Comment by varispeed
10 days ago
That until it becomes illegal to have or use open source models without approval and licence from government. With more talks about on device scanning, this could be easily plumbed in. If OS detects there is open source model, it could brick your device or alert authorities. Then next step will be limiting what operating system you can install. Likely only those where you cannot remove client side scanning.
It can't easily be plumbed in, though. I can spin up my own Linux build with none of that plumbing and do what I want with it. I can grab China's best models and use them or distilled versions on my own terms because OSS allows for that. Until hardware comes fully locked down and the models cannot be run on old hardware, both a long ways off, OSS is a way out.
Linux will be an illegal operating system in 2027 because it doesn't do age verification.
Good joke. The whole US industry would collapse before even destroying one of the main pillars of AWS, Google and whatnot. It would be funny if all the billion-sized corpos suing the judges en masse for trillions of dollars in losses. Lots of them would even ask for bodyguards because for sure the bribed ones would dissapear in really "weird" ways.
1 reply →
In the not so distant future, all your coding edits (for work anyway) will be through centralized gateways. Think remote desktop environment where pasting from the client is disabled.
Many workplaces already to this.
Politicians can act very quickly if there is good enough "incentive".
I wonder how you can reliably detect an open source model though. It can be stored in any binary format, and the weights can be modified slightly so that the float values are completely different while the network works the same. The binary that runs it can be obfuscated as well. Maybe the hardware could detect common LLM inference patterns at runtime? That would probably produce many false positives.
It's been illegal forever to run a pirated copy of Windows or Photoshop. Even 30 years ago people weren't worried that their pirated copies would tattle on them, businesses did not use pirated copies because vendors would report them/not work on their systems, legal discovery could find them, etc and then they would get ridiculous fines.
It's one thing to get a copy of "illegal" software and use it yourself. The stakes are basically zero and you almost certainly will not get caught
It's a completely different thing to run a business on it with dozens of employees and requiring the employees to break the law to do their job.
Completely different to run a business on massive copyright infringement. A business like... OpenAI.
You don’t need to detect it, you just need to incentivize employees and competitors to snitch on companies using unapproved models.
You don't need a blacklist.
Maintaining a blessed whitelist is the way to go.
Antivirus companies have large expertise in this.
Not gonna happen; the incentives for bypassing this are too high.
This is getting terrifying pretty quick...
Eh probably easier ways to do this. Just sanction all entities that release open weight models for "illegal distillation". Enough to cross the risk threshold for most businesses in the west, and reduce future monetization opportunities.