Preventing AI Plagiarism With .ASS Subtitling

Around two years ago, the world was inundated with news about how generative AI or large language models would revolutionize the world. At the time it was easy to get caught up in the hype, but in the intervening months these tools have done little in the way of productive work outside of a few edge cases, and mostly serve to burn tons of cash while turning the Internet into even more of a desolate wasteland than it was before. They do this largely by regurgitating human creations like text, audio, and video into inferior simulacrums and, if you still want to exist on the Internet, there’s basically nothing you can do to prevent this sort of plagiarism. Except feed the AI models garbage data like this YouTuber has started doing.

At least as far as YouTube is concerned, the worst offenders of AI plagiarism work by downloading the video’s subtitles, passing them through some sort of AI model, and then generating another YouTube video based off of the original creator’s work. Most subtitle files are the fairly straightfoward .srt filetype which only allows for timing and text information. But a more obscure subtitle filetype known as Advanced SubStation Alpha, or .ass, allows for all kinds of subtitle customization like orientation, formatting, font types, colors, shadowing, and many others. YouTuber [f4mi] realized that using this subtitle system, extra garbage text could be placed in the subtitle filetype but set out of view of the video itself, either by placing the text outside the viewable area or increasing its transparency. So now when an AI crawler downloads the subtitle file it can’t distinguish real subtitles from the garbage placed into it.

[f4mi] created a few scripts to do this automatically so that it doesn’t have to be done by hand for each one. It also doesn’t impact the actual subtitles on the screen for people who need them for accessibility reasons. It’s a great way to “poison” AI models and make it at least harder for them to rip off the creations of original artists, and [f4mi]’s tests show that it does work. We’ve actually seen a similar method for poisoning data sets used for emails long ago, back when we were all collectively much more concerned about groups like the NSA using automated snooping tools in our emails than we were that machines were going to steal our creative endeavors.

Thanks to [www2] for the tip!

Continue reading “Preventing AI Plagiarism With .ASS Subtitling”

Making The Longest-Distance Radio Contact Possible

One of the more popular activities in the ham radio world is DXing, which is attempting to communicate with radio stations as far away as possible. There are some feats that will earn some major credibility in this arena, like two-way communication with Antarctica with only a few watts of power, long-path communication around the globe, or even bouncing a signal off the moon and back to a faraway point on Earth. But these modes all have one thing in common: they’re communicating with someone who’s also presumably on the same planet. Barring extraterrestrial contact, if you want to step up your DX game you’ll want to try to contact some of our deep-space probes (PDF warning).

[David Prutchi] aka [N2QG] has been doing this for a number of years now and has a wealth of knowledge and experience to share. He’s using both a 3.2 meter dish and a 1.2 meter dish for probing deep space, as well as some custom feed horns and other antennas to mount to them. Generally these signals are incredibly small since they travel a long way through deep space, so some amplification of the received signals is also needed. Not only that, but since planets and satellites are all moving with respect to each other, some sort of tracking system is needed to actively point the dish in the correct direction.

With all of that taken care of, it’s time to see what sort of signals are coming in. Compared to NASA’s 70-meter antennas used to communicate with deep space, some signals received on smaller dishes like these will only see the carrier wave. This was the case when an amateur radio group used an old radio telescope to detect one of the Voyager signals recently. But there are a few cases where [David] was able to actually receive data and demodulate it, so it’s not always carrier-only. If you’re sitting on an old satellite TV dish like these, we’d certainly recommend pointing it to the sky to see what’s out there. If not, you can always 3D print one.

Bouncing Signals Off Of Satellites Other Than The Moon

The moon is a popular target for ham radio operators to bounce signals since it’s fairly large and follows a predictable path. There are some downsides, though; it’s not always visible from the same point on Earth and is a relatively long way away. Thinking they could trade some distance for size, an amateur radio group from the Netherlands was recently able to use a radio telescope pointed at a geostationary satellite to reflect a signal back down to Earth, using this man-made satellite to complete the path instead of the more common natural one.

While there are plenty of satellites in orbit meant for amateur radio communication (including the International Space Station, although it occasionally does other things too), these all have built-in radio transmitters or repeaters specifically meant for re-transmitting received signals. They’re also generally not in geostationary orbit. So, with a retired radio telescope with a 20-meter dish aimed directly at one of the ones already there, they sent out a signal which bounced off of the physical body of the satellite and then back down where it was received by a station in Switzerland. Of course, the path loss here is fairly extreme as well since the satellite is small compared to the moon and geostationary orbit is a significant distance away, so they used the Q65 mode in WSJT-X which is specifically designed for recovering weak signals.

Don’t break out the tape measure Yagi antenna to try this yourself just yet, though. This path is not quite as reliable as Earth-Moon-Earth for a few reasons the group is not quite sure about yet. Not every satellite they aimed their dish at worked, although they theorize that this might be because of different shapes and sizes of the satellites or that the solar panels were not pointing the correct direction. But they were able to make a few contacts using this method nonetheless, a remarkable achievement they can add to their list which includes receiving a signal from one of the Voyager spacecraft.

Bosch Starter Motor Freed From Mercedes Prison

Vehicle alternators are interesting beasts. Produced on a massive scale, these electric machines are available for a minimum of cost and contain all kinds of great parts: some power electronics and a belt-driven generator are generally standard fare. This generator can also be used as a motor with only minor changes to the machine as a whole, so thanks to economies of scale it’s possible to get readily-available, powerful, compact, and cheap motors for all kinds of projects using alternators as a starting point. [LeoDJ] noticed that this starter motor in a modern Mercedes had some interesting benefits beyond all of these perks, but it took a bit more work to get up and running than a typical alternator would have.

The motor, built by Bosch, can be found in the Mercedes E200 EQ Boost. This is a hybrid car, but different than something like a Prius in that it doesn’t have an electric motor capable of powering the car on its own. Instead it uses a combination starter motor/alternator/generator to provide extra power to the engine during acceleration, increasing efficiency and performance. It can also charge the small battery bank when the car slows down. Vehicles that use this system need much beefier alternators than a standard car, but liberating it from the car means that it has much more power available than a typical alternator would.

There were a number of issues to solve, though. Being that the motor/alternator has to do all of this extra work (and that it came out of a car whose brand is known for being tedious to work on in the first place) it is much more complicated than an off-the-shelf alternator. [LeoDJ] has been able to get his to spin by communicating with it over the CAN bus, but there’s still some work to be done before it goes into something like an impressively fast electric bicycle.

Thanks to [RoganDawes] for the tip!

Small Print Bed? No Problem!

One of the major limitations of 3D printers is the size of the printable area. The robotic arm holding the printer head can only print where it can reach, after all. Some methods of reducing this constraint have been tried before, largely focusing on either larger printers or printer heads that are mobile in some way. Another approach to increasing the size of prints beyond the confined space typical of most consumer-grade 3D printers is to create some sort of joinery into the prints themselves so that larger things can be created. [Cal Bryant] is developing this jigsaw-based method which has allowed him to produce some truly massive prints.

Rather than making the joints by hand, [Cal]’s software will cut up a model into a certain number of parts when given the volume constraints of a specific 3D printer so it can not only easily print the parts, but also automatically add the jigsaw-like dovetail joints to each of the sections of the print. There were a few things that needed prototyping to get exactly right like the tolerance between each of the “teeth” of the joint, which [Cal] settled on 0.2 mm which allows for a strong glued joint, and there are were some software artifacts to take care of as well like overhanging sections of teeth around the edges of prints. But with those edge cases taken care of he has some working automation software that can print arbitrarily large objects.

[Cal] has used this to build a few speaker enclosures, replacing older MDF designs with 3D printed ones. He’s also built a full-size arcade cabinet which he points out was an excellent way to use up leftover filament. Another clever way we’ve seen of producing prints larger than the 3D printer is to remove the print bed entirely. This robotic 3D printer can move itself to a location and then print directly on its environment.

Electric Vehicle Charging Heats Up

As the electric vehicle takeover slowly lumbers along, marginally increasing efficiencies for certain applications while entrenching car-centric urban design even further, there are some knock-on effects that are benefiting people and infrastructure beyond simple transportation. Vehicle-to-grid technology has applications for providing energy from the car back to the grid for things like power outages or grid leveling. But [Technology Connections] is taking this logic one step further. Since a large number of EV owners have charging stations built into their garages, he wondered if these charging stations could be used for other tasks and built an electric heater which can use one for power.

This project uses a level 2 charger, capable of delivering many kilowatts of power to an EV over fairly standard 240V home wiring with a smart controller in between that and the car. Compared to a level 1 charger which can only trickle charge a car on a standard 120V outlet (in the US) or a DC fast charger which can provide a truly tremendous amount of energy in a very short time, these are a happy middle ground. So, while it’s true a homeowner could simply wire up another 240V outlet for this type of space heater or other similar application, this project uses the existing infrastructure of the home to avoid redundancies like that.

Of course this isn’t exactly plug-and-play. Car chargers communicate with vehicles to negotiate power capabilities with each other, so any appliance wanting to use one as a bulk electric supply needs to be able to perform this negotiation. To get the full power available in this case all that’s needed is a resistor connected to one of the signal wires, but this won’t work for all cases and could overload smaller charging stations. For that a more complex signalling method is needed, but since this was more of a proof-of-concept we’ll still call it a success. For those wanting to DIY the charger itself, building one from the ground up is fairly straightforward as well.

Thanks to [Billy] for the tip!

Continue reading “Electric Vehicle Charging Heats Up”

DIY Drones Deliver The Goods With Printed Release

It seems like the widespread use of delivery drones by companies like Amazon and Wal-Mart has been perpetually just out of reach. Of course robotics is a tricky field, and producing a fleet of these machines reliable enough to be cost effective has proven to be quite a challenge. But on an individual level, turning any drone into one that can deliver a package is not only doable but is something [Iloke-Alusala] demonstrates with their latest project.

The project aims to be able to turn any drone into a delivery drone, in this case using a FPV drone as the platform. Two hitch-like parts are 3D printed, one which adds an attachment point to the drone and another which attaches to the package, allowing the drone to easily pick up the package and then drop it off quickly. The real key to this build is the control mechanism. [Iloke-Alusala] used an ESP32 to tap into the communications between the receiver and the flight controller. When the ESP32 detects a specific signal has been sent to the flight controller, it can activate the mechanism on the 3D printed hitch to either grab on to a package or release it at a certain point.

While this is a long way from a fully autonomous fleet of delivery drones, it goes a long way into showing that individuals can use existing drones to transport useful amounts of material and also sets up a way for an ESP32 to decode and use a common protocol used in drones, making it easy to expand their capabilities in other ways as well. After all, if we have search and rescue drones we could also have drones that deliver help to those stranded.

Continue reading “DIY Drones Deliver The Goods With Printed Release”