Our Picks
Our Picks
What caught our eye, outside the queue: articles, posts, threads. Hand-picked, with a line on why.
On X· @h4nkdog · Hank· September 21
A living-room wall becomes a game you control with your hand
Hank projects a game onto his wall and uses a camera to track hand movements. His post shows the demo and how he built it with Codex.
On X· @superalesha · Alexey Fateev· September 21
Impossible scenes, generated on a home computer
Alexey Fateev imagines everyday places interrupted by something surreal. He says he created the images with Qwen on his own computer and shares the prompts.
On X· Mark Zuckerberg · Meta· September 21
Ask an assistant to use services on your behalf
Mark Zuckerberg introduces Muse connectors: developers can connect their services to the assistant, turning chat requests into actions in apps.
On X· @konstipaulus · konstantinpaulus· September 21
Will Smith and spaghetti, this time drawn with code
Konstantin Paulus revisits the familiar experiment with a code-driven animation. A playful test whose rough edges are part of the exploration.
On X· @icreatelife · Kris Kashtanova· September 21
Turn your pet into a little paper character
Kris Kashtanova shares a prompt for imagining cube-shaped paper pets. These are AI-generated images; the thread explains the extra steps needed to make a foldable template.
Article· Claude Mellan · LACMA· collections.lacma.org· September 21
A face engraved almost entirely with one line
In 1649, Claude Mellan made the Holy Face with a spiral starting at the nose. The line widens and narrows to create light and shade.
Article· Michael Levin · IAI· iai.tv· September 21
What if we looked for intelligence beyond the brain?
In this essay, Michael Levin argues for seeing intelligence in organised processes, rather than only in brains. A theoretical perspective to explore and debate.
Article· Disney Research · ETH Zürich· la.disneyresearch.com· September 21
The robot that drives from the floor onto a wall
VertiGo uses two tilting propellers to press its wheels against a wall. This 2015 project from Disney Research and ETH Zürich is still striking to watch.
Article· On · LightSpray· on.com· September 21
A shoe takes shape under a robot’s spray
With LightSpray, On makes a shoe upper by spraying a filament onto a mould. The video shows the manufacturing process.
Article· MIT· news.mit.edu· September 21
A flat sheet folds itself into a tiny robot
This MIT prototype, presented in 2015, folds when heated and moves using external magnetic fields. It can walk and swim.
Article· MIT · Adam Zewe· news.mit.edu· September 21
A thin fabric that helps stop noise
In a 2024 study, MIT researchers and collaborators controlled a fabric’s vibrations to reduce transmitted sound. A laboratory demonstration, not a curtain already on sale.
Article· Harvard SEAS· seas.harvard.edu· September 21
Knitted fabric that changes shape and stays that way
Harvard researchers created textiles that snap between stable shapes. By adding conductive yarns, they also demonstrated small switches and sensors.
Article· Studio Koba · Aeternum Game Studios· store.steampowered.com· September 21
Old LCD handheld games return as a modern adventure
Future Knight borrows the look of classic LCD handheld games for a modern action game, created by Studio Koba and Aeternum Game Studios.
On X· Adaptyv Bio· September 21
Design proteins with AI and test them in an automated lab
Adaptyv and Anthropic launch a protein-design competition with funded experiments and open results. The proposed designs still have to pass laboratory testing.
Article· afosxone · Titanic animation· titanic-opus-5-2.vercel.app· September 21
The Titanic’s voyage becomes a browser animation
A code-built animation credited to afosxone in the post sharing it. A creative experiment you can explore in your browser.
Article· Figure· figure.ai· September 21
Robots try household chores in 30 unfamiliar homes
Figure shows its robots tidying, folding towels and making beds. In the company’s tests they complete 56% of tasks: a demonstration still far from a reliable household helper.
On X· Lumina· September 21
A PlayStation controller drawn with code
Lumina shows a vector-graphics recreation of a PlayStation controller made with AI assistance. A visual test shared by its creator, rather than a ranking of model capabilities.
Article· Exa · video di Donald Jewkes· exa.ai· September 21
Search the web as though it were the past
Exa Snapshot searches historical copies of web pages. Its launch video was made by Donald Jewkes, who also describes his work with AI tools.
On X· @thebuggeddev · The Bugged Dev· September 21
Embroider with your mouse, right in the browser
Stitchcraft turns cursor strokes into digital stitches, knots and threads. Its creator demonstrates the experiment and shares a version you can try online.
On X· OpenAI· September 21
An AI assistant alongside your Word document
OpenAI demonstrates ChatGPT inside Word, helping turn notes into drafts, review paragraphs and suggest edits without leaving the document.
On X· OpenAI· September 21
An AI assistant designed for lawyers’ work
OpenAI introduces Astra for Law, with tools and context for legal practice. The video presents the company’s offering as support for professional work.
On X· @chaseleantj · Chase Lean· September 21
A desktop aquarium whose fish follow your cursor
Chase Lean shows an interactive Mac wallpaper built with AI assistance. Moving the cursor changes how the fish behave.
On X· Kevin Ngo· September 21
Web photographs become little animated doodles
Kevin Ngo asks an AI assistant to choose images and draw over them with JavaScript. The video collects the results of the experiment.
On X· @mattparavel · Matt· September 21
Pokémon comes to life on a Daylight screen
Matt adapts Pokémon to a Daylight device with touch controls and a wider map view. The video shows a test session from his personal project.
Article· Fingers as Legs · autori dello studio· arxiv.org· September 21
A robotic hand walks on its own fingers
Fingers as Legs teaches a robotic hand to move and manipulate objects using the same fingers. The researchers demonstrate the prototype and describe their method in a paper.
On X· @philz1337x · philz1337x· September 16
Enhancing an image, between recovered and invented detail
Phil shares a demonstration of his upscaler, revealing increasingly sharp details as the view zooms in. Generative enhancement does not guarantee that added details existed in the original.
On X· @GoogleAI · Google AI· September 15
A conversation that also uses what the camera sees
Google introduces Gemini 3.8 Live with a demonstration of voice assistance informed by a camera feed. The announcement also outlines an Extended Thinking variant for multistep tasks.
On X· @ElevenLabs · ElevenLabs· September 14
Voice, music and video in the same assistant
ElevenLabs demonstrates an expansion of its MCP to voice, music, images and video, bringing media generation and transformation into the assistant used for work.
On X· @andrew_n_carr · Andrew Carr 🤸· September 13
Twelve animation principles to give movement weight
Andrew Carr shares an animation refined through several exchanges with Astra. Cartwheel MCP helps with foot movement, while effects and impact frames are drawn in Grease Pencil.
On X· @auteurhouse · Auteur House· September 12
Stalker sped up reveals the movement of the camera
A sped-up edit shared by Auteur House reveals the rhythm of Stalker’s long camera movements: another way to study composition and space in Tarkovsky’s film.
On X· @DavidStalmarck · David Stålmarck· September 11
Designing hardware with a modular system
David Stålmarck shares a demonstration of atech.dev for building hardware projects with less manual wiring. The link leads to the project’s original presentation.
On X· @joshmillgate · Josh Millgate· September 11
An animated mockup for trying out a phone design
Josh Millgate demonstrates placing an interface in Ultramock’s iPhone Duo mockup, complete with a blur transition. The video documents the tool’s workflow.
On X· @higgsfield · Higgsfield AI 🧩· September 11
An assistant for working inside After Effects
Higgsfield introduces Motion Designer, a plugin connecting ChatGPT to After Effects. Its demo shows changes to animation and expressions inside an editable project.
On X· @higgsfield_ai · Higgsfield AI 🧩· September 11
A simulated crowd built from individual collisions
Higgsfield’s Houdini demonstration combines running characters, collisions and ragdoll dynamics, showing how local interactions can produce chaotic collective motion.
On X· @higgsfield_ai · Higgsfield AI 🧩· September 10
A butterfly animated through 600 Photoshop layers
Higgsfield shares a butterfly animation built, according to the team, with 600 editable Photoshop layers. The post documents an Astra-assisted production experiment.
On X· @OpenAIDevs · OpenAI Developers· September 10
A voice assistant that listens while speaking
OpenAI introduces GPT-Live-1 in its API with a demonstration of two-way voice conversation. The announcement is aimed at developers integrating these interactions into applications.
On X· @higgsfield_ai · Higgsfield AI 🧩· September 10
From a photograph to a short video inside a chat
Higgsfield introduces Effects 2.0 with examples of turning a photograph into video through ChatGPT. This is a product demonstration, not a promise about results or reach.
On X· @eminimnim · nim· September 10
From concept to assembly for a small DJ controller
The author describes an experiment linking a concept, component research, a CAD model and a Blender assembly animation. The video concerns the controller’s design, not a test of a finished product.
On X· @mrjonfinger · Jon Finger· September 10
Visualizing a set with real-time AI: an imperfect experiment
Jon Finger shares a real-time visualization test and points out its limitation: in his view, it is not yet useful enough on set. The experiment is interesting for what it does not yet solve.
On X· @The_Alex · Alex· September 10
An experiment in assisted editing in DaVinci Resolve
Alex shows Astra working in DaVinci Resolve and links to an account of his method. The video is the author’s demonstration, not an independent comparison of tools.
On X· @realankitp · Ankit· September 9
A visual identity started with a single prompt
Ankit shares an experiment in visual identity made with ChatGPT Images 2.5. The images show the graphic proposal produced in his test.
On X· @higgsfield_ai · Higgsfield AI 🧩· September 9
Flowers changing shape in continuous stop motion
Higgsfield shares a floral animation generated with Images 2.5 alongside an experiment with a printed photobook, exploring the connection between digital sequences and paper objects.
On X· @samagra_sharma · samagra14· September 9
A home for interactive experiences on the web
Samagra Sharma introduces Whip Web as a place to publish and explore interactive experiences. The demo proposes an alternative to content that is only watched.
On X· @suno · Suno· September 9
Suno introduces its v6 model family
Suno announces a new family of music-creation models. The link leads to the company’s own presentation, examples and description of the project.
On X· @CreatedByJannn · Jan· September 9
A brick-built game, right down to the weather
Jannn shares a brick-style version of a game prototype made with Astra and 3D AI Studio. The transformation extends to the weather effects.
On X· @nett0eth · Nett0· September 9
A clay cat on a skateboard, in a small GIF
Nett0 shares a short animation with a stop-motion look: a clay cat in dark glasses rides a skateboard. The author describes the prompt used to generate it.
On X· @henry19840301 · Henry Li· September 9
Rebuilding a short video as a Three.js scene
Henry demonstrates an experiment rebuilding geometry, materials, lighting and motion from a video. The project includes a web demo.
On X· @TormanJeremy · Jeremy Torman (Minting Now)· September 9
Landscapes that morph with the two channels of music
Jeremy Torman shares an audio-reactive StyleGAN experiment: the stereo channels independently drive motion while landscapes change with the beat.
On X· @derrickcchoi · Derrick Choi· September 9
A pelican on a bicycle, animated in 96 images
Derrick Choi shares eight seconds of animation assembled, according to his account, from 96 individually generated images.
On X· @higgsfield_ai · Higgsfield AI 🧩· September 9
A painting where every landscape contains another
Higgsfield shares a continuous-zoom experiment: moving into the painting reveals one landscape after another. A small study in composition and visual transitions.
On X· @EHuanglu · el.cine· September 9
From storyboard to animated scene
The author combines storyboard and character preparation in Images 2.5 with video generation in Seedance. The post brings together a demonstration and notes on the workflow.
On X· @higgsfield · Higgsfield AI 🧩· September 9
Describing a character and wearing it on camera
Higgsfield demonstrates an experiment combining character generation with a real-time face filter. The video follows the transition from a prompt to an on-screen transformation.
On X· @axultan · Axultan· September 9
From a cartoon clip to a scene rebuilt in Blender
Axultan shares an experiment reconstructing a 3D scene from a short cartoon clip. The post also points to the Blender file for exploring the result.
On X· @higgsfield_ai · Higgsfield AI 🧩· September 9
When Excel cells become animation frames
Higgsfield shares an experiment using Excel cells to animate generated images, treating a spreadsheet as a visual surface.
On X· @runwayml · Runway· September 8
Runway joins the Adobe editing workflow
Runway presents an integration of its tools into Adobe workflows. The announcement video shows generated content and footage entering the same editing process.
On X· @Alissonerdx · Alisson Pereira· September 8
Sharpening a video without increasing its resolution
Alisson publishes an experimental LoRA for MiniMax H3 with a ComfyUI workflow. The author stresses that it is not an upscaler: resolution stays unchanged while sharpness strength can be adjusted.
On X· @nickfromlater · Probably Nick· September 8
An interactive model railway inside a browser
Nick shares a model railway built with Astra and Three.js. The original post also links to the interactive project.
On X· @cdngdev · thijs· September 8
A robot tries painting the Golden Gate Bridge
The author connects a robot, a paintbrush and a camera to Astra and shares a timelapse of its attempts. The experiment shows a system working on a physical painting.
On X· @cleondesigns · cleon· September 8
An imaginary shop for rediscovering MiniDiscs
Cleon imagines an Apple-style store devoted to Sony MiniDiscs. The scene is modeled in Three.js: a design exercise, not a real shop.
On X· @ty_kra_lab · Tykra· September 8
Designing a camera assembly piece by piece
Ty Kra Lab documents a parametric hardware-design prototype focused on camera and sensor clusters, with separately editable components and space requirements.
On X· @prisonmikeio · PrisonMike· September 7
A modular space station with human review of its assets
The author shares a Unity environment built with AI assistance and reviewed by a 3D artist. The account distinguishes model production from the validation still needed in the engine.
On X· @onihachi · Michiyoshi Shirota· September 7
Building a forest in Blender with Megascans assets
Onihachi experiments with building a forest from Megascans assets in Blender using a reference image. The post shows the test’s result, not a new scanning technique.
On X· @dewitim · Dewi-Tim· September 6
A modular synthesizer to play in virtual reality
Tim shares a simulation of the Moog System 55 that, according to his account, runs directly on Meta Quest 3. The video shows a prototype built over a weekend.
On X· @NURM_Dima · Dima Nurm· September 6
A father builds a platformer starring his son
Dima describes creating a personalized platformer with ten worlds, enemies and bosses. The project runs on his iPhone in developer mode; at the time of the post it was not yet published on the App Store.
On X· @shinboson · 𝞍 Shin Megami Boson 𝞍· September 6
Imagining the movement of an extinct animal
Shin Boson shares an animated Lystrosaurus study that moves from a description to a rigged 3D model. It is an experimental reconstruction, not evidence of how the animal actually moved.
On X· @kevin_t_ngo · Kevin Ngo· September 15
A fruit fly's life, animated frame by frame in JavaScript
Kevin Ngo shows a short animation about a fruit fly. The author says Claude Opus 5 drew the frames in JavaScript: the result is a creative demonstration, not a validated biological simulation.
A small visual story built with code.
Article· Kirill Mazur e colleghi· agenticstar.github.io· September 14
Reconstruct an object in 3D as it moves behind a hand
AgentSTAR reconstructs the shape and movement of articulated objects from monocular video. The project page shows the process and examples, with links to the work and code; the footage is the video shared by researcher Kirill Mazur.
The geometry of everyday objects becomes a scene whose movement can be followed and reconstructed.
On X· @toyoshi · とよし🍅株式会社トクイテン代表· September 12
Scan a missing part and print a replacement
Toyoshi shows a repair experiment: a 3D scan made with an iPhone is used to prepare a part for printing. According to the author, Astra produced the print data without needing to be explicitly told which part was missing.
AI applied to a practical task: repairing rather than replacing.
On X· @yungcontent · bloomy btw !! 🌸· September 10
To race in this game, you first have to draw the horse
Bloomy presents a racing game in which players draw their own horse before the start. The creator says Astra built it in response to a request for an entertaining website that could spread on social media.
An absurd idea that makes immediate sense: the drawing becomes the game.
On X· @Mayz1169 · Kiki· September 9
A dog turned into a small pixel-art animation
Kiki shows how a photograph of a dog was turned into a pixel-art animation with GPT Image 2.5. The original post also points to the process; the animation can be played directly here.
A personal photograph becomes a video game character.
On X· @dkundel · dominik kundel· September 8
From two photographs to a custom doorstop
Dominik Kundel needed a custom doorstop. He says he took two photographs with a measuring tape beside the object and obtained a 3D-printable design in around ten minutes; the video shows the result of his experiment.
Digital design meets one of those small problems we all encounter at home.
On X· @rohancalum · Rohan Nuttall· September 7
A photograph of a bookshelf becomes an interactive catalogue
Rohan Nuttall shows a demonstration prompted by a simple request: make a photograph interactive and identify the books. The short video documents the author's experiment.
A different way to explore what is already in front of us.
On X· @hturan · harley turan· September 6
An e-ink screen scrolling at 60 Hz
Harley Turan shows a high-density e-ink screen in a video, describing it as capable of refreshing at 60 Hz. This is a visual demonstration of the device, not our own measurement of its performance.
The appearance of paper meets the fluid motion we associate with other screens.
On X· @mathelirium · Mathelirium· September 5
Every step the spider takes changes the tension in its web
Mathelirium presents a simulation in which a spider's movements pull on the threads and deform the web. The video makes geometry and mechanical constraints visible: it is a digital model, not wildlife footage.
A few mathematical rules produce movement that is surprisingly easy to follow.
On X· @delafilms · Dela· September 5
Pillars: a short animation made in Blender
Dela shares Pillars, an animation and motion design work made in Blender. The artist states that no AI was used; the link leads directly to the original post.
A place for artists' work alongside experiments with new tools.
Article· International Commission for the Study of Clouds · Internet Archive· archive.org· September 16
Browse a cloud atlas from 1930 online
The International Atlas of Clouds and of States of the Sky, published in Paris in 1930, is available on Internet Archive. GRRL's rediscovery leads to plates and images designed to help observers study and classify the sky.
A meteorological collection that can also be enjoyed as a book of graphic design.
Article· World Labs· www.worldlabs.ai· September 1
Move the camera through a scene reconstructed from photographs
World Labs presents Atlas, a model that uses images and video to reconstruct scenes and generate new views. The demonstrations include controlled camera movements and three-dimensional scenes; unseen areas are generated, not recovered from hidden footage.
Photography becomes more than a single viewpoint: it becomes a space to explore.
On X· @iruletheworldmo · 🍓🍓🍓· September 5
may i remind people what’s happening?
may i remind people what’s happening? agi is eating everyone’s lunch so they’re fighting it with everything they’ve got because this will unlock our ability dramatically to take the value they usually charge us for don’t let them destroy our brave new world chat
On X· @BreathingByEd · Ed· August 25
Micro-Dosing Breathwork is The Future Of Nervous System Regulation
When most people think about breathwork, they picture one of two things. A 20-minute morning session on a yoga mat, or one of those Instagram reels you see where individuals are having an intense
On X· @0xQuasark · Paul Brown· August 24
The UK's first publicly funded psilocybin trial just dropped.
The UK's first publicly funded psilocybin trial just dropped. They recruited 60 people with depression who already failed at least 2 treatments. Half got placebo. Half got one 25mg dose of psilocybin. Six weeks later: 50% of the psilocybin group met the threshold for treatment response. Placebo? 𝟯%. From one dose! People whose conventional treatment had already failed...actually got relief. This is exactly why this research deserves to be taken seriously.
On X· @Polymarket · Polymarket· August 22
JUST IN: U.S. researchers successfully transmit quantum information wirelessly through open air over 13 miles, advancing efforts to build…
JUST IN: U.S. researchers successfully transmit quantum information wirelessly through open air over 13 miles, advancing efforts to build a future quantum internet.
Article· ScienceDaily· sciencedaily.com· September 6
3 minutes of sprinting does something 90 minutes of exercise does not
A few minutes of sprinting may reshape the bloodstream far more dramatically than a much longer moderate workout. Six 30-second sprints altered nearly a quarter of the blood proteins measured and more than 200 metabolites, while 90 minutes of moderate cycling produced far smaller immediate changes. Many of the proteins affected by sprinting were also associated with lower risks of obesity, type 2 diabetes, and other metabolic diseases.
On X· @alexeixbt · alexei· August 8
Neuroplasticity: How Repetition Quietly Decides Who You Become
Somewhere out there is a version of you that's wealthier, calmer, more confident, better with people, more disciplined, less afraid. And here's the part almost nobody explains properly: the gap
Article· News-Medical· news-medical.net· August 18
Music activates brain networks for imagination and meaning
In a new paper appearing Aug. 18 in Nature Communications, Princeton University researchers show how music produces distinct patterns in the regions of the brain associated with imagination and meaning, similar to those prompted by spoken narratives.
Article· Scientific American· scientificamerican.com· September 6
How psilocybin alters consciousness to create a sense of ‘oneness’
Psychedelics can create a new “hidden order” in brain activity that could explain how these drugs alter our sense of self