To make sure that AI never uses your work (source code or binaries) for training, I see two possible solutions:
- the software would run only on your own server and clients will access it remotely through an interface
- you will sell the software together with the hardware and a proprietary operating system, which does not allow anyone to get the ones and zeros of your software
So, the only way to prevent GTA6 being "decompiled" and "forked" by AI models is to run it remotely (with a video interface), or sell it together with a proprietary console :D
>most oss is written with the expectation of no financial reward.
Most open source software is written by corporations which have a complimentary motive. Think hardware manufacturers writing drivers because it makes their products more saleable or Google writing Go because they needed a systems programming language.
It isn't true ownership if you depend on an LLM you don't own to modify the code in ways you don't understand. That's just proprietary software with extra steps.
By that standard, couldn't I say you don't truly own open source software like Chromium or OpenSSL because very few are capable of understanding the codebase well enough to implement changes?
That's always been the bar for FOSS, we've just failed as a community at education because we became elitist and decided the muggles didn't actually deserve to understand anything. The end goal of the free software movement should be that everyone has that understanding, that code literacy should be as essential as reading and math. But yes, in short, you don't truly own code you don't understand. Simply having the code available is better than not, but a freedom you can't exercise really isn't a freedom, and all of the freedoms of free software presuppose the ability to comprehend software.
And there's a difference between few being capable of implementing changes, and anyone. In a world where most or all code is written and edited by LLMs, and CS courses are focused on how to prompt LLMs and no longer on fundamentals, the opportunities for humans to learn programming will all but vanish. And this assumes LLMs will even expose code at all. They may simply emit precompiled applications in the future. If humans remain in the loop even when LLMs are used then this becomes less of a problem. But the culture around LLMs converges on vibe-coding - as little direct human understanding or effort as possible. As such it converges on sacrificing freedom for convenience.
Who is going to write that software if there is no reward.
To make sure that AI never uses your work (source code or binaries) for training, I see two possible solutions:
- the software would run only on your own server and clients will access it remotely through an interface
- you will sell the software together with the hardware and a proprietary operating system, which does not allow anyone to get the ones and zeros of your software
So, the only way to prevent GTA6 being "decompiled" and "forked" by AI models is to run it remotely (with a video interface), or sell it together with a proprietary console :D
most oss is written with the expectation of no financial reward. there is more to life than money.
not to mention there are alternative revenue sources anyways.
>most oss is written with the expectation of no financial reward.
Most open source software is written by corporations which have a complimentary motive. Think hardware manufacturers writing drivers because it makes their products more saleable or Google writing Go because they needed a systems programming language.
Could you please mention the alternative revenue sources.
It isn't true ownership if you depend on an LLM you don't own to modify the code in ways you don't understand. That's just proprietary software with extra steps.
That's a pretty high bar.
By that standard, couldn't I say you don't truly own open source software like Chromium or OpenSSL because very few are capable of understanding the codebase well enough to implement changes?
That's always been the bar for FOSS, we've just failed as a community at education because we became elitist and decided the muggles didn't actually deserve to understand anything. The end goal of the free software movement should be that everyone has that understanding, that code literacy should be as essential as reading and math. But yes, in short, you don't truly own code you don't understand. Simply having the code available is better than not, but a freedom you can't exercise really isn't a freedom, and all of the freedoms of free software presuppose the ability to comprehend software.
And there's a difference between few being capable of implementing changes, and anyone. In a world where most or all code is written and edited by LLMs, and CS courses are focused on how to prompt LLMs and no longer on fundamentals, the opportunities for humans to learn programming will all but vanish. And this assumes LLMs will even expose code at all. They may simply emit precompiled applications in the future. If humans remain in the loop even when LLMs are used then this becomes less of a problem. But the culture around LLMs converges on vibe-coding - as little direct human understanding or effort as possible. As such it converges on sacrificing freedom for convenience.
3 replies →