3 ms·

This anti-AI talk is a coping mechanism, but completely unproductive. People don't like the possible or likely consequences of AI that delivers even 10% of its

by fmajid 1y ago

This anti-AI talk is a coping mechanism, but completely unproductive. People don't like the possible or likely consequences of AI that delivers even 10% of its boosters' outlandish promises, and instead of planning how to organize politically to mitigate them, just plug their ears and say "la la la I can't hear you". I have a friend who is a translator, and all her translation work has dried out. All she gets now is tedious jobs reviewing and correcting machine translations. As Simon Willison points out, writing code is much more gratifying than code reviews, and I can easily imagine a future where the only work left for most human coders is code reviews of LLM-generated code, which would be terrible for job satisfaction. The early 19th Century Luddites, skilled craftspeople and weavers understood the benefits of mechanized looms, but demanded the profits be shared equitably with the displaced workers. When the proto-capitalists, who largely overlapped with the feudal aristocracy that rules Britain to this day, refused, they engaged in a guerrilla campaign of breaking looms, and the elite establishment responded by sending more soldiers to quell the uprising than were fighting Napoleon, and hanging people for breaking frames: https://en.wikipedia.org/wiki/Destruction_of_Stocking_Frames,_etc._Act_1812 https://en.wikipedia.org/wiki/Destruction_of_Stocking_Frames... Ironically the kinds of jobs hardest for AI to replace are those requiring manual dexterity and skill, often derided as "unskilled work" by snobbish white-collar academics and workers.

thefz 1y ago
> This anti-AI talk is a coping mechanism No, it' being tired of the trite bullshit new tech to rob people of agency, culture and knowledge and move more value towards the shareholders. The whole nature of LLMs is built upon exploiting other people's work. Without it they will be useless. > I have a friend who is a translator, and all her translation work has dried out. Can a machine translate prose?
- munksbeer 1y ago
  Disclaimer: I feel that your post is overly angry and reactive, but I would like to have meaningful discourse about it, and not just talk past each-other with downvotes. So with that said, what if we started from a position where I accept you are correct about the exploitation stuff, and went from there? Do you see a way back from where we are now? I don't. The genie is out the bottle, and depending on your view, that means "big tech" has won again. So what do we do? Do you think LLMs are useless, or are they not useless? Will they be useless in the future? Do you think they will be an empowering tool, or one that disempowers? > Can a machine translate prose? Maybe not now, but can it in the future? I'm going to guess yes. So of what use is arguing that it is rubbish because it can't translate prose now?
  - thefz 1y ago
    > Do you see a way back from where we are now? I don't. Do not use it. Do not promote or talk about it. Would be a great start. Do you really believe LLMs are a force for good?
    - munksbeer 1y ago
      This is like asking me if steam power was a force for good, or the combustion engine, or computers. The answer is yes and no, but the more important point is that innovation and progress is inevitable. We lead vastly better lives than people 100 years ago. Why? Because of innovation. I use LLM code completion in my IDE, and it is fairly useful, but not yet ideal. I use it all the time to ask technical questions rather than searching documentation first. It is extremely helpful for that task, and most of the time is correct - I always double check after it points me in the right direction. I see a path where AI leads to the destruction of humanity, but it could equally lead to a post scarcity utopia. You are being completely dogmatic in you view, and I can understand. If you truly believe AI is a force for evil, then it makes sense for you to rage against it. But keep in mind, no-one knows yet, and be open to the possibility that you're wrong.

by fmajid 1y ago

thefz 1y ago

> This anti-AI talk is a coping mechanism No, it' being tired of the trite bullshit new tech to rob people of agency, culture and knowledge and move more value towards the shareholders. The whole nature of LLMs is built upon exploiting other people's work. Without it they will be useless. > I have a friend who is a translator, and all her translation work has dried out. Can a machine translate prose?

munksbeer 1y ago
Disclaimer: I feel that your post is overly angry and reactive, but I would like to have meaningful discourse about it, and not just talk past each-other with downvotes. So with that said, what if we started from a position where I accept you are correct about the exploitation stuff, and went from there? Do you see a way back from where we are now? I don't. The genie is out the bottle, and depending on your view, that means "big tech" has won again. So what do we do? Do you think LLMs are useless, or are they not useless? Will they be useless in the future? Do you think they will be an empowering tool, or one that disempowers? > Can a machine translate prose? Maybe not now, but can it in the future? I'm going to guess yes. So of what use is arguing that it is rubbish because it can't translate prose now?
- thefz 1y ago
  > Do you see a way back from where we are now? I don't. Do not use it. Do not promote or talk about it. Would be a great start. Do you really believe LLMs are a force for good?
  - munksbeer 1y ago
    This is like asking me if steam power was a force for good, or the combustion engine, or computers. The answer is yes and no, but the more important point is that innovation and progress is inevitable. We lead vastly better lives than people 100 years ago. Why? Because of innovation. I use LLM code completion in my IDE, and it is fairly useful, but not yet ideal. I use it all the time to ask technical questions rather than searching documentation first. It is extremely helpful for that task, and most of the time is correct - I always double check after it points me in the right direction. I see a path where AI leads to the destruction of humanity, but it could equally lead to a post scarcity utopia. You are being completely dogmatic in you view, and I can understand. If you truly believe AI is a force for evil, then it makes sense for you to rage against it. But keep in mind, no-one knows yet, and be open to the possibility that you're wrong.