Conversation

Notices

1. Embed this notice
   Erik Jonker (erikjonker@mastodon.social)'s status on Tuesday, 03-Feb-2026 00:26:59 JST Erik Jonker

   • Kevin Beaumont
   • Rich Felker

   @dalias @GossiTheDog ...wow, your are totally locked-up in your own thinking, the person i mentioned did nothing even close to what you just mentioned... he was editing his OWN work...., he is not on reddit or social media....

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Tuesday, 03-Feb-2026 00:26:59 JST Rich Felker

     in reply to

     • Kevin Beaumont

     @ErikJonker @GossiTheDog I was being a bit abrasive and sarcastic, but even if you're starting with your own images, you're now also appropriating other people's as soon as you start using "AI" tools on them.

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Tuesday, 03-Feb-2026 00:46:19 JST Rich Felker

     in reply to

     • Kevin Beaumont

     @ErikJonker @GossiTheDog The fix is not to use it.

   • Embed this notice
     Erik Jonker (erikjonker@mastodon.social)'s status on Tuesday, 03-Feb-2026 00:46:20 JST Erik Jonker

     in reply to

     • Kevin Beaumont
     • Rich Felker

     @dalias @GossiTheDog ...i see, you mean because of their training data, i agree, but that is hard to fix now...

   • Embed this notice
     Erik Jonker (erikjonker@mastodon.social)'s status on Tuesday, 03-Feb-2026 00:52:04 JST Erik Jonker

     in reply to

     • Kevin Beaumont
     • Rich Felker

     @dalias @GossiTheDog ...true, but empirically we can establish that ship has sailed, looking at usage numbers. Then we would also have to stop using models trained on those models etcetera, impossible...

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Tuesday, 03-Feb-2026 00:52:04 JST Rich Felker

     in reply to

     • Kevin Beaumont

     @ErikJonker @GossiTheDog No, we absolutely do not have to accept that claim. "Usage numbers" are all lies told for investors. The vast majority of "users" are unwilling ones who had it put into the systems they use against their will and who are being counted as a user because they didn't or couldn't properly opt out.

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Tuesday, 03-Feb-2026 02:08:34 JST Rich Felker

     in reply to

     • Kevin Beaumont

     @ErikJonker @GossiTheDog Those are not open source models. The definition of open source demands license to actually use everything the source is made up of. No model trained on stolen data, directly or indirectly, is "open source".

   • Embed this notice
     Erik Jonker (erikjonker@mastodon.social)'s status on Tuesday, 03-Feb-2026 02:08:35 JST Erik Jonker

     in reply to

     • Kevin Beaumont
     • Rich Felker

     @dalias @GossiTheDog sure, but opensource models are trained on those models trained with stolen data, you can not make them disappear, they are used in backends, for example powerful Chinese opensource models

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Tuesday, 03-Feb-2026 03:19:16 JST Rich Felker

     in reply to

     • Kevin Beaumont

     @ErikJonker @GossiTheDog Sure we will. Any application that is actually open source will have the user know what's being used in it, always, and will not have any of this nonsense. If by "applications" you mean a broader sense not of software but rather "this stuff getting applied to social systems we're a part of" that is a threat, one we deal with by organizing and legislation, opposing the technofascists who want to impose what you're obediently describing as inevitable.

   • Embed this notice
     Erik Jonker (erikjonker@mastodon.social)'s status on Tuesday, 03-Feb-2026 03:19:17 JST Erik Jonker

     in reply to

     • Kevin Beaumont
     • Rich Felker

     @dalias @GossiTheDog i know but my point is we will not prevent open weight models like Qwen and Kiwi being used in all kind of applications, without users even knowing it

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Tuesday, 03-Feb-2026 03:22:03 JST Rich Felker

     in reply to

     @ErikJonker @GossiTheDog And here's where I get to the point of validation for my initial abrasive response: your behavior is a pattern I'm very familiar with. Jumping in to a discussion where folks are establishing shared values of rejecting a technofascist future, and innocuously making a pro-AI comment. But when poked, you can't resist going into the all-out inevitability narrative that's what you actually were hoping to push.

     Sorry not sorry, folks like me are going to speak up and forcefully reject that whole narrative and pattern of behavior. We see it and we won't let it go quietly.