> an adversary could trick Apple’s algorithm into erroneously matching an existing image
This is a very real, possible attack. Apple ships its CSAM model on device so any attacker can have a copy of the model. Then the attacker creates an image that triggers CSAM but looks like a panda [1]. Now the attacker sends tons of triggering photos to the unsuspecting victim, who now gets questioned by the FBI.
I wonder if shops will still install used (stolen) units regardless? Or maybe they are just being shipped out of state or broken down for raw materials.
Are you asking what is the advantage of having ML optimized hardware on device? Yes, running inferences is expensive too, especially for video, photo, and speech processing. I would expect this phone to have user noticeable improvements in those three areas.
> The average cost of housing has gone up over time, but this trend is mainly driven by the fact that modern houses are larger and built better than houses in the past.
Oh, that explains why the house that sold for $300k in 1990 sold today for $2M (only 3X adjusted for inflation). It must have gotten 3X larger and 3X better built, even though it is the exact same house with a couple extra layers of paint and a redone kitchen.
Nice! I did something like this a long time ago: http://theartbot.com
I wrote a couple algorithms but the best one basically treated the darkness of the image like the altitude in a map, then drew contour lines. I also physically built the plotter.
According to my page, you called it "Super Cool".