Oh, look, it’s the latest mass storage tech that will never be commercially produced!
Microsoft has etched palm-sized slabs of ordinary glass into data “books” capable of storing 4.8 terabytes — the equivalent of roughly 2M books or 200 4K movies
Submitted 1 day ago by Innerworld@lemmy.world to technology@lemmy.world
https://www.nature.com/articles/s41586-025-10042-w
Comments
jordanlund@lemmy.world 1 day ago
sin_free_for_00_days@sopuli.xyz 1 day ago
Reminds me a bit of this older “breakthrough” petapixel.com/…/new-5d-disc-storage-can-store-500…
Wildmimic@anarchist.nexus 6 hours ago
The Microsoft Team cites Yuhao Lei, the researcher behind your article, 3 times - regarding the techniques of writing in different types of silica incl. polarization. So yes, it’s the continuation of the breakthrough.
jordanlund@lemmy.world 23 hours ago
“Holographic Storage” - en.wikipedia.org/…/Holographic_Data_Storage_Syste…
cmnybo@discuss.tchncs.de 22 hours ago
It’s for archival of data that needs to be stored for thousands of years, not for consumer use.
XLE@piefed.social 23 hours ago
MetalSlugX@piefed.social 22 hours ago
Weird bow Microslop hasn’t lie-boasted about quantum computing in a year.
FauxLiving@lemmy.world 20 hours ago
Finally, we can have sci-fi future that is just weird, rather than dystopian.
Image
MatSeFi@lemmy.liebeleu.de 19 hours ago
Not sure what to think about that a CNN is involved in the reading process.
FauxLiving@lemmy.world 19 hours ago
It’s a way to infer the data without having to create some human engineered and fragile detection method.
The problem of dealing with unreliable signal transmission (i.e. a CNN’s error rate at inferring the data based on their imaging) is well explored. A CNN that fails to correctly read some measurable percentage of time is not much different than a wireless data transmission on a noisy channel.
You solve the problem by encoding the signal so that you can check the data as it comes in to discover and correct for errors. A simple example would be writing the data 3 times so that you could compare the inference on each of the 3 places where the data is written. Modern error checking algorithms can do a lot better than this, space-wise.
CNNs can be trained to have a very high accuracy rate on these kinds of image recognition tasks (especially with a limited symbol set) and they can tune their error correction around the CNN’s error rates so the net result would be a clean and error check and corrected output.
Not to mention that CNNs may not be required of future persons with better imaging technology.
pastermil@sh.itjust.works 19 hours ago
CNN as in “Central News Network” or “Convoluted Neural Network”?