You can run deepseek-v4.1-flash on a local server with 4xH200@141GB for around 180K EUR. It can serve more than one user so it can be a good investment for a company, especially if you need your data to stay local (although I would advise do go a step up and populate with 8xH200 to serve many more users, ~300K EUR). You can run whatever test you want and you won't be limited by a front-end that pre-filters your requests.
Expect an error but this will depend on the brand.
"Smudging" is a common trick. Just dab some face oil on the lens, just enough so it can't get detail but not so much that the system can tell there's a covering.
Maybe you'll be able to buy a box to plug into the CAN bus and simulate pressing the button to deactivate it. Sorta like the auto-stop eliminator for that horrid feature (which saves less than 5 gallons of gas per year in my dad's Subaru - thankfully mine is one year too old for that).
No, the regulation requires that the system can be turned off or just the warnings; however the manufacturer wants to do it.
> 3.1.2. It shall be possible for the driver to manually deactivate either the ADDW warning or the ADDW system, depending on which of the two possibilities (or both) the vehicle manufacturer chose to make possible.
The full text of the regulation is linked at the bottom of the article.
Those nudges are gentle and totally safe in every car I've ever had. And no "random" nudges outside road construction work with dubious lane markings where you need to have a grip on the wheel anyway. A regular firm grip always overrides lane keeping.
I mostly agree in my 2024 Ioniq 5, but not in my 2019 Subaru Outback. You can definitely override the lane keep if you have a firm grip and are ready for it, but it tries to throw me off the road often enough that I don't use it anymore.
The scariest was when I had to swerve into another lane to avoid some trash that was sticking into the road from the highway. It tried to force me back into it twice! Luckily I was ready but it gave me a fright for sure.
Not sure about the systems on cars in the EU, but I got a loaner 2025 Hyundai Tuscon when my EV was in the shop. It had some driver attention monitoring feature with a camera above the steering wheel staring me in the eyes. I covered it with a piece of black electrical tape. It popped a little warning on the main display (IIRC, a crossed out eye, but maybe I'm confusing with Subaru Eyesight) when the car first started up, showing that the camera wasn't working, then proceeded to be silent for the rest of the drive.
I dunno if that'll fly going forward. I know I'll test it in every new car with this feature that I test drive though!
You are entitled to your opinion, obviously, as is anyone else. Restomods have been around since forever, and this is hardly the first EV conversion (almost certainly not the first classic Mustang EV restomod either).
Personally I think it's pretty damn cool. But I have always been a Mustang fan, and I know that this era of Mustang is not especially collectable. They made quite a large number of them and plenty are still running.
To be fair though, this is beyond typical restomod territory. This isn’t an LS swap with modern wheels and suspension. This most certainly changes the spirit of the car.
There's no question the first generation of Mustangs are the most collectible.
In practice, very short texts don't carry very high value so watermarking is (usually) less important. For longer text false positives are not an issue at all since you have a large amount of data to extract your signal from.
I wonder where you got that impression. Several professional watermarking systems for movie studio type content I have worked with (and on) are highly resistant to noise removal while remaining imperceptible.
Based on my research experience and judgment, I have published several top-conference papers in both the detection and diffusion domain, but I haven’t explored the engineering/product side. I believe that if such a system hasn’t been invented yet, it wouldn’t be difficult to create one to remove that watermark using an open-source image/video model and maintain the high quality. Would you be interested in having a further discussion on this?
reply