Conversation

Notices

1. Embed this notice
   Anthropy :verified_dragon: (anthropy@mastodon.derg.nz)'s status on Saturday, 11-Jan-2025 08:19:16 JST Anthropy :verified_dragon:

   Severely tired of people whining about offline opensource AI models as if they are somehow as offensive as large online corporate capitalist models.

   Once again people are screaming murder about things they just don't understand, because they refuse to have the slightest bit of nuance in their opinions.

   This is basically like going back in time and being upset about weather prediction models because "AI"

   forever relevant: https://mastodon.derg.nz/@anthropy/111471337222055082

   edit, context: https://mas.to/@alternativeto/113805097886731681

   • Embed this notice
     Anthropy :verified_dragon: (anthropy@mastodon.derg.nz)'s status on Saturday, 11-Jan-2025 08:19:06 JST Anthropy :verified_dragon:

     in reply to

     VLC's ML-derived speech to text is undoubtedly imperfect. whether you're going to call it AI, ML, speech to text, digital hamsters or literally anything else, is, as literally as it gets, arguing semantics.

     And I'd really really love to move past the semantics so we can have an actual discussion on the meaningful parts, like, how good is the model actually (rather than assuming it's 'biased'), is it going to bloat the player significantly or is it going to be an optional extra download? etc

     narcolepsy and alcoholism :flag: likes this.
   • Embed this notice
     Anthropy :verified_dragon: (anthropy@mastodon.derg.nz)'s status on Saturday, 11-Jan-2025 08:19:08 JST Anthropy :verified_dragon:

     in reply to

     Would I prefer people not calling it AI? Sure, but I kind of doubt if we're going to call the whole pile "ML" it's going to be any different.

     Also: IMHO we should stop with this silly "but biased" thing as soon as the word "AI" is dropped; that's no different for any other ML implementation, if you've ever worked with early days OCR you know exactly what poorly trained/biased "AI"/ML looks like.
     Is that meh? Sure, any type of (bayesian, or other statistical) inference is, just ask spamfilters 😅

   • Embed this notice
     Anthropy :verified_dragon: (anthropy@mastodon.derg.nz)'s status on Saturday, 11-Jan-2025 08:19:10 JST Anthropy :verified_dragon:

     in reply to

     another lukewarm AI take in this thread: if you "hate AI" you're no better than the tone deaf marketeers that are labelling everything as "AI", because neither of you actually wants to acknowledge the two letter acronym has long lost its value and meaning, and you are both thus saying nothing.

     And you can't just say "but then don't call it AI" either, because we are well and far past that station, too.

     Just simply have a nuanced take instead of going all out zealot or foaming savage.

   • Embed this notice
     Anthropy :verified_dragon: (anthropy@mastodon.derg.nz)'s status on Saturday, 11-Jan-2025 08:19:11 JST Anthropy :verified_dragon:

     in reply to

     I much understand people being upset about 'the AI hype', I am with you there, I just really think that we need better words for this, if we actually want to make a tangible impact on the companies doing these immoral things.

     I don't know many 'AI haters' that have explicitly talked about how OpenAI has literally taken up defense contracts, or called for specifically e.g making regulation to force datacenters to only generate their own (green) power, it's almost always 'datacenters & AI bad' :/

   • Embed this notice
     Anthropy :verified_dragon: (anthropy@mastodon.derg.nz)'s status on Saturday, 11-Jan-2025 08:19:12 JST Anthropy :verified_dragon:

     in reply to

     AI is just an easy scapegoat and companies LOVE IT when you do this because they would MUCH RATHER you attack the incoherent vague term "AI" and get written off as a lunatic, than actually make a well-reasoned argument why e.g OpenAI are a bunch of lunatics that are doing grossly immoral things at scale.

     And more importantly: THESE PEOPLE DON'T NEED AI TO EXTORT, HURT AND EXPLOIT YOU, THAT'S JUST A FACADE, AI/ML is just a tool, you wouldn't blame a knife for the acts of a murderer that's absurd

   • Embed this notice
     Anthropy :verified_dragon: (anthropy@mastodon.derg.nz)'s status on Saturday, 11-Jan-2025 08:19:13 JST Anthropy :verified_dragon:

     in reply to

     I fully understand if you hate it when:
     - companies scrape the entire internet for training data without consent
     - use these tools to put actual humans that have worked hard to become good at their arts and crafts out of a job so that some CEO can get a multi-million bonus
     - in general make money over other people's backs
     - plagiarize copyrighted works and somehow get fucking away with it too
     - etc

     But:
     - Blame those who actually do these immoral things
     - and preferably don't limit it to "AI" 🤷

   • Embed this notice
     Anthropy :verified_dragon: (anthropy@mastodon.derg.nz)'s status on Saturday, 11-Jan-2025 08:19:15 JST Anthropy :verified_dragon:

     in reply to

     Flaming hot take: Being against "AI" is useless because it's become a generic marketing term to denominate all of ML, and I SINCERELY HOPE people can agree that being against weather prediction models, spam filters, OCR, and search engines in general, just because you don't like ChatGPT and the likes thereof (which I do fully understand!), is insane?

     Can people *PLEASE* come up with a more nuanced take than "AI bad >:c"? Please?

   • Embed this notice
     narcolepsy and alcoholism :flag: (hj@shigusegubu.club)'s status on Saturday, 11-Jan-2025 08:19:30 JST narcolepsy and alcoholism :flag:

     in reply to
     @anthropy god I love the "digital hamsters" term so much now