent decades building intuitive tools for editing images, documents, and software, but not model behavior ↓Ronak Malde @rronak_Imagine if everyone had the tools to shape intelligence itself. That day is today. Today we launch the Trajectory platform - the easiest and most intuitive tool to post-train models of your own. The foundation for continual learning. With just a few agent prompts, you get access to a full creative suite to train, eval, and deploy trillion parameter models. On top of that, we believe in a future where you don’t just shape your model once, it also keeps learning on its own from the world around it. We’re releasing two research previews. 1. Continual grading scales your taste, so your model learns what good looks like from far more than you could ever grade by hand. 2. GS1, a new algorithm that lets models learn directly from what they do in production. We’ve now raised 60M, with a new round led by @Sequoia at a 300M valuation to democratize the most powerful technology of today - training models - so that the lawyers, doctors, and engineers of the world can own their own intelligence. You know what good looks like, now go build it. Own your own value function with Trajectory.34311.9K
Imagine if everyone had the tools to shape intelligence itself. That day is today. Today we launch the Trajectory platform - the easiest and most intuitive tool to post-train models of your own. The foundation for continual learning. With just a few agent prompts, you get access to a full creative suite to train, eval, and deploy trillion parameter models. On top of that, we believe in a future where you don’t just shape your model once, it also keeps learning on its own from the world around it. We’re releasing two research previews. 1. Continual grading scales your taste, so your model learns what good looks like from far more than you could ever grade by hand. 2. GS1, a new algorithm that lets models learn directly from what they do in production. We’ve now raised 60M, with a new round led by @Sequoia at a 300M valuation to democratize the most powerful technology of today - training models - so that the lawyers, doctors, and engineers of the world can own their own intelligence. You know what good looks like, now go build it. Own your own value function with Trajectory.2061211.3K791K
Ok, so initially I thought that what happened here was like, distillation of the human brain. So, the same way open source models have been (allegedly) deriving weights by treating a more advanced model as a black box (feed data in, observe the output and tune), we treat the brain as a teacher/black box and train a new model. That sounds fascinating & crazy & futuristic That's not whats happening here. Instead, they've studied how electrical signals propagate through neurons and used those findings to design a small neural-network adapter for an existing video model. They then train the adapter's weights on regular video data. So you can't really say the model they end up with has any of the intelligence inherent in the brain. Its kind of just like adding a biologically inspired computational component to an existing model. Which is still pretty cool! So then I was like, well is anyone thinking about the brain as a distillation source? Turns out yeah that is a topic of research. But the problem is, distillation of a brain would be extremely slow and difficult. Distillation works with LLMs because you can hit them with millions and millions of inputs and collect their outputs automatically. We don't have an equivalent interface for a living brain, especially if we want to observe its neural activity rather than just what a person says. Also, networks of brain cells grown in a lab haven't developed the complex connections or accumulated the decades of experience of an adult human brain. Even if we could perfectly reproduce their behavior in software, we wouldn't necessarily be transferring any human intelligence. I suppose what it comes down to is learning how biological neurons compute and learning what an intelligent brain knows are very different problems.Alex Ksendzovsky @AlexKsendzovskytoday, we introduce a little mischief to the world. for four years, we’ve been brewing real brain cells on electrodes to understand how they compute. now, we’ve turned those discoveries into a video model anyone can use. from 0.5¢ a second, mischief 1 is at least 5x faster than its base and the cheapest image-to-video model on @fal and @OpenRouter. soon to be on @awscloud marketplace. it’s not a delusion anymore. biocompute is real, commercial and this is just the beginning. happy halloween.002186
today, we introduce a little mischief to the world. for four years, we’ve been brewing real brain cells on electrodes to understand how they compute. now, we’ve turned those discoveries into a video model anyone can use. from 0.5¢ a second, mischief 1 is at least 5x faster than its base and the cheapest image-to-video model on @fal and @OpenRouter. soon to be on @awscloud marketplace. it’s not a delusion anymore. biocompute is real, commercial and this is just the beginning. happy halloween.8541.1K12.5K3.8M
4...3...2.... PC interface proposition! Rate my desirability on your screen. #interface #GameDesign #MMORPG
7130176- Show more posts
