> ...never asked why were child safety features not a concern for Apple Inc for the last 20+ years.
This is remarkably disingenuous and can be applied to any company who hasn't done any number of things to help prevent any number of criminal or immoral activities on their platforms. Did we have the need or the technology to implement such features in 2001? I wasn't using cloud services to store photos in 2001. Were servers, let alone devices powerful enough to generate the hashes? Did NCMEC have the capability to generate these hashes or even have a database from which they could help create these technologies? PhotoDNA's development didn't start until 2009.
Capability has expanded because the devices we carry around are super computers with built-in ML chips to enable these sorts of technologies. This would not have been possible with a feature phone in 2001. Advancements in machine learning have also enabled these sorts of technologies.
Concerns have expanded because our use of third party platforms for cloud storage has significantly expanded which creates a liability for all these platforms. E2EE would of course remove their liability since they would have no idea what was being hosted on their platforms but the problem of ensuring that people don't lose access to their photos in the case of losing their device or password remains.
If Apple's users had intrinsic trust in the company then Apple wouldn't be concerned about the push back against this tech. All you're doing here is outing yourself as the opposite of the "apple sheep."
> All you're doing here is outing yourself as the opposite of the "apple sheep."
Fair point. I feel its a deflection from the point though. Nothing about this whole saga has much to do with tech. Understandbly since this is an engineering medium - fair play.
> This would not have been possible with a feature phone in 2001.
Perceptual hashes (photoDNA) can be executed on pretty much any device, the requirements for it are quite low as far as I understand it. Probably not a great idea to run it on a 40mHz Nokia, but it would actually be rather viable with devices swinging past the 1gHz threshold.
This is remarkably disingenuous and can be applied to any company who hasn't done any number of things to help prevent any number of criminal or immoral activities on their platforms. Did we have the need or the technology to implement such features in 2001? I wasn't using cloud services to store photos in 2001. Were servers, let alone devices powerful enough to generate the hashes? Did NCMEC have the capability to generate these hashes or even have a database from which they could help create these technologies? PhotoDNA's development didn't start until 2009.
Capability has expanded because the devices we carry around are super computers with built-in ML chips to enable these sorts of technologies. This would not have been possible with a feature phone in 2001. Advancements in machine learning have also enabled these sorts of technologies.
Concerns have expanded because our use of third party platforms for cloud storage has significantly expanded which creates a liability for all these platforms. E2EE would of course remove their liability since they would have no idea what was being hosted on their platforms but the problem of ensuring that people don't lose access to their photos in the case of losing their device or password remains.
If Apple's users had intrinsic trust in the company then Apple wouldn't be concerned about the push back against this tech. All you're doing here is outing yourself as the opposite of the "apple sheep."