I've been saying "if AI is making you so productive then where is all this great new software" and I guess the answer is the software is out there it's just not great, it's terrible, and nobody is using it
-
@eniko This weekend my old demoscene crew had a meetup at my place. We're from the 80s home computer scene, the greybeards that speak assembler natively and there's not a single algorithm we're not able to implement from scratch.
All of us now use LLM-aids with the experiences that we're much more productive.
(And we're not having them forced upon us, we're choosing to use them and we choose for which projects)
Here are three of my programs that wouldn't have existed without the productivity boost using LLMs have given me:
https://codeberg.org/troed/meshnetmon
https://www.curseforge.com/hytale/mods/lightautomation
https://www.curseforge.com/hytale/mods/pathinator
... a fourth one is coming up shortly, for Nextcloud. Additionally I've done one feature development and fixed one bug in open source projects.
The truth is that LLMs got good at software development around six months ago. Those who claim differently seem to not be developers ...
@troed yeah man I do os dev and programming language development and write triangle rasterizers as a hobby because I'm not actually a developer. Pay no heed to the award winning indie game running on a custom engine and framework of my own design
Get the fuck outta here with your greybeard rank pulling nonsense
-
this post has made all the AI users Big Mad
@eniko evergreen post
-
this post has made all the AI users Big Mad
@eniko My condolences for your mentions, and also, wow digging through replies makes it easy to block a bunch of slopbros all at once.
-
@eniko This weekend my old demoscene crew had a meetup at my place. We're from the 80s home computer scene, the greybeards that speak assembler natively and there's not a single algorithm we're not able to implement from scratch.
All of us now use LLM-aids with the experiences that we're much more productive.
(And we're not having them forced upon us, we're choosing to use them and we choose for which projects)
Here are three of my programs that wouldn't have existed without the productivity boost using LLMs have given me:
https://codeberg.org/troed/meshnetmon
https://www.curseforge.com/hytale/mods/lightautomation
https://www.curseforge.com/hytale/mods/pathinator
... a fourth one is coming up shortly, for Nextcloud. Additionally I've done one feature development and fixed one bug in open source projects.
The truth is that LLMs got good at software development around six months ago. Those who claim differently seem to not be developers ...
-
@ehproque @gabrielesvelto the even worse news is that the bug fixes introduce 5x as many new bugs
@eniko @ehproque @gabrielesvelto In our glorious ai-forward nation, none shall go hungry! Double prion rations for all!
-
RE: https://mastodon.social/@jplebreton/116709297088742776
I've been saying "if AI is making you so productive then where is all this great new software" and I guess the answer is the software is out there it's just not great, it's terrible, and nobody is using it
@eniko The AI software is out there, it's terrible, and we are being forced to use it. Every bug you run into in your bank or transit system: that's AI.
-
@eniko This weekend my old demoscene crew had a meetup at my place. We're from the 80s home computer scene, the greybeards that speak assembler natively and there's not a single algorithm we're not able to implement from scratch.
All of us now use LLM-aids with the experiences that we're much more productive.
(And we're not having them forced upon us, we're choosing to use them and we choose for which projects)
Here are three of my programs that wouldn't have existed without the productivity boost using LLMs have given me:
https://codeberg.org/troed/meshnetmon
https://www.curseforge.com/hytale/mods/lightautomation
https://www.curseforge.com/hytale/mods/pathinator
... a fourth one is coming up shortly, for Nextcloud. Additionally I've done one feature development and fixed one bug in open source projects.
The truth is that LLMs got good at software development around six months ago. Those who claim differently seem to not be developers ...
I'll take this from your bio:
> discuss based in facts and actively despise "but my gut feeling .." arguments.
Now what you have stated is more of a "feels" statement rather than anything of fact.
In addition, given the commit history and timelines, I wouldn't say I am impressed by the work in the time given - However, hard for myself factor in what things you had going on and what-not. I'll give benefit of the doubt but I am also non-caring of it.
However, since you want facts and I like linking papers to people
* "From Technical Debt to Cognitive and Intent Debt: Rethinking Software Health in the Age of AI" - https://arxiv.org/abs/2603.22106
* "Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task" - https://arxiv.org/abs/2506.08872
* "How AI impacts Skill Format" - https://arxiv.org/pdf/2601.20245, heck Anthropic paid for this one
* "AI Assistance Reduces Persistence and Hurts Independent Performance" - https://arxiv.org/abs/2604.04721
Now, onto another note:
Be mindful that it is awfully like you are just going around agitating people and being completely unaware or oblivious to the kind of experience some of these people have. Not everyone is here to have a discussion on these things with you and that if you want opposing opinions that would be happy to debate you, you are better off elsewhere.
Maybe these LLMs work for you and not for others but I for sure would say that you are no expert in education or learning and that LLMs "got good at software development" seems like an overstatement and one of those "feels" you seem to have while very likely ignoring many other aspects of software development where they haven't proven themselves or appear harmful.
-
I use local LLMs. They use as much energy as when you play a game.
Read more. Post less.
-
@troed yeah man I do os dev and programming language development and write triangle rasterizers as a hobby because I'm not actually a developer. Pay no heed to the award winning indie game running on a custom engine and framework of my own design
Get the fuck outta here with your greybeard rank pulling nonsense
@eniko I mean, you're just wrong. It's ok to be wrong.
-
That was indeed a paper from way back _before_ the models got good at development. I laughed at them then - but I have no issues admitting to them now being good and changing my mind.
-
I'll take this from your bio:
> discuss based in facts and actively despise "but my gut feeling .." arguments.
Now what you have stated is more of a "feels" statement rather than anything of fact.
In addition, given the commit history and timelines, I wouldn't say I am impressed by the work in the time given - However, hard for myself factor in what things you had going on and what-not. I'll give benefit of the doubt but I am also non-caring of it.
However, since you want facts and I like linking papers to people
* "From Technical Debt to Cognitive and Intent Debt: Rethinking Software Health in the Age of AI" - https://arxiv.org/abs/2603.22106
* "Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task" - https://arxiv.org/abs/2506.08872
* "How AI impacts Skill Format" - https://arxiv.org/pdf/2601.20245, heck Anthropic paid for this one
* "AI Assistance Reduces Persistence and Hurts Independent Performance" - https://arxiv.org/abs/2604.04721
Now, onto another note:
Be mindful that it is awfully like you are just going around agitating people and being completely unaware or oblivious to the kind of experience some of these people have. Not everyone is here to have a discussion on these things with you and that if you want opposing opinions that would be happy to debate you, you are better off elsewhere.
Maybe these LLMs work for you and not for others but I for sure would say that you are no expert in education or learning and that LLMs "got good at software development" seems like an overstatement and one of those "feels" you seem to have while very likely ignoring many other aspects of software development where they haven't proven themselves or appear harmful.
"Not everyone is here to have a discussion on these things"
I'm sure people can post in places that aren't public and open to comments in such cases.
'Now what you have stated is more of a "feels" statement rather than anything of fact.'
Strange - I so very much linked to the facts there, those source code repositories you know.
-
"Not everyone is here to have a discussion on these things"
I'm sure people can post in places that aren't public and open to comments in such cases.
'Now what you have stated is more of a "feels" statement rather than anything of fact.'
Strange - I so very much linked to the facts there, those source code repositories you know.
> I'm sure people can post in places that aren't public and open to comments in such cases.
Sure and people do not have to respect your opinion. At the moment, that is all you have here and given this reply, I think not many people should respect your opinion on this topic.
I gave you research articles on the topic which you can read up on and dispute if you doubt their claims or want to challenge those findings - This is much more concrete evidence than what you have provided here.
> Strange - I so very much linked to the facts there, those source code repositories you know.
The repos tell me nothing and aren't anything interesting nor do they really say much about the LLM usage really.
They just tell me how you use them and honestly I am not finding anything interesting from that - The only statement you made here was "three of my programs that wouldn't have existed without the productivity boost using LLMs have given me"
If you "feel" that way, sure! You can feel that way but it is meaningless outside of that.
At the moment, you seem to be replying to anti-llm sentiment with not much more than "you feel like it speeds you up". Take a moment to reflect on this, it is okay to be wrong.
-
@eniko I mean, you're just wrong. It's ok to be wrong.
-
> I'm sure people can post in places that aren't public and open to comments in such cases.
Sure and people do not have to respect your opinion. At the moment, that is all you have here and given this reply, I think not many people should respect your opinion on this topic.
I gave you research articles on the topic which you can read up on and dispute if you doubt their claims or want to challenge those findings - This is much more concrete evidence than what you have provided here.
> Strange - I so very much linked to the facts there, those source code repositories you know.
The repos tell me nothing and aren't anything interesting nor do they really say much about the LLM usage really.
They just tell me how you use them and honestly I am not finding anything interesting from that - The only statement you made here was "three of my programs that wouldn't have existed without the productivity boost using LLMs have given me"
If you "feel" that way, sure! You can feel that way but it is meaningless outside of that.
At the moment, you seem to be replying to anti-llm sentiment with not much more than "you feel like it speeds you up". Take a moment to reflect on this, it is okay to be wrong.
@ahto I'll try to dumb it down:
The claim was that LLMs don't help in creating programs that people actually use.
I gave three examples of apps that people use, and that wouldn't have existed without LLMs. Their existence is not in doubt, their usage is not in doubt.
I'm going to make a wild guess: There's nothing anyone can show you, ever, that will make you change your mind.
Correct?
(no need to tag eniko - just like all other antivaxxers, sorry, anti-AI fanatics, they block all facts that oppose their convictions)
-
@ahto I'll try to dumb it down:
The claim was that LLMs don't help in creating programs that people actually use.
I gave three examples of apps that people use, and that wouldn't have existed without LLMs. Their existence is not in doubt, their usage is not in doubt.
I'm going to make a wild guess: There's nothing anyone can show you, ever, that will make you change your mind.
Correct?
(no need to tag eniko - just like all other antivaxxers, sorry, anti-AI fanatics, they block all facts that oppose their convictions)
@troed Alright, now lets look at the statement:
> "if AI is making you so productive then where is all this great new software" and I guess the answer is the software is out there it's just not great, it's terrible, and nobody is using it
and look at the chart where usage is lowering and app reviews are also lowering but more app releases are occurring.... Which is drawing a conclusion that, yeah given the data.. these things aren't being used.
Now to quote you and what you are asserting as the claim.
> The claim was that LLMs don't help in creating programs that people actually use.
That statement is different from the original one. Cute that you tried to change it though!
There is a weighing on the software quality and the usage of it, along with the graph indicating that there is less usage overall given more application releases.
Fwiw, if I wanted to fully clarify what the claim is, I would have asked the original poster to clarify rather than just guessing, which is what you did.
I can't state what their full claim is but I can see semblance of truth in what is being stated given:
* Statement
* Chart
* Journals
* Social consensus (Not reliable but hey, I'll throw it in there)> I gave three examples of apps that people use
I guess you did, dang! I guess if I was to at least track some accounts on there it would be roughly 7 people out of those 3 repos? acknowledging the 5 followers on one and 2 comments left on two of them.
However, that is being generous that those people are using them.
> and that wouldn't have existed without LLM
Maybe for you but I could see someone else making these things or possibly they may already exist. I'm not finding these things special or requiring an LLM for them to exist.
> I'm going to make a wild guess: There's nothing anyone can show you, ever, that will make you change your mind.
If you want to come to that conclusion, that's fine. I understand how you could feel here and it is okay to also be wrong.
-
@troed @ZanaGB @eniko and what hardware was used to train those models? Whose data was in the training set? Was it properly licensed or was it obtained without consent, compensation or attribution? And what did it cost to gather that training data? Externalities matter, these things don't appear out of thin air
-
@troed @ZanaGB @eniko and what hardware was used to train those models? Whose data was in the training set? Was it properly licensed or was it obtained without consent, compensation or attribution? And what did it cost to gather that training data? Externalities matter, these things don't appear out of thin air
If you want to make completely different points to the one answered - sure!
Has any artist ever compensated another from having looked at their paintings while learning to draw?
Has any budding coder ever compensated others when having studied their code to learn how to do things?
I'm all for lambasting shitty tech bro AI companies, but that's not the same as claiming that any and all LLM usage is bad. I suggest looking at Mistral AI as a european company that's building datacenters using fully renewable energy and ethically sourced data.
-
@troed Alright, now lets look at the statement:
> "if AI is making you so productive then where is all this great new software" and I guess the answer is the software is out there it's just not great, it's terrible, and nobody is using it
and look at the chart where usage is lowering and app reviews are also lowering but more app releases are occurring.... Which is drawing a conclusion that, yeah given the data.. these things aren't being used.
Now to quote you and what you are asserting as the claim.
> The claim was that LLMs don't help in creating programs that people actually use.
That statement is different from the original one. Cute that you tried to change it though!
There is a weighing on the software quality and the usage of it, along with the graph indicating that there is less usage overall given more application releases.
Fwiw, if I wanted to fully clarify what the claim is, I would have asked the original poster to clarify rather than just guessing, which is what you did.
I can't state what their full claim is but I can see semblance of truth in what is being stated given:
* Statement
* Chart
* Journals
* Social consensus (Not reliable but hey, I'll throw it in there)> I gave three examples of apps that people use
I guess you did, dang! I guess if I was to at least track some accounts on there it would be roughly 7 people out of those 3 repos? acknowledging the 5 followers on one and 2 comments left on two of them.
However, that is being generous that those people are using them.
> and that wouldn't have existed without LLM
Maybe for you but I could see someone else making these things or possibly they may already exist. I'm not finding these things special or requiring an LLM for them to exist.
> I'm going to make a wild guess: There's nothing anyone can show you, ever, that will make you change your mind.
If you want to come to that conclusion, that's fine. I understand how you could feel here and it is okay to also be wrong.
@ahto So you're _guessing_ that if I hadn't created those programs someone else would've and that means that the argument that LLMs indeed produce apps that people use is wrong?
I'm sure you believe that you're great at debating.
-
If you want to make completely different points to the one answered - sure!
Has any artist ever compensated another from having looked at their paintings while learning to draw?
Has any budding coder ever compensated others when having studied their code to learn how to do things?
I'm all for lambasting shitty tech bro AI companies, but that's not the same as claiming that any and all LLM usage is bad. I suggest looking at Mistral AI as a european company that's building datacenters using fully renewable energy and ethically sourced data.
@troed not the same thing and you know it. People looking at things and storing copies of someone else's potentially copyrighted data for training are two completely different things. Is it so hard to admit that there are externalities and they are bad no matter how you slice it?
Mistral AI is your run-of-the-mill AI company that does not disclose what's in their training sets, just like everybody else: https://help.mistral.ai/en/articles/347390-does-mistral-disclose-its-training-datasets
-
@troed not the same thing and you know it. People looking at things and storing copies of someone else's potentially copyrighted data for training are two completely different things. Is it so hard to admit that there are externalities and they are bad no matter how you slice it?
Mistral AI is your run-of-the-mill AI company that does not disclose what's in their training sets, just like everybody else: https://help.mistral.ai/en/articles/347390-does-mistral-disclose-its-training-datasets
@gabrielesvelto LLMs don't "store copies" when they train. What happens in their neural networks is very similar to what happens in a human brain when learning.
"run of the mill": https://mistral.ai/news/our-contribution-to-a-global-environmental-standard-for-ai/
