Florian Maury on Nostr: google/magika: "Detect file content types with deep learning" Fucking hell. Just ...
google/magika: "Detect file content types with deep learning"
Fucking hell. Just because we could not create formats with explicit internal types (like PDF, PNG or WEBVTT which start with the strings PDF, PNG or WEBVTT), we are now wasting vast quatities of energy to "predict the type of a file".
OMG, we are so fucking doomed.
(also, throwing yet another AI bullshit into the mix, when libmagic was doing it with static rules; how difficult can it be to just quarantine all files with "data" (i.e. unknown) data type?)
Fucking hell. Just because we could not create formats with explicit internal types (like PDF, PNG or WEBVTT which start with the strings PDF, PNG or WEBVTT), we are now wasting vast quatities of energy to "predict the type of a file".
OMG, we are so fucking doomed.
(also, throwing yet another AI bullshit into the mix, when libmagic was doing it with static rules; how difficult can it be to just quarantine all files with "data" (i.e. unknown) data type?)