Lawyer-turned-software-developer reporting in! (American, to be clear…)
My concern with AI generated code has to do with software licensing. Mainly…. What is the license of AI generated code??? What if it regurgitates entire sections of proprietary code? Unless I’ve totally missed some legal/tech news, these questions are FAR from settled.
I feel like I’m taking crazy pills sometimes, but that seems like something the Linux kernel community should be at least mildly curious about….
Look, to be fair, maybe they are! I don’t follow the kernel mailing list; I don’t know what they talk about. I’m just a humble backend developer/infrastructure guy.
But I know it worries me… personally, I mostly use LLMs in my own work via asking questions about code I’ve already written. Then I look at suggestions it’s made, or potential issues it points out, and decide what to do accordingly. But that’s just me!
Yeah, that is a pretty shitty consequence, but it’s nothing new or different from the strandard clean room techniques the industry has been using for decades.
Unironically a AI’s biggest win. Absolutely fair criticism on art and weak imitations of human creativity, but code is just math with syntactic sugar. “Owning” algorithms and hiding implementations is such an absurd concept.
I often wonder how much better and more accessible software would be if everyone wasn’t reinventing nearly every wheel all the time. What if every phone could copy-paste an Airdrop implementation? What if printers just worked? What if you could put adblockers on any device? What if every online game could be self hosted and modded? AI is (for the moment) giving an approximation of that timeline and I’m here for it.
I do write code. Code only exists as a high level abstraction that codifies logic and allows us to talk about the function of the silicone without getting bogged down in details. The utility of code depends entirely on the context of the people looking at it (eg. different code works better for a small shop of experienced devs vs a large scale enterprise vs a FOSS project).
I can show you a for loop and talk in simple English about what it’s doing or write a formal proof showing that it meets the problem parameters. I could show you the compiled ASM and we would struggle to have the same conversation even with the exact same functionality. Same math, different syntactic sugar.
The process of building software is where humans will always be needed (identify the human meat space problems, define features, manage risk and cost, set delivery expectations, etc). The specifics of how we create the code within that process isn’t as crucial.
I’m incredibly against the big LLM SaaS providers and material waste besides. Anyway I don’t have a crystal ball and I’m highly sceptical that this technology will work out for average people, but it sure would be interesting if it somehow lead to less corporate control and cultural enclosure rather than how it seems to be going right now.
What if every phone could copy-paste an Airdrop implementation?
That’s apple’s decision, one of many to force their users to only get their devices, similarly not being able to host your own game servers benefits the publishers/ devs. (Edit for clarity: by this I mean that this state of things isn’t caused by unavailability of open source solutions but rather by corporate greed.)
“Owning” algorithms and hiding implementations is such an absurd concept.
Not any different from protecting one’s art creations. You spend a bunch of time learning how to code and are doing it for a living so you don’t want others/ competition to take it and sell as their own. Not sure what’s hard to understand.
That’s true of course. And it’s true for a lot of people in other areas as well.
In vacuum I’m all for open source software, in fact I would like it a lot if everything was open. But reality is that as a developer in a smaller software company we’d be out of our jobs immediately as other, larger companies would take everything that’s worth taking and use their larger budgets to out advertise their copy.
Lawyer-turned-software-developer reporting in! (American, to be clear…)
My concern with AI generated code has to do with software licensing. Mainly…. What is the license of AI generated code??? What if it regurgitates entire sections of proprietary code? Unless I’ve totally missed some legal/tech news, these questions are FAR from settled.
I feel like I’m taking crazy pills sometimes, but that seems like something the Linux kernel community should be at least mildly curious about….
Look, to be fair, maybe they are! I don’t follow the kernel mailing list; I don’t know what they talk about. I’m just a humble backend developer/infrastructure guy.
But I know it worries me… personally, I mostly use LLMs in my own work via asking questions about code I’ve already written. Then I look at suggestions it’s made, or potential issues it points out, and decide what to do accordingly. But that’s just me!
I’m more miffed about GPL or copyleft code going to proprietary software and then being profited from without any attribution. It’s a poison pill.
Yeah, that is a pretty shitty consequence, but it’s nothing new or different from the strandard clean room techniques the industry has been using for decades.
If it regurgitates proprietary code the model should not have had access to for training data… well…
That street goes both ways, there’s more GPLed code in the training corpus than there is leaked proprietary code.
Unironically a AI’s biggest win. Absolutely fair criticism on art and weak imitations of human creativity, but code is just math with syntactic sugar. “Owning” algorithms and hiding implementations is such an absurd concept.
I often wonder how much better and more accessible software would be if everyone wasn’t reinventing nearly every wheel all the time. What if every phone could copy-paste an Airdrop implementation? What if printers just worked? What if you could put adblockers on any device? What if every online game could be self hosted and modded? AI is (for the moment) giving an approximation of that timeline and I’m here for it.
“code is just math with syntactic sugar”
Wow!
Tell me you don’t write code wihout telling me you don’t write code.
I do write code. Code only exists as a high level abstraction that codifies logic and allows us to talk about the function of the silicone without getting bogged down in details. The utility of code depends entirely on the context of the people looking at it (eg. different code works better for a small shop of experienced devs vs a large scale enterprise vs a FOSS project).
I can show you a
forloop and talk in simple English about what it’s doing or write a formal proof showing that it meets the problem parameters. I could show you the compiled ASM and we would struggle to have the same conversation even with the exact same functionality. Same math, different syntactic sugar.The process of building software is where humans will always be needed (identify the human meat space problems, define features, manage risk and cost, set delivery expectations, etc). The specifics of how we create the code within that process isn’t as crucial.
I’m incredibly against the big LLM SaaS providers and material waste besides. Anyway I don’t have a crystal ball and I’m highly sceptical that this technology will work out for average people, but it sure would be interesting if it somehow lead to less corporate control and cultural enclosure rather than how it seems to be going right now.
But what about the poor corporations making those games that already fired all the devs that made it after overworking them?
That’s apple’s decision, one of many to force their users to only get their devices, similarly not being able to host your own game servers benefits the publishers/ devs. (Edit for clarity: by this I mean that this state of things isn’t caused by unavailability of open source solutions but rather by corporate greed.)
Not any different from protecting one’s art creations. You spend a bunch of time learning how to code and are doing it for a living so you don’t want others/ competition to take it and sell as their own. Not sure what’s hard to understand.
Most programmers earn a wage, not the full value of their labor.
That’s true of course. And it’s true for a lot of people in other areas as well. In vacuum I’m all for open source software, in fact I would like it a lot if everything was open. But reality is that as a developer in a smaller software company we’d be out of our jobs immediately as other, larger companies would take everything that’s worth taking and use their larger budgets to out advertise their copy.
Software licensing is dead, you should really catch up because what are even talking about.