Data is written in a square glass platter with ultrafast femtosecond lasers through voxels. These are permanent modifications to the physical structure of the glass, and allow for multiple bits of data to be written in layers across the surface of the glass. These layers are then stacked vertically in their hundreds.
To read data, they employ polarization microscopy technology to image the platter, while the read drive scans sectors in a Z-pattern. The images are then sent to be processed and decoded, which leans on machine learning model to convert analog signals to digital data.
You must log in or register to comment.
Digital archival storage, like digital storage that really lasts on a long scale, would be huge if it is brought on line at scale. Bitrot and the gradual loss of digital data is a massive cultural problem. We’re going to lose so much to the gradual decay of currently existing digital media.
Is that a significant problem with everything being stored on AWS/Azure? They’re pretty good about backing everything up
My understanding is that, over a long enough time period, even with good error correction, bitrot is going to creep in. I think you’re right, right now they’re really, really good at it. But I don’t think there’s an equivalent of like acid-free museum quality paper or whatever that they use to write important documents on. Reading and writing to disc drives and SSDs degrades the hardware and errors creep in. And so as far as I know, if you’re talking about storing digital data for decades or centuries, we don’t really have anything right now.
This sounds like they’re using lasers to write the data in to the glass permanently, which sounds like it’ll be there for a long, long, long time and not subject to some fo the issues hard discs and SSDs have. That could be a big help to the problem of really long term storage.
Without a single degree, they built us papyrus that lasted thousands of years
And then the engineers arrived and gave us bitrot
deleted by creator
which leans on machine learning model to convert analog signals to digital data.
No fucking way does this work
Machine learning is used in so many evil ways, employed stupidly, and depended on in too complex situations because capitalists are greedy but this is extremely trivial for what we can accomplish today
Even disregarding that, it either works consistently or it doesn’t, and that’s easily verifiable with checksums and enough testing.
Exact data and machine learning sounds like a horrible combination. I hope they’re using checksums for everything, but how do you store the checksums long term?
The AI hallucinates anime tiddies on every picture of a historical figure. I hate how bazinga brains these days are too stupid to implement basic signal processing and have to use machine learning to brute force every problem they come across.
I’d like to believe it. But if companies and governments refuse to change or implement the simplest thing after being being pentested 5000 times, then the technology really doesn’t matter. I suppose it might help that Microsoft has a monopoly on this and most organizations use MS products, so the next pentest report might just suggest “call your Microsoft rep and buy the new glass storage.”
I been wanting to make a homebrew version of this tech for a few years now. you can get femtosec pulse lasers on alibaba for like 500 USD. other than not having the money, I haven’t thought of a way to read it back after writing. seems fun tho
