AI and Creative Works in Progress
Are tech companies allowing their insatiable LLM's to consume everything in their sphere of influence?
[Note to readers, I have not published anything the last week because substack no longer functions on my linux computer. I put this together on my phone. Thank you for your patience while I figure out how to proceed.)
A strange thing happened in my Substack dashboard recently. As a writer I have access to a lot of different metrics, subscriber numbers, followers, views of recent posts etc. One on these metrics is the 30 day count. I see how much higher or lower my numbers are compared to the previous thirty days.
My 30 day number went up 300%, just like that, about two weeks ago. But there was no equivalent jump in any of my recent posts or my subscriber numbers.
I only know because the 30 day view number was cut in half after the election. That was not surprising, as I made some changes to the content here, my political writing during the Biden era was more popular than my esoteric writing, I was writing more regularly then, and I think people have been on Substack less, generally.
The only explanation I have is, the many iterations of LLMs, Large Language Model AI, consuming everything in this blog equally. There is three years of content.
I suppose substack admin might have made some sort of change, maybe adding notes traffic to the count? Good luck finding out though. The other thing I have learned in this process is, Substack Support is as remote as a deity.
Shortly after I noticed that uptick in the 30 day count, substack stopped working entirely on my Linux computer. And
has been absolutely silent about that.Now, I am not a techie, so owning an open-source computer is an odd choice for me. Nor am I a luddite, I use tech to my advantage in whatever way I am able. I have built websites using software for that purpose, I could not build the software or the website from scratch using code. When there is an issue I have at times been able to find command tutorials online and work it out in the terminal.
I chose Linux in part because the last time I checked, I could not use a lot of word processing software, particularly Microsoft Word, unless I was online. Not only is that inconvenient, color me paranoid, there seems something nefarious, Microsoft requiring me to be online if I am using their product.
My Linux is only a few years old. But now it feels a bit like, if I am to use Substack as a writing platform, I have to be using PC, Apple, or Chromebook. This is not the first time substack went offline on my computer, just the longest.
PC is Microsoft based, Azure is their AI. Apple has Intelligence, Chromebook is formerly do-no-evil Google, which is Gemini.
Which has me wondering, is my work being consumed by AI as I write it? Is the Substack writing platform secure? Is any platform secure on the internet?
Is AI privy to my works in progress? And if so, is my work going to appear somewhere in the world, before I have even published it? If so, what possible recourse would I have? Who would even care?
But I am pretty certain, if content creators feel like their works in progress are being stolen even before they are finished, that will break the internet.
In that respect, I have a writing project I have been working on for the last eight months, 10,000+/- words, a series of poems. I hand wrote the first draft. Normally the second draft of projects like this, I write on a computer. But this time, the second draft will be by hand, and probably too, all the drafts. I intend to bind and publish my own books.
As long as I am able AI will not have access to that language.
There is not anything incendiary about that work, it is not political at all. It is simply my Art, and this is a matter of principle.
Maybe I do not really understand how all this tech really works, but I am increasingly certain Substack Support is not going to be much help understanding.
These are the people I know on Substack, who can speak to this more accurately and eloquently than I can, and if I have brought up any worthwhile points, I hope you will expand. If I am a fool I hope you will say so. If you have any advice how I should proceed, so I don’t have to write Substack posts on my phone, I would be happy to hear it. For all of Substack’s faults, I still feel it is the best venue around, for writers to connect and talk about these issues.
, , , , , ,
I'm still using a copy of Word from a CD that I bought in 1997. It runs offline just fine, go figure. My current desktop runs Windows 10 and has started interrupting me to warn that support is ending soon. I don't particularly care, I can run my own antivirus software, but it's just a reminder that my computer is getting long in the tooth. I have a backup one, new in the box, also running Windows 10 that I'll switch to if/when this one dies. But I can feel the day coming when drastic decisions will have to be made. Their efforts to herd us into the cattle shoot get more overt with time. You can feel the walls closing in. My bet is still on the humans, but not without a fight.
When you say no longer works on you Linux box, is that the app or via the browser? The form factor of the App is not all that great, I haven't tried writing on it yet.
Re AI snagging our stuff, the ToS seems to be better that FaceSpook and heX. But then again, you have to catch someone violating a license, then you have to litigate, and that's an impediment. The stuff out in the public is going to pilfered by the Bro-ligarchs, because our wanting to retain intellectual property rights prevents the Overmind AI from coming into being. So they'll continue to steal and hope that Congress passes the AI moratorium in some format.
As far as stuff behind the paywall, you'd hope that would be kept away from the AI scrapers.