12 comments

[ 2.9 ms ] story [ 45.8 ms ] thread
This article completely ignores how technology becomes decentralized and open-sourced. There's already a massive community dedicated to creating and sharing LLMs, and that's not going to stop if OpenAI is forced to shut down ChatGPT because of copyright laws. FWIW, shutting down Napster didn't stop music torrenting at all. What really stopped widespread torrenting was easy to use streaming services, which the music industry was forced to adapt to.
I think the past twenty or so years of technology development has mostly pointed to relentless consolidation and re-centralization.
They did mention that if you read the entire article. The argument is that free sharing of music started by Napster didn’t end up being the predominate model because copyright law resulted in steaming services starting with Apple dominating. So we should be leery of claims about what LLMs will do without considering what the legal ramifications of various lawsuits will lead to.
Foreign Policy is a trash site, since its inception.

I'm no AI stan, but in no manner is AI similar to file sharing. The article amounts to legacy media whining about their moat.

It seems pretty ignorant to expect rights holders to lay down and die while your use their data to make massively profitable systems.
I think it's pretty rational given capitalism's explicit focus on short-term profit at the expense of negative externalities. Workers are replaceable. The US outsourced its manufacturing sector. Now the same people who oversaw that process would like to outsource the service sector.

I'm not a fan of the status quo, but it's clearly functioning as designed. The cruelty of this mostly affecting individual artists and their ability to generate a living rather than behemoths like Disney is absolutely the point.

"Ignorant" is strong language given that for-profit media companies, who are complaining, are rarely citing let alone paying their sources.

The takeaway not being that they should start, but that their complaint is likely on shaky ground.

it's absolutely similar to file sharing, a bunch of free riders who rationalize that an extra copy of spider-man never hurt anybody
By that logic when you regurgitate a summary of your understanding of anything, you are potentially violating the copyright of your sources.

Or when a writer plys their trade, the same. Now what if they charge for the article?

Much of what Ai is ingesting is on the web for the purpose of ingestion.

>Much of what Ai is ingesting is on the web for the purpose of ingestion.

ingestion/consumption by humans, not automated ingestion for the purpose of reselling. the economics would not work longterm.

Humans ingest said inforrmation, and then write further articles based on it to resell.

The economics may not work, but my point is that media companies are going to have a lot of difficulty in parsing exactly who and what is able to utilize published information; and how they / it are permitted to utilize it. In the end, their economics won't be factored.