home.social

#openvoiceos — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #openvoiceos, aggregated by home.social.

  1. Some time ago, an artist on Fedi:

    I wanted robots to do the dishes and the chores for me, so I can do art. Not they doing the art and chores left for me...

    I am a software developer, game developer, a writer and a translator(for FOSS and Wikipedia). And I know most of the time, LLMs create slop for a considerable amount of work. I'm not against AI nor I am against LLM. I am trying to imply that these tools are best for assisting the human. Not replacing them.

    Perhaps in future, we'll have more intelligent Artificial Intelligence tools. But as I told everyone, I highly doubt a #Transformer alone will help us travel long.

    I fully support automating boring tasks. We have such these in every field. But there is a line between a task which must be done by a human and a task which can be automated.

    BTW, meanwhile in offtopic chatroom of #OpenVoiceOS, one of the folks shared these two very valuable insights:

    sites.cc.gatech.edu/computing/

    nickbostrom.com/papers/vulnera

    As I already have written a post on me blog[1], I strongly oppose both the #AIHype and the #AIAntiHype. I see many #FOSS projects put a blanket ban on any use of LLMs. I don't think this is good. Many projects let LLMs flourish all around their codebase. And I'm very afraid of the long term consequences it *might* have. I mean maybe, there won't be any consequences. But do you want to risk on that and after like 3 years realize how the project is ruined? I am not talking about vibe coding. I am talking about professional programmers overwhelmingly trust LLMs.

    Imagine there is a muddy water and you cannot see what's in there. Will you take your chances to swim in? Or you wait till the mud goes down and the water becomes clear?

    [0] far.chickenkiller.com/computin

    #AI #LLM #LargeLanguageModels #programming #robots #robot #vibecoding

  2. Some time ago, an artist on Fedi:

    I wanted robots to do the dishes and the chores for me, so I can do art. Not they doing the art and chores left for me...

    I am a software developer, game developer, a writer and a translator(for FOSS and Wikipedia). And I know most of the time, LLMs create slop for a considerable amount of work. I'm not against AI nor I am against LLM. I am trying to imply that these tools are best for assisting the human. Not replacing them.

    Perhaps in future, we'll have more intelligent Artificial Intelligence tools. But as I told everyone, I highly doubt a #Transformer alone will help us travel long.

    I fully support automating boring tasks. We have such these in every field. But there is a line between a task which must be done by a human and a task which can be automated.

    BTW, meanwhile in offtopic chatroom of #OpenVoiceOS, one of the folks shared these two very valuable insights:

    sites.cc.gatech.edu/computing/

    nickbostrom.com/papers/vulnera

    As I already have written a post on me blog[1], I strongly oppose both the #AIHype and the #AIAntiHype. I see many #FOSS projects put a blanket ban on any use of LLMs. I don't think this is good. Many projects let LLMs flourish all around their codebase. And I'm very afraid of the long term consequences it *might* have. I mean maybe, there won't be any consequences. But do you want to risk on that and after like 3 years realize how the project is ruined? I am not talking about vibe coding. I am talking about professional programmers overwhelmingly trust LLMs.

    Imagine there is a muddy water and you cannot see what's in there. Will you take your chances to swim in? Or you wait till the mud goes down and the water becomes clear?

    [0] far.chickenkiller.com/computin

    #AI #LLM #LargeLanguageModels #programming #robots #robot #vibecoding