New rule at BSD Cafe: Do not promote AI
-
@troed the most important models are produced and sold by the US companies. And there are some powerful, self-hostable Chinese models. Still, made and controlled by some specific companies.
The average Joe can adapt them, but cannot create a "new model". And the hardware prices are skyrocketing, making self hosting those models much more difficult than it should be.
@stefano I'd say Qwen 4 Next Flash is probably the most important model since it's so much cheaper to host (which you can do yourself) that the tiny delta between it and the main US models means it's much more cost effective to use it even though it needs slightly more work to get to the same result. We also have European models that are fully ethically trained with the whole dataset being open source that some companies opt to use.
Personally I've never used a single US cloud model after I realised I was wrong about LLMs back in March. Only local ones - and I'm in full control over what they do.
I also don't think there's data supporting that we're seeing large scale worker replacement - that sounds more like some of the anti AI propaganda that is extremely prevalent here on Mastodon.
So far I've been the target of harassment for daring to post that LLMs are useful tools within my areas of expertise (software dev, cybersec and IT adm) so I'm slightly wary that there are those that consider ANYTHING that's not pure anti AI hatred to be "promoting AI".
-
New rule at BSD Cafe: Do not promote AI
BSD Cafe prefers to define itself by what it supports, rather than by what it opposes.
We support people, open technologies, decentralisation, technological independence, privacy, user control, the sharing of knowledge, and communities where technology remains a tool rather than becoming the authority.
For these reasons, we cannot support the current direction of AI and the ecosystem surrounding it.
Much of today's AI technology is developed and controlled by a very small number of extremely large companies, through an unprecedented concentration of computing power, data, infrastructure, and capital. This creates dependencies that are fundamentally at odds with many of the values this community was built around.
We are also seeing increasingly direct consequences for people. Skilled professionals are losing work to automated systems. Organisations are becoming dependent on services they cannot inspect, understand, or control. AI-generated technical advice is often presented with confidence while lacking essential context. Automation is being used to scale abuse, manipulation, spam, and attacks.
At the same time, fear of these threats can itself push people and organisations towards a handful of large centralised providers, further reinforcing the same concentration of power.
The problem is not simply the existence of artificial intelligence as a technology. The problem is how these technologies are currently being developed, deployed, controlled, marketed, and increasingly presented as an inevitable replacement for human skills, judgement, creativity, and interaction.
Too often, the message coming from the companies selling these systems is essentially: *this is inevitable, it will replace what you do, and you will have to adapt to it whether you want to or not.*
At the BSD Cafe, we believe there must always be another choice.
For this reason, we ask members **not to use BSD Cafe to promote, advertise, or evangelise AI services, or the idea that replacing human work and decision-making with AI is inherently desirable, inevitable, or something everyone should simply accept.**
This does not prevent discussion.
But please do not use this community to promote the current AI-driven model of technological centralisation and human replacement.
**"Own your data" and the current state of AI are simply not compatible.**
Bravo!
-
New rule at BSD Cafe: Do not promote AI
BSD Cafe prefers to define itself by what it supports, rather than by what it opposes.
We support people, open technologies, decentralisation, technological independence, privacy, user control, the sharing of knowledge, and communities where technology remains a tool rather than becoming the authority.
For these reasons, we cannot support the current direction of AI and the ecosystem surrounding it.
Much of today's AI technology is developed and controlled by a very small number of extremely large companies, through an unprecedented concentration of computing power, data, infrastructure, and capital. This creates dependencies that are fundamentally at odds with many of the values this community was built around.
We are also seeing increasingly direct consequences for people. Skilled professionals are losing work to automated systems. Organisations are becoming dependent on services they cannot inspect, understand, or control. AI-generated technical advice is often presented with confidence while lacking essential context. Automation is being used to scale abuse, manipulation, spam, and attacks.
At the same time, fear of these threats can itself push people and organisations towards a handful of large centralised providers, further reinforcing the same concentration of power.
The problem is not simply the existence of artificial intelligence as a technology. The problem is how these technologies are currently being developed, deployed, controlled, marketed, and increasingly presented as an inevitable replacement for human skills, judgement, creativity, and interaction.
Too often, the message coming from the companies selling these systems is essentially: *this is inevitable, it will replace what you do, and you will have to adapt to it whether you want to or not.*
At the BSD Cafe, we believe there must always be another choice.
For this reason, we ask members **not to use BSD Cafe to promote, advertise, or evangelise AI services, or the idea that replacing human work and decision-making with AI is inherently desirable, inevitable, or something everyone should simply accept.**
This does not prevent discussion.
But please do not use this community to promote the current AI-driven model of technological centralisation and human replacement.
**"Own your data" and the current state of AI are simply not compatible.**
@stefano This makes me happy, thank you very much!
-
@stefano I'd say Qwen 4 Next Flash is probably the most important model since it's so much cheaper to host (which you can do yourself) that the tiny delta between it and the main US models means it's much more cost effective to use it even though it needs slightly more work to get to the same result. We also have European models that are fully ethically trained with the whole dataset being open source that some companies opt to use.
Personally I've never used a single US cloud model after I realised I was wrong about LLMs back in March. Only local ones - and I'm in full control over what they do.
I also don't think there's data supporting that we're seeing large scale worker replacement - that sounds more like some of the anti AI propaganda that is extremely prevalent here on Mastodon.
So far I've been the target of harassment for daring to post that LLMs are useful tools within my areas of expertise (software dev, cybersec and IT adm) so I'm slightly wary that there are those that consider ANYTHING that's not pure anti AI hatred to be "promoting AI".
@troed I'm not against any technology, and AI is one of those technologies that, if correctly developed, ethically trained and properly used, can be useful for many use cases. So I perfectly understand your point.
I'm not really considering data here, but mostly what I'm seeing happening to me, to people around me and to friends. I've always thought that data about these things can be biased in one direction or another, so I tend to take it with a grain of salt.
And you should definitely never be harassed because you're not anti-AI. If people disagree with you, they can explain why they think you're wrong, or simply block or mute you and walk away. Shouting at strangers won't help in any way.
As I wrote in the original post, discussion is absolutely allowed.
What I'm asking people not to do on BSD Cafe is turn discussions into "You need to use AI or you'll be a Luddite" arguments, or promote the idea that AI should replace people, skills, judgement or creativity by default. Statements such as "Claude should be the author or co-author of any Open Source tool", for example, are very different from simply saying that you use an LLM and find it useful.
Open Source has always been about sharing, learning, doing things for the pleasure of doing them, and understanding what we build. We are not in a rush. I don't think we should give up the pleasure of hosting, implementing, learning and experimenting just because an AI can do something "faster."
And I totally agree with you on the last point: the Fediverse can sometimes be quite hostile towards anyone who doesn't take the most absolute position on a subject. I don't think attacking people for using a tool, or for having a different opinion about it, helps anyone.
Sometimes the quieter people are doing much more useful work than those spending their time arguing with a random person who's wrong on the Internet
And I'm not talking about AI here, but more generally. Shouting against something can be useful, but in my experience, going out into the world and showing people who use it why you believe it's wrong is often much more effective. -
New rule at BSD Cafe: Do not promote AI
BSD Cafe prefers to define itself by what it supports, rather than by what it opposes.
We support people, open technologies, decentralisation, technological independence, privacy, user control, the sharing of knowledge, and communities where technology remains a tool rather than becoming the authority.
For these reasons, we cannot support the current direction of AI and the ecosystem surrounding it.
Much of today's AI technology is developed and controlled by a very small number of extremely large companies, through an unprecedented concentration of computing power, data, infrastructure, and capital. This creates dependencies that are fundamentally at odds with many of the values this community was built around.
We are also seeing increasingly direct consequences for people. Skilled professionals are losing work to automated systems. Organisations are becoming dependent on services they cannot inspect, understand, or control. AI-generated technical advice is often presented with confidence while lacking essential context. Automation is being used to scale abuse, manipulation, spam, and attacks.
At the same time, fear of these threats can itself push people and organisations towards a handful of large centralised providers, further reinforcing the same concentration of power.
The problem is not simply the existence of artificial intelligence as a technology. The problem is how these technologies are currently being developed, deployed, controlled, marketed, and increasingly presented as an inevitable replacement for human skills, judgement, creativity, and interaction.
Too often, the message coming from the companies selling these systems is essentially: *this is inevitable, it will replace what you do, and you will have to adapt to it whether you want to or not.*
At the BSD Cafe, we believe there must always be another choice.
For this reason, we ask members **not to use BSD Cafe to promote, advertise, or evangelise AI services, or the idea that replacing human work and decision-making with AI is inherently desirable, inevitable, or something everyone should simply accept.**
This does not prevent discussion.
But please do not use this community to promote the current AI-driven model of technological centralisation and human replacement.
**"Own your data" and the current state of AI are simply not compatible.**
@stefano I'm really glad that AI/ML is perceived this way, especially an observation that the models themselves aren't harmful on their own. Companies which promote them are.
-
@troed I'm not against any technology, and AI is one of those technologies that, if correctly developed, ethically trained and properly used, can be useful for many use cases. So I perfectly understand your point.
I'm not really considering data here, but mostly what I'm seeing happening to me, to people around me and to friends. I've always thought that data about these things can be biased in one direction or another, so I tend to take it with a grain of salt.
And you should definitely never be harassed because you're not anti-AI. If people disagree with you, they can explain why they think you're wrong, or simply block or mute you and walk away. Shouting at strangers won't help in any way.
As I wrote in the original post, discussion is absolutely allowed.
What I'm asking people not to do on BSD Cafe is turn discussions into "You need to use AI or you'll be a Luddite" arguments, or promote the idea that AI should replace people, skills, judgement or creativity by default. Statements such as "Claude should be the author or co-author of any Open Source tool", for example, are very different from simply saying that you use an LLM and find it useful.
Open Source has always been about sharing, learning, doing things for the pleasure of doing them, and understanding what we build. We are not in a rush. I don't think we should give up the pleasure of hosting, implementing, learning and experimenting just because an AI can do something "faster."
And I totally agree with you on the last point: the Fediverse can sometimes be quite hostile towards anyone who doesn't take the most absolute position on a subject. I don't think attacking people for using a tool, or for having a different opinion about it, helps anyone.
Sometimes the quieter people are doing much more useful work than those spending their time arguing with a random person who's wrong on the Internet
And I'm not talking about AI here, but more generally. Shouting against something can be useful, but in my experience, going out into the world and showing people who use it why you believe it's wrong is often much more effective.@stefano
- agree on all points. -
@troed I'm not against any technology, and AI is one of those technologies that, if correctly developed, ethically trained and properly used, can be useful for many use cases. So I perfectly understand your point.
I'm not really considering data here, but mostly what I'm seeing happening to me, to people around me and to friends. I've always thought that data about these things can be biased in one direction or another, so I tend to take it with a grain of salt.
And you should definitely never be harassed because you're not anti-AI. If people disagree with you, they can explain why they think you're wrong, or simply block or mute you and walk away. Shouting at strangers won't help in any way.
As I wrote in the original post, discussion is absolutely allowed.
What I'm asking people not to do on BSD Cafe is turn discussions into "You need to use AI or you'll be a Luddite" arguments, or promote the idea that AI should replace people, skills, judgement or creativity by default. Statements such as "Claude should be the author or co-author of any Open Source tool", for example, are very different from simply saying that you use an LLM and find it useful.
Open Source has always been about sharing, learning, doing things for the pleasure of doing them, and understanding what we build. We are not in a rush. I don't think we should give up the pleasure of hosting, implementing, learning and experimenting just because an AI can do something "faster."
And I totally agree with you on the last point: the Fediverse can sometimes be quite hostile towards anyone who doesn't take the most absolute position on a subject. I don't think attacking people for using a tool, or for having a different opinion about it, helps anyone.
Sometimes the quieter people are doing much more useful work than those spending their time arguing with a random person who's wrong on the Internet
And I'm not talking about AI here, but more generally. Shouting against something can be useful, but in my experience, going out into the world and showing people who use it why you believe it's wrong is often much more effective. -
-
New rule at BSD Cafe: Do not promote AI
BSD Cafe prefers to define itself by what it supports, rather than by what it opposes.
We support people, open technologies, decentralisation, technological independence, privacy, user control, the sharing of knowledge, and communities where technology remains a tool rather than becoming the authority.
For these reasons, we cannot support the current direction of AI and the ecosystem surrounding it.
Much of today's AI technology is developed and controlled by a very small number of extremely large companies, through an unprecedented concentration of computing power, data, infrastructure, and capital. This creates dependencies that are fundamentally at odds with many of the values this community was built around.
We are also seeing increasingly direct consequences for people. Skilled professionals are losing work to automated systems. Organisations are becoming dependent on services they cannot inspect, understand, or control. AI-generated technical advice is often presented with confidence while lacking essential context. Automation is being used to scale abuse, manipulation, spam, and attacks.
At the same time, fear of these threats can itself push people and organisations towards a handful of large centralised providers, further reinforcing the same concentration of power.
The problem is not simply the existence of artificial intelligence as a technology. The problem is how these technologies are currently being developed, deployed, controlled, marketed, and increasingly presented as an inevitable replacement for human skills, judgement, creativity, and interaction.
Too often, the message coming from the companies selling these systems is essentially: *this is inevitable, it will replace what you do, and you will have to adapt to it whether you want to or not.*
At the BSD Cafe, we believe there must always be another choice.
For this reason, we ask members **not to use BSD Cafe to promote, advertise, or evangelise AI services, or the idea that replacing human work and decision-making with AI is inherently desirable, inevitable, or something everyone should simply accept.**
This does not prevent discussion.
But please do not use this community to promote the current AI-driven model of technological centralisation and human replacement.
**"Own your data" and the current state of AI are simply not compatible.**
@stefano
Lately there's a lot of fear and fundamentalism, and that's completely contrary to independent critical thinking. Mastodon is becoming almost like a church, where you can only think and argue in one direction. Not to mention the childish political arguments, full of insults and threats. And for me, politics and religion are matters for intellectually impoverished people.
All technology can be a tool or a weapon, depending on how each person uses it. That can't be controlled. -
New rule at BSD Cafe: Do not promote AI
BSD Cafe prefers to define itself by what it supports, rather than by what it opposes.
We support people, open technologies, decentralisation, technological independence, privacy, user control, the sharing of knowledge, and communities where technology remains a tool rather than becoming the authority.
For these reasons, we cannot support the current direction of AI and the ecosystem surrounding it.
Much of today's AI technology is developed and controlled by a very small number of extremely large companies, through an unprecedented concentration of computing power, data, infrastructure, and capital. This creates dependencies that are fundamentally at odds with many of the values this community was built around.
We are also seeing increasingly direct consequences for people. Skilled professionals are losing work to automated systems. Organisations are becoming dependent on services they cannot inspect, understand, or control. AI-generated technical advice is often presented with confidence while lacking essential context. Automation is being used to scale abuse, manipulation, spam, and attacks.
At the same time, fear of these threats can itself push people and organisations towards a handful of large centralised providers, further reinforcing the same concentration of power.
The problem is not simply the existence of artificial intelligence as a technology. The problem is how these technologies are currently being developed, deployed, controlled, marketed, and increasingly presented as an inevitable replacement for human skills, judgement, creativity, and interaction.
Too often, the message coming from the companies selling these systems is essentially: *this is inevitable, it will replace what you do, and you will have to adapt to it whether you want to or not.*
At the BSD Cafe, we believe there must always be another choice.
For this reason, we ask members **not to use BSD Cafe to promote, advertise, or evangelise AI services, or the idea that replacing human work and decision-making with AI is inherently desirable, inevitable, or something everyone should simply accept.**
This does not prevent discussion.
But please do not use this community to promote the current AI-driven model of technological centralisation and human replacement.
**"Own your data" and the current state of AI are simply not compatible.**
-
That would be a slippery slope given the lack of ethically-trained models. The policy is fine where it stands.
Also, I appreciate that it speaks to the social problems, where most people stick to the copyright, environmental, and cognitive problems in trying to warn against it.
-
That would be a slippery slope given the lack of ethically-trained models. The policy is fine where it stands.
Also, I appreciate that it speaks to the social problems, where most people stick to the copyright, environmental, and cognitive problems in trying to warn against it.
@mason "the lack of ethically-trained models"
Here you go: https://www.apertus-ai.org/
-
New rule at BSD Cafe: Do not promote AI
BSD Cafe prefers to define itself by what it supports, rather than by what it opposes.
We support people, open technologies, decentralisation, technological independence, privacy, user control, the sharing of knowledge, and communities where technology remains a tool rather than becoming the authority.
For these reasons, we cannot support the current direction of AI and the ecosystem surrounding it.
Much of today's AI technology is developed and controlled by a very small number of extremely large companies, through an unprecedented concentration of computing power, data, infrastructure, and capital. This creates dependencies that are fundamentally at odds with many of the values this community was built around.
We are also seeing increasingly direct consequences for people. Skilled professionals are losing work to automated systems. Organisations are becoming dependent on services they cannot inspect, understand, or control. AI-generated technical advice is often presented with confidence while lacking essential context. Automation is being used to scale abuse, manipulation, spam, and attacks.
At the same time, fear of these threats can itself push people and organisations towards a handful of large centralised providers, further reinforcing the same concentration of power.
The problem is not simply the existence of artificial intelligence as a technology. The problem is how these technologies are currently being developed, deployed, controlled, marketed, and increasingly presented as an inevitable replacement for human skills, judgement, creativity, and interaction.
Too often, the message coming from the companies selling these systems is essentially: *this is inevitable, it will replace what you do, and you will have to adapt to it whether you want to or not.*
At the BSD Cafe, we believe there must always be another choice.
For this reason, we ask members **not to use BSD Cafe to promote, advertise, or evangelise AI services, or the idea that replacing human work and decision-making with AI is inherently desirable, inevitable, or something everyone should simply accept.**
This does not prevent discussion.
But please do not use this community to promote the current AI-driven model of technological centralisation and human replacement.
**"Own your data" and the current state of AI are simply not compatible.**
-
J jwcph@helvede.net shared this topic
agree and fully understand. So what is your opinion about last