I feel really sorry for this person. This is purely my own view of the future, and I am no fortune teller. But the reality he is desperately clinging to is just gone. AI in software development is...
I feel really sorry for this person. This is purely my own view of the future, and I am no fortune teller. But the reality he is desperately clinging to is just gone.
AI in software development is here to stay. Denying this is genuine mental denial. I wont say that non-AI code will disappear, and hands-off vibecoding is still genuinely reckless, but I believe the amount of AI generated code is certainly already outpacing humans, and this is just the beginning.
As I mentioned in a cousin post: AI software development is not inevitable, it is not mandatory to accept it. If a project or community doesn't want it, the AI coders should shut up, fork, and...
As I mentioned in a cousin post:
AI software development is not inevitable, it is not mandatory to accept it. If a project or community doesn't want it, the AI coders should shut up, fork, and form their own community instead of trying to force their values into the community that is rejecting theirs.
I personally find a disclosure policy a good middle ground, allowing for that flexibility while still fostering a default stance of being human-first.
Yes. But there will be space for them to co-exist, and there is no reason to try to force acceptance where it is not wanted. Which is the biggest problem a lot of projects and communities are facing.
Yes. But there will be space for them to co-exist, and there is no reason to try to force acceptance where it is not wanted. Which is the biggest problem a lot of projects and communities are facing.
I'm not familiar with Bevy or it's community, but it sounds like perhaps this is actually the inverse case? The above goes both ways.
If a project or community doesn't want it, the AI coders should shut up, fork, and form their own community instead of trying to force their values into the community that is rejecting theirs.
I'm not familiar with Bevy or it's community, but it sounds like perhaps this is actually the inverse case? The above goes both ways.
I don't disagree, but given the initial AI policy of 'hard ban' I think that is not the case. My read is that the culture has shifted, and that making life easier on the maintaners was the primary...
I don't disagree, but given the initial AI policy of 'hard ban' I think that is not the case.
My read is that the culture has shifted, and that making life easier on the maintaners was the primary goal. The author may well have lost their space already, but I think that is a worthy and justified bemoaning if the pro-AI crowd is newer and displaced the critics entirely.
Are cultures not allowed to shift or people change their minds? I think the nazi bar analogy is a bad one for obvious reasons. Cultures of bars shift all the time as people age out, kids...
Are cultures not allowed to shift or people change their minds?
I think the nazi bar analogy is a bad one for obvious reasons. Cultures of bars shift all the time as people age out, kids re-inventing underused watering holes is almost a national pastime here. Not every shift in culture is bad as in nazi bad.
I agree, and I think this is going to be one of the big generational divides going forward. It'll be very interesting to see how the 'AI-natives' turn out... I really hope that most of them can...
I agree, and I think this is going to be one of the big generational divides going forward. It'll be very interesting to see how the 'AI-natives' turn out... I really hope that most of them can find interesting and novel ways to make use of AI instead of just letting their skills atrophy (or never develop at all). I think there's all kinds of potential for them if they can reach it.
An observation of how a change in AI policy has affected the Bevy game engine community in the months since a new AI policy came into effect. As can be inferred from the title, the author, a...
An observation of how a change in AI policy has affected the Bevy game engine community in the months since a new AI policy came into effect.
As can be inferred from the title, the author, a long-time Bevy contributor, has had a very negative experience with the changes in the community. Their post describes the new policy's effects on the behavior and health of the development community, and how the changes in the community have directly decreased the author's willingness to continue participating in the Bevy project.
The author makes their disgust with AI very obvious, but disgust is not a good reason for prejudice. Apparently they want a "safe space" for their anti-AI prejudices? They object to a "no dunking"...
The author makes their disgust with AI very obvious, but disgust is not a good reason for prejudice.
Apparently they want a "safe space" for their anti-AI prejudices? They object to a "no dunking" rule that would prevent people from freely expressing their anti-AI prejudices using memes?
I wouldn't want to be in that kind of community.
I expect that it's only going to get worse though. Anti-AI populism is on the rise and uncritical boosters don't help either. They're going to feed each other.
Maybe we will need safe spaces for reasonable people who prefer evidence-based discussions?
That's not really what I took away from this at all. And your response does feel somewhat like you're validating this line: I think there a few key points: LLM-coded projects are not exactly great...
That's not really what I took away from this at all. And your response does feel somewhat like you're validating this line:
Meanwhile, almost all arguments against "AI" use were more or less thrown out or ignored.
I think there a few key points:
LLM-coded projects are not exactly great for the purposes of sharing and learning from each other. If the LLM spits something out, and a human asks me a question about it, I now basically have to answer that question for myself first. Handcoding it myself means I'm answering from myself. And I can get the frustration if the goal of a space was sharing lessons learned on the journey and not 'look at the end product.' I can't really speak to whteher that space really is/was that, but overall that sentiment is reasonable IMO.
Having done it for several months now, LLM coding is largely orthagonal to learning, especially when coming in to unknown territory. You don't really learn much until the LLM screws up. Then you're mostly bughunting rather than writing.
A no dunking rule does invite the Nazi Bar problem, particularly in an open space. If Bevey was an anti-AI (or at least AI-skeptical) community, placing the newer AI boosters on level ground with the old community is not a neutral move; it's practically an invitation to displace the old culture.
The bullet list at the end is the typical laundry list of a community getting hit with an Eternal September.
For my own views that echo the author's:
I think it's reasonable to allow AI-assisted contributions with a disclosure policy, while also allowing for Tildes-style "We're rejecting this because we decided to" PR policy. As well as not necessarily further welcoming AI boosrers into a space that previously rejected them.
In this era when any project can easily become flooded, project owners need to be given the grace to straight up reject anything without writing up a massive well-reasoned rejection note.
AI use is not inevitable. If a given community doesn't want it, and they ban it, then maybe the AI users should respect that, fork, and form their own community instead of trying to force their stance on the existing community. If their AI methods truely are superior, then they'll leave the old projects in the dust and both communities can live in peace apart from each other.
Reconciling this is more important on giant projects like the Linux kernel. Less so on a boutique project that I'm hearing about for the first time.
Sorry to dig into your take, but it already seems like the rest of the world is the safe space for people who want to discuss how LLMs/generative AI can be deployed for use in day to day life? The...
Sorry to dig into your take, but it already seems like the rest of the world is the safe space for people who want to discuss how LLMs/generative AI can be deployed for use in day to day life? The idea that there's widespread negative perception is largely an effect of the anglosphere bubble that many of us live in -- especially for those in the tech sector (which foolishly for decades decided that unionization was for dumb losers and continuously refused to organize >:| zero sympathy there from me tbh). The article author's take that the average person isn't accepting of this stuff is not reflective of reality -- the average American, sure, but there's more than just one country in the world that can buy things.
And anti-AI populism seems economically self defeating overall, so it's probably not something to be concerned with directly? I'd be moreso worried about it spiraling off and turning into another antivax -> Q-anon pipeline, but I don't know if a politically compatible conspiracy theory group exists. If all they do is burn down local data centres and protest in discord, it's not going to be a practical concern for anyone else.
IMO as I'd noted in another thread, the main impact from this will be excluding otherwise very motivated voices from discussions on how to realistically address the fallout from AI tooling existing. That seems like an objectively bad thing, but if people want to opt out, I don't see a reason not to support them in doing so?
I frequently go to live drawing sessions, cameras are forbidden in this spaces out of respect to the model. Nobody goes crazy about it, claiming that people are living in the past and that there's...
I frequently go to live drawing sessions, cameras are forbidden in this spaces out of respect to the model. Nobody goes crazy about it, claiming that people are living in the past and that there's no longer value in drawing with a pencil when a camera can do it faster. At the same time, I can guarantee that most people on those session also use photographs to study the human on their own. A good artists understands that drawing from live and drawing from a reference are completely different experiences. A camera will flatten the color and values not matter how good it is. At the same time, a good artist understands that is not necessary to do everything from scratch and that the right tool will save you a lot of time of money.
Whenever a conversation like this comes out, where people want their own space, others are quick to say that "they are trying to hold on the past, that world no longer exists". I wonder if this is unique to software. Do people tell others that there's no longer value in learning to draw, learning how an engine works and built your own model, or learning how to create a mechanical computer from scratch?
People understand that things are different in the professional world and the hobbyist communities. Why can't software be different. There's still value in struggling, learning things on your own and talking with other people, sharing ideas. Otherwise why are you in tildes and not talking to an LLM?
Wanting a space, free of LLMs should be an uncontroversial opinion, the same way that wanting a space free of smart phones should be. I understand that this a game engine that can be used in real games, not just hobbyist projects, but I still don't see the point on dismissing a person that wants to create a space only for humans. If someone wants to create that space, the only thing there's to do is respect their decision.
I feel really sorry for this person. This is purely my own view of the future, and I am no fortune teller. But the reality he is desperately clinging to is just gone.
AI in software development is here to stay. Denying this is genuine mental denial. I wont say that non-AI code will disappear, and hands-off vibecoding is still genuinely reckless, but I believe the amount of AI generated code is certainly already outpacing humans, and this is just the beginning.
As I mentioned in a cousin post:
AI software development is not inevitable, it is not mandatory to accept it. If a project or community doesn't want it, the AI coders should shut up, fork, and form their own community instead of trying to force their values into the community that is rejecting theirs.
I personally find a disclosure policy a good middle ground, allowing for that flexibility while still fostering a default stance of being human-first.
In a project, not per se, in general, it is inevitable
Yes. But there will be space for them to co-exist, and there is no reason to try to force acceptance where it is not wanted. Which is the biggest problem a lot of projects and communities are facing.
Sounds like it would be the other way around if the author thinks the admins are compromised.
I'm not familiar with Bevy or it's community, but it sounds like perhaps this is actually the inverse case? The above goes both ways.
I don't disagree, but given the initial AI policy of 'hard ban' I think that is not the case.
My read is that the culture has shifted, and that making life easier on the maintaners was the primary goal. The author may well have lost their space already, but I think that is a worthy and justified bemoaning if the pro-AI crowd is newer and displaced the critics entirely.
Are cultures not allowed to shift or people change their minds?
I think the nazi bar analogy is a bad one for obvious reasons. Cultures of bars shift all the time as people age out, kids re-inventing underused watering holes is almost a national pastime here. Not every shift in culture is bad as in nazi bad.
I agree, and I think this is going to be one of the big generational divides going forward. It'll be very interesting to see how the 'AI-natives' turn out... I really hope that most of them can find interesting and novel ways to make use of AI instead of just letting their skills atrophy (or never develop at all). I think there's all kinds of potential for them if they can reach it.
Truly unprecedented times we live in.
An observation of how a change in AI policy has affected the Bevy game engine community in the months since a new AI policy came into effect.
As can be inferred from the title, the author, a long-time Bevy contributor, has had a very negative experience with the changes in the community. Their post describes the new policy's effects on the behavior and health of the development community, and how the changes in the community have directly decreased the author's willingness to continue participating in the Bevy project.
The author makes their disgust with AI very obvious, but disgust is not a good reason for prejudice.
Apparently they want a "safe space" for their anti-AI prejudices? They object to a "no dunking" rule that would prevent people from freely expressing their anti-AI prejudices using memes?
I wouldn't want to be in that kind of community.
I expect that it's only going to get worse though. Anti-AI populism is on the rise and uncritical boosters don't help either. They're going to feed each other.
Maybe we will need safe spaces for reasonable people who prefer evidence-based discussions?
That's not really what I took away from this at all. And your response does feel somewhat like you're validating this line:
I think there a few key points:
For my own views that echo the author's:
I think it's reasonable to allow AI-assisted contributions with a disclosure policy, while also allowing for Tildes-style "We're rejecting this because we decided to" PR policy. As well as not necessarily further welcoming AI boosrers into a space that previously rejected them.
In this era when any project can easily become flooded, project owners need to be given the grace to straight up reject anything without writing up a massive well-reasoned rejection note.
AI use is not inevitable. If a given community doesn't want it, and they ban it, then maybe the AI users should respect that, fork, and form their own community instead of trying to force their stance on the existing community. If their AI methods truely are superior, then they'll leave the old projects in the dust and both communities can live in peace apart from each other.
Reconciling this is more important on giant projects like the Linux kernel. Less so on a boutique project that I'm hearing about for the first time.
Sorry to dig into your take, but it already seems like the rest of the world is the safe space for people who want to discuss how LLMs/generative AI can be deployed for use in day to day life? The idea that there's widespread negative perception is largely an effect of the anglosphere bubble that many of us live in -- especially for those in the tech sector (which foolishly for decades decided that unionization was for dumb losers and continuously refused to organize >:| zero sympathy there from me tbh). The article author's take that the average person isn't accepting of this stuff is not reflective of reality -- the average American, sure, but there's more than just one country in the world that can buy things.
And anti-AI populism seems economically self defeating overall, so it's probably not something to be concerned with directly? I'd be moreso worried about it spiraling off and turning into another antivax -> Q-anon pipeline, but I don't know if a politically compatible conspiracy theory group exists. If all they do is burn down local data centres and protest in discord, it's not going to be a practical concern for anyone else.
IMO as I'd noted in another thread, the main impact from this will be excluding otherwise very motivated voices from discussions on how to realistically address the fallout from AI tooling existing. That seems like an objectively bad thing, but if people want to opt out, I don't see a reason not to support them in doing so?
I frequently go to live drawing sessions, cameras are forbidden in this spaces out of respect to the model. Nobody goes crazy about it, claiming that people are living in the past and that there's no longer value in drawing with a pencil when a camera can do it faster. At the same time, I can guarantee that most people on those session also use photographs to study the human on their own. A good artists understands that drawing from live and drawing from a reference are completely different experiences. A camera will flatten the color and values not matter how good it is. At the same time, a good artist understands that is not necessary to do everything from scratch and that the right tool will save you a lot of time of money.
Whenever a conversation like this comes out, where people want their own space, others are quick to say that "they are trying to hold on the past, that world no longer exists". I wonder if this is unique to software. Do people tell others that there's no longer value in learning to draw, learning how an engine works and built your own model, or learning how to create a mechanical computer from scratch?
People understand that things are different in the professional world and the hobbyist communities. Why can't software be different. There's still value in struggling, learning things on your own and talking with other people, sharing ideas. Otherwise why are you in tildes and not talking to an LLM?
Wanting a space, free of LLMs should be an uncontroversial opinion, the same way that wanting a space free of smart phones should be. I understand that this a game engine that can be used in real games, not just hobbyist projects, but I still don't see the point on dismissing a person that wants to create a space only for humans. If someone wants to create that space, the only thing there's to do is respect their decision.