Microsoft develops ultra durable glass plates that can store several TBs of data for 10000 years::Project Silica’s coaster-size glass plates can store unaltered data for thousands of years, creating sustainable storage for the world
Microsoft develops ultra durable glass plates that can store several TBs of data for 10000 years::Project Silica’s coaster-size glass plates can store unaltered data for thousands of years, creating sustainable storage for the world
It seems like it would make for a great replacement for Tape Backups that are currently used for long term storage. They are easy to write to but hard to read from and restore. It’ll probably be a great technology to put backups on especially if it lasts as long as they say. The challenge will probably come in with the specialized reading and writing laser / microscopes being expensive.
According to the article, they’re using their AI cloud service to decode the data, so it’s also likely so computationally expensive to decode that it won’t be practical. Seems more like a gimmick to woo investors that won’t actually ever see real world use, at least not any time soon. I suppose you could make the argument that you can back up data on it now, and hope reading it becomes more practical later, but then it’s more of a supplement to tape backup, rather than a replacement.
There is certainly an element of this being PR for Microsoft. But it is worth considering that a huge amount of computing is done in large data centers.
I think this fact could easily jump-start the use of a technology such as this. If it starts out where every large to mid-sized data center has a reader and writer shared among their thousands of customers it certainly would make it more viable.
I would guess the AI service is MS’s way of trying to make sure they control the technology. Hopefully, it eventually can get replaced by a local AI model rather than MS’s proprietary AI.
The hell does that even mean? Is it a model that convinces people it’s decrypting data while taking guesses based on the training set?
My guess is it’s an attempt to build long term a subscription service model behind the idea. No subscription, equals it can’t be read or some contrived bs to leech more money out of users/governments of the encoding/decoding technology.