Conversation

Notices

1. Embed this notice
   Gareth Halfacree (ghalfacree@mastodon.social)'s status on Friday, 16-Feb-2024 19:58:51 JST Gareth Halfacree

   Google, following the industry trend to AI-all-the-things, has released Magika - a machine learning model which can identify file types. It claims it can outperform traditional methods by 20 per cent.

   I pitted it against BSD File on something I figured Google hadn't included in its million-file-strong corpus: CU Amiga's Mega CD-ROM coverdisc from November 1995.

   Magika identified... one file correctly, a plain-text document. File? File got 'em all, and quicker too.

   (An unfair test, I know!)

   • Puniko ? and narcolepsy and alcoholism :flag: like this.
   • Embed this notice
     Gareth Halfacree (ghalfacree@mastodon.social)'s status on Friday, 16-Feb-2024 21:17:00 JST Gareth Halfacree

     in reply to

     Oh, and how long did each take to run?

     Magika:
     real: 0m0.289s
     user: 0m1.282s
     sys: 0m0.717s

     File:
     real: 0m0.006s
     user: 0m0.003s
     sys: 0m0.003s

     The Future™!

     Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     Polychrome :blabcat: (polychrome@poly.cybre.city)'s status on Friday, 16-Feb-2024 21:17:00 JST Polychrome :blabcat:

     in reply to
     @ghalfacree yes but one of these will excite investors and shareholders :blobcatgooglytrash:
     
     Also, I love that font!
   • Embed this notice
     Gareth Halfacree (ghalfacree@mastodon.social)'s status on Friday, 16-Feb-2024 23:43:08 JST Gareth Halfacree

     in reply to

     • Polychrome :blabcat:

     @Polychrome It's great, isn't it? Tamzen, a fork of Tamsyn - both pixel-perfect bitmap fonts, great for the terminal and readable down to a very small size!

     https://github.com/sunaku/tamzen-font

     Polychrome :blabcat: likes this.
   • Embed this notice
     feld (feld@bikeshed.party)'s status on Saturday, 17-Feb-2024 00:24:11 JST feld

     in reply to
     @ghalfacree yeah but if I have to choose between using some AI crap to identify the files or rely on file(1) which really depends on the libmagic database -- it would be easier to use the AI tool's API.
     
     libmagic does not have a stable ABI. It sucks balls if you want to use it in your software but not ship a copy of it built for every OS you want people to run your software on. You really can't just dynamically link to it and ship your code with any expectation it will work
   • Embed this notice
     Gareth Halfacree (ghalfacree@mastodon.social)'s status on Saturday, 17-Feb-2024 00:48:51 JST Gareth Halfacree

     in reply to

     • feld

     @feld That's fair: my experience of libmagic has (thankfully) been purely as a consumer.

   • Embed this notice
     feld (feld@bikeshed.party)'s status on Saturday, 17-Feb-2024 00:48:51 JST feld

     in reply to
     @ghalfacree it's crazy because the software is *really good* even if it is a giant mess of hacks (or so I'm told)
     Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     Haelwenn /элвэн/ :triskell: (lanodan@queer.hacktivis.me)'s status on Saturday, 17-Feb-2024 14:36:16 JST Haelwenn /элвэн/ :triskell:

     in reply to

     • feld
     @feld @ghalfacree That said if Unix would have a standard way to point commands to a shared memory location, then file(1) could be used instead of libmagic, avoiding a lot of issues.
     feld likes this.