Post

i think the scummiest part of LLM codegen is that you never know if you're getting an amalgamation of some common idea it's been trained on a million copies of, or just getting One Guy's passion project with the identifiers changed a bit, a thing i have overheard happening more than once

you'd get this same sort of problem generating art or writing, of course, but a number of programmers do not seem to understand that the thing they do is a creative pursuit, as if the existence of a test suite makes it a hard science

so we land instead on this industry-wide authorship laundering, encouraged by having the output come from a natural language interface with a cutesy name that makes it seem like a Guy rather than a fuzzy database with all the attribution filed off

and this isn't about copyright or whatever, just about common human decency between people who Make Things. artists of all stripes seem to understand this pretty intuitively, that you give nods to those you borrow from, because... because that's why we do any of this, to share and inspire

and in fact the theft is arguably worse with programming — you can't share artwork without sharing its pixels, and you can't share writing without sharing its words, but you CAN share a program without sharing its code and in many cases that is the default thing that happens

so codegen LLMs /must/ mostly be trained on code that someone actively decided to share, that its authors wished to contribute to the commons! this industry was just recently fretting over how to support the open source projects it relies on, too. now it exploits them

and hey it turns out that the industry is quite happy to pay millions of dollars for open source code! just not to its developers. no, all that money goes to fucking GOOGLE AND MICROSOFT because they're the ones operating the vacuum cleaners. fucking ludicrous state of affairs.

if you have a clever or novel idea, or if you just do the grunt work for some obscure task, your reward is now to have your name erased, to have "Co-authored-by: Claude" stapled on in its place, to have it treated like it just appeared from the æther by some cynical asshole who hates their career

and if you have even the faintest notion of how this technology works, you KNOW all this. if you've even used autocomplete on your fucking phone you know this! it's not some secret niche behavior. all of the output comes from somewhere

of course sometimes i'll see some clownshoes comment like "well people have always stolen code" as if that's an adequate excuse for an entire industry rallying around paying for a machine that automates stealing code and pats you on the back and promises that's not what you're doing

this has attracted a few replies from the "imagine if we never cared about attribution and just had a Nexus of Ideas, man" faction, which i'm blocking on sight because i already know their ideas and don't need to hear from them as named individuals, never speak to me again thanks

kind of infantile opinion you can only develop if you've never made anything, thought of anything, or worked with anyone

"what if an endless parade of other, faceless people worked for me for free, and i could benefit off of it indefinitely without ever giving them a second thought?" wow so insightful, what a thinker you are

i’m in a discord for decompilation of GC/Wii games and it’s been increasingly ensloppened i saw someone talking about getting AI to decompile the compiler so they can learn how it works. i’m sure it’ll do pretty well since they’ve all ingested the decomp i painstakingly did and published years ago!

I saw an AI bro recently talking up how neat it was that his AI was able to decode this level format for this one game. Interestingly, the AI labeled the sections of the file using the names I made up a decade prior.

I even wrote a whole Blender add-on for it?? What really gets me is how utterly uninterested these people are in engaging with primary sources or collaborating with subject matter experts. They just want to be amazed at their oracle of truth.