A surprisingly robust approach to human verification would be to abandon visual classification entirely and instead exploit culturally embedded call-and-response patterns. There are obvious concerns regarding age and demographic bias, but these could potentially be addressed through a rotating corpus of challenges. More interesting is the adversarial question: while an LLM can trivially retrieve the expected response, can it distinguish between merely knowing the response and knowing when to come in?
When I say uh you say 'ah' Uh ah Uh Ah
Now when i say freeze y'all stop on a tidime
when I say freeze you just freeze one time
when i say freeze y'all stop on a dime
FREEZE
Now - all the ladies in the place, if you got real hair, real fingernails, if you got a job, you going to school and yall need nobody to help you handle yo business make some noise
NOOIISE
So - Dr. D. J. Kool has clearly been working on this thesis for some time
I found that rather entertaining, thank you. It reminds me of Rick n Morty, S01 E04, when they start rapping on stage and making more and more specific requests to get the matrix they're stuck in to glitch out.
I’m going to get pwned one of these days, considering my threshold for giving a random website access to my webcam is apparently mild curiosity and entertainment for 30s.
But more generally, if I’m the sort of person who would allow this I’d probably allow some other stupid permissions if they promised me something novel on the other side.
Maybe try hiding your face below the camera and throwing your hands up in the sky so the camera only sees your hands then play/sing Metallica.
[Chorus]
Never cared for what they do
Never cared for what they know
But I know
[Refrain]
So close, no matter how far
It couldn't be much more from the heart
Forever trusting who we are
And nothing else matters
Well they have your IP....and browser fingerprints....rough time of day if a window, direction, weather....products you bought....your age...possibly health issues that come up....essentially any information that can be used to schedule and stalk you...
That's hilarious side by side with stories about cops abusing surveillance to stalk their ex, and women generally.
So you don't fear being victimised and don't plan to ever stand up for those who do. Not very unique, I'd say that's the primordial slime we all come from... but why stay in it?
"An overdriven webcam sensor reconfigures its pixels into an AI-steered optical phased array, discharging the laptop's total electrical power as a lethal, coherent laser pulse the instant your hand moves."
Then it rendered the page load with '<html><h1>"You have been theremin ated".'
Very impressed by the responsiveness of this. But I gotta say, it's much more fun to play a physical theremin, which has slightly different controls (two antennae, one controlling volume and the other pitch).
The other is a (slightly) more serious exploration of multivariate time signals as a control signal for an instrument. In the latter case, the instrument is learned/inferred from a segment of input audio:
Although its still kind of working in roughly the same manner as the original theramin - just using a different frequency of electromagnetic radiation.
Ermm, theremin came first, so actually the air guitar is a theremin-version of a guitar.
So actually this project should be called “Cameramin”, since it has very little do to with air guitar, other than air guitar being derived from the theremin ..
Both front/rear cameras simultaneously? How would that work? You’d either have to have it balancing upright, or hold it in your mouth or something else awkward no?
Exactly, just 3D print a little inverted π-shaped stand, plonk the Cameramin on it, and get on with the performance.
Sounds like a great app idea, no matter that someone else might’ve already done something similar. I’d love to see it personally (although I have my own theremin to play with) .. an app like this would be very fun indeed.
An app which functions when you place the iPhone vertically in such a fashion that both hands are on each side of the phone, facing both cameras.
You don’t need a stand, you could also use a salt and pepper shaker, a wedge of plasticine, your girlfriend/boyfriend’s knees...
Sheesh, try to be less pedantically rigid, honestly .. it’s a cool app idea. All it takes is for the user to mount their phone vertically while they’re using it.
> We were talking about an app hence my comment on usability. Now we’re discussing something entirely different.
The misunderstanding is, yes, for this app you would need the phone to stand vertically, so yeah, some kind of mounting adapter would be needed, or at least the user has to find a way to expose both cameras on the horizontal plane while also having their hands free ..
It would be a pretty cool app, but yeah there is the requirement that the user mount their phone somehow. My thoughts went immediately to 3D printing a little holder, but yeah, even a bit of ducttape would do it. ;P
Really cool. I remember one local musician here in Lithuania used a theremin like 20 years ago and i was amazed at the interactivity and the alien sound of it. Now sitting in front of my computer doing the same that many years later brings memories back!
Something tells me this is all a ploy to attract great talented theremin players, ML their talents...and then sell you Air play back....on subscription of course
There's a button next to "gyro" labeled "hands." Click on it, and the browser will ask for the camera permission. For me (Chrome, mac), I had to click on the "cal" button next to it the first time to actually switch to the camera.
Yeah, it's really easy to spot an AI made app. I don't want to discount the project for being coded with AI, but you can tell when the UI is basically one shotted due to all the clutter and bloated features. I don't think AI understands that "less is more" when it comes to UI. It always seems overly eager to expose every single control and add UI flourishes to the point of being gaudy.
That pedantic ship has sailed. In everyday communication people use "AI" as a shorthand for saying "LLM," even if they know perfectly well that other forms of AI exist. You can tell this by context, no need to tell people that there are other kinds of AI.
This is extremely cool. I activated the camera and it recognized the position of my hands and I was able to play it as shown in the instructions. It's very fun.
The camera watches your palms. Move them apart and it gets louder. Bring them together and it goes quiet. Raise them both and the note goes up. There is a sound wave strung between your hands, and it slackens as they close.
Two palms give you four numbers, and one of them is best left alone. Their mean height is the pitch. The distance between them is the volume. The angle of the line joining them sat unused for months and is now the vibrato: see-saw your hands and the note begins to sing. The fourth is where the pair sits horizontally, and it stays free on purpose, so that nobody is pinned to one spot in front of the camera. Their apparent size is a fifth thing and not a position at all; it moves a lowpass and a reverb send, so leaning back makes the room bigger. The angle is independent of the distance, so the vibrato leaves the volume alone, and the string drawn between the hands is already a picture of it.
It began as a phone toy — tilt left and right for volume, forward and back for pitch. The camera came later and turned out to be the better instrument. The phone mode is still there.
No video leaves your machine. The hand tracking is MediaPipe's Hand Landmarker compiled to WASM, running on your own GPU. The tracker and its model are fetched once from a CDN, about 8 MB, and after that you can unplug the network and it still plays.
Most of the work was not the sound. It was the hand that disappears. MediaPipe finds a palm by its plane, so it loses one when you turn your wrist edge-on, or when the room is dark and the webcam falls to fifteen frames a second and a moving hand becomes a smear. At first I held the note for a second after a hand vanished. People said it lagged. I cut it to half a second. People said the sound kept dropping out. Both were right. Now the hold depends on where the hand was last seen: near the edge of the frame a person has genuinely taken their hand away, so 400 ms; deep in frame nobody teleports and the detector merely blinked, so a full second, and no one hears the gap.
The other stubborn thing was the bass. A phone speaker produces nothing below roughly 150 Hz, and raising the gain there does nothing at all — I tried ten decibels and heard silence. So the bottom octave is split off, saturated gently to breed harmonics, and its fundamental is then filtered away. The ear reconstructs a pitch it never heard.
That trick then caused a bug I misread for a week. A tester said the pitch moved in steps in the phone mode, worst in the bass, and guessed it was a frame rate problem. It was not. The phone reports its tilt angle quantised to something close to a whole degree, and I had packed six octaves into eighty degrees of tilt — so one tick of the sensor was most of a semitone. It was audible in the bass specifically because of the exciter: at 70 Hz the ear resolves pitch poorly, but the exciter re-voices the note through harmonics at two and three hundred hertz, which is exactly where the ear resolves cents best. So the fix was not the sound and not the frame rate. It was a narrower range, a wider sweep, and reading the sensor at frame rate instead of once per event.
Nice work. "No video leaves your machine." Either Claude specifically or AI models in general use this phrasing now and No teeth are left un-gnashed as a result. Uggh I wonder why it's so aggravating.
My bad, I used Claude to create the description. The thing is, I'm really bad at presenting a project properly. But thanks for the criticism, I'll take it into account next time.
Best way to get better is practice, and practice is best without the assistant. Think weight lifting, rather than a forklift: the hard part is the point.
Obviously I appreciate this is hard to do when it's so tempting to get the easy convenient and free AI to do it for you, but the more practiced you are, the better you can do when you do need a metaphorical forklift.
I think the main objection is that hacker news supposedly doesn't want AI generated comments, which is going to paradoxically produce the situation where the mods will be overwhelmed with policing AI generated comments and they themselves will be replaced by AI who will then be incentivized to accuse human comments as being AI (because they're not team AI duh) and then there will be nothing but bots here.
this is really cool. i especially like the gesture mapping... using the distance between your hands for volume, their average height for pitch, and the angle between them for vibrato feels unusually intuitive. the fact that those dimensions can move independently makes this feel like it could support a lot more than a theremin sound
they're LLM-generated, web-embeddable synths, and the real-time controllable parameters can be defined in the prompt — so your hand tracking could map directly onto whatever controls make sense for the instrument
reply