I have just launched http://weowntheglass.com, it features an oled-safe screensaver steady stream of public domain images of violent uprisings against an oppressive class. The ACR should enjoy that!
replyWhat if it was a quick succession of screenshots from actual TV shows? Random pictures probably don’t have any ACR match but TV shows would.
replyI took a look at some of the ACR patents I could find and it seems they typically use a perceptual hash time series to positively ID content, with short match rejection to avoid 'channel surfing' attributions. So the effective strategy would be a random feed of conflicting demographic content streamed for at least a few minutes each -- something like 5 minute shuffle of Miss Rachel, Game of Thrones, This Old House, 30 Rock, The Decalogue, Gunsmoke, CSPAN, Sesame Street, QVC, something like that. Problem is... who wants that actually streaming on their screen all day? My site as built probably doesn't actually poison the ACR data, it's more like waging memetic warfare/raising awareness.
replyI like this enough to actually considering buying one to get it to identify the endpoints. Maybe better to see first if we can find it in the firmware update downloads.
replyHave Claude or Astra reverse engineer the firmware?
replyI started asking Astra and Terra about this, including the idea of spoofing the data that was mentioned in this thread and got slammed with a warning for Cyber Abuse on my account, the thread is here https://chatgpt.com/s/cx_6a9fc635825481918c84187765645d5c.
replyI will be very careful going forward with what I use open ai for and what I phrase how.
How do you go about generating the garbage data? Bunch of ESP32s with rngs in the same subnet?
replyAlways speak in Klingon within hearing range of the TV. Qapla'!
replyI watch all the stations I hate, and avoid watching all my favorites.
reply