Silly Robots Contest is a San Francisco event where the goal is not usefulness or commercial value, but simply making people laugh. Entries are invited to be whimsical, silly, funny or completely unnecessary, with the event scheduled for September 10, 2026.
The call explicitly welcomes “cursed kinematics,” “unnecessary degrees of freedom” and “beautifully overengineered mechanisms,” celebrating robots that move in the opposite direction from conventional performance competitions.
A robot contest judged by “is it funny?” instead of “is it useful?” feels like a very healthy use of engineering.
A small robot car has been built to detect a face with its camera, turn until that face is centered and then drive toward the person it sees. The project upgrades an Elegoo robot car kit with an Arduino UNO Q and webcam, using computer vision to locate faces and adjust steering. It was created by Iulia Feroli and featured on Arduino's official blog.
Face detection is familiar on a screen; connecting the detection result directly to motion gives the same software a much stronger physical presence.
Face detection feels harmless on a screen, then you add wheels and it starts closing the distance. Giving AI mobility really changes the mood.
A small rechargeable air blower meant for cleaning desks and keyboards was taken apart to test whether its internal EDF (electric ducted fan) could propel an RC airplane. The fan and ESC form a convenient compact package, and testing reportedly produced up to about 110 grams of thrust at 16.5 amps.
One unit may not provide enough thrust for the intended model, so using two is also being considered. The project treats a consumer gadget less as a product and more as a convenient container of reusable components.
This is the mindset where you see a product and immediately ask “what's inside that I can repurpose?” The blower probably didn't expect an aviation career.
During a cheerleading performance at the World Humanoid Robot Games, roughly ten humanoids were dancing in sync when one suddenly collapsed. Its control sequence apparently kept running, because the fallen machine continued moving its arms and legs from the floor.
The other robots carried on upright while staff approached to retrieve it, leaving one participant performing the same routine at an entirely different height.
The body has fallen, but the program is still cheerful. It's like nobody updated the robot's situation.
Wheeled delivery robots struggle with stairs, curbs and thresholds. R-dog tackles those obstacles by using a quadruped body that can step over them, while also carrying deliveries and potentially displaying advertising on its sides.
The current product page adds another layer: pet the robot's head and a sensor makes its tail wag. The dog-like body is not only functional; the design also commits to dog behavior for charm.
Four legs make practical sense, but then the design continues to “well, a dog should wag its tail.” That's serious commitment to the role.
ClankerVids is a site that collects large numbers of robot-failure videos into categories such as Fails, Humanoids Eating It, Robot Dogs Being Weird and Factory Chaos. It also runs a daily Robot Digest and trending sections.
Instead of treating robot fails as clips you happen to encounter on social media, the site organizes them as a video genre to search, categorize and browse intentionally.
Go looking for robot fails and discover the genre already has something like its own channel. Apparently the audience is real.
Camera with Digital Film lets you physically pull the captured image out of the camera as a cartridge. A Raspberry Pi 3 A+ and 5 MP camera write each shot to a removable module containing a 1.54-inch four-color e-paper display.
The cartridge can then be displayed as a physical photograph. Because it is e-paper, the image remains visible without power and can later be overwritten. Modern digital hardware brings back one very specific analog pleasure: taking the picture out of the camera.
Digital cameras spent years eliminating film, and now the joy of pulling the picture out has come back around. A very elegant detour.
Larger robot vacuums can struggle to fit beneath furniture and into other tight spaces. Roomba Duo tackles that problem with a large cleaning robot that carries a smaller companion robot and lowers it to the floor using a mechanical arm on top. iRobot presented the two-robot system as a concept at IFA 2026.
The main unit handles regular vacuuming and mopping, while the roughly 8.35 cm-tall companion can enter places the larger robot cannot. Both share the same home map, dividing the job between two machines of different sizes.
Robot vacuums now have something like parent-and-child roles. The appliance is starting to look more like a tiny team.
Library for Ants is a storage device where a tiny robotic mechanism physically carries 12 microSD cards and inserts them into a reader one by one. Instead of switching solid-state storage electronically, it deliberately moves the media itself.
Maker svofski describes it as a robot that wants to grow up to become a tape library. It is essentially an old large-scale storage behavior carefully transplanted onto one of today's smallest memory formats.
Instead of using microSD because it's tiny, the project uses that fact to say “great, a tiny robot can carry it.” Wonderful reversal.
Learning biped robots are going to fall. With a full-size humanoid, each failure can mean more damage to the machine and its surroundings. Microduck approaches that problem by being a roughly 25 cm-tall, sub-800 gram biped small enough for falls to be much less dramatic. Built by Pollen Robotics as a research and development platform, it carries 15 motors, a camera, LiDAR and IMUs and is intended for sim-to-real learning workflows.
It can walk, kick, grab objects with its beak and get itself back up after falling. It can even roller-skate, choosing manageable scale rather than pretending experimental robots never fail.
It's a research robot where a failed experiment can stay roughly desk-sized. Being a duck also makes the tumbles look a little friendlier.
If a toothbrush is going to target the gaps between teeth precisely, it first needs to see exactly where the liquid should go. CameraJet tackles that with a roughly 100,000-pixel macro camera, machine-learning-assisted gap detection and a targeted liquid jet. Dyson introduced the device in September 2026, and a live view of the mouth can be displayed through the MyDyson app.
To keep the camera's view clear, the system also uses specially designed low-foaming toothpaste. Adding imaging, recognition and liquid control to brushing turns even the toothpaste into part of the system design.
At this point, even the toothpaste feels like part of the hardware stack. Oral care is starting to look like systems engineering.
Whenever birds call outside, Fugleramme uses AI to identify them and displays only recently heard species as 19th-century illustrations on an e-paper frame. The Norway-based project runs BirdNET-Go on a Raspberry Pi 5.
Modern audio recognition works behind the scenes, while the visible result looks like an old natural-history book. Digital detections are transformed into a quiet, daily-changing bird guide for the immediate surroundings.
Modern AI is working hard in the background so the wall can show something that looks like an old bird book. That's luxuriously understated.
A robot contest judged by “is it funny?” instead of “is it useful?” feels like a very healthy use of engineering.