Post

p sure folk here will hate this, but i think that open models above a certain strength (Mythos level??) should indeed be restricted access below that threshold i’m all in on open models, but above leads to some serious social risks idk maybe i’ll think different after significant diffusion

the thing i want is controls — the ability for access to be revoked, content guards, etc. open models inherently don’t have these things. like if they had them they wouldn’t be open

Say more about the risks? Cybersecurity? Something else? An unknown unknown? If it's cybersecurity, is it sufficient for institutions to have a month advance warning of new capabilities? Fwiw, I strongly suspect that what's happening rn is US gov protecting its own *offensive* edge.

i’m being vague on purpose it’s the ability to pull back a model that’s important. the actual crime will probably evolve a heck of a lot over the next many months, but you can’t even begin to address any of them without basic controls

Hmm. The idea of controlling information to protect against an unspecified future threat is indeed going to run against the grain of many people's instincts. The old Oliver Wendell Holmes standard for restricting speech was "clear and present danger."

but LLMs are not “information”. they merely act on information

0 reposts2 likes

Uhm Yes, but reading datasheets for a mcu obsolete in 2000 is extremely hard Writing sleigh is even harder Reverse complete firmware even more But yes, the information is out there Not sure of your point at this time😉🤗

My undergraduate philosophy degree, normally pretty calm and well-behaved, just started bouncing back and forth inside my skull like a caged chimpanzee that sees a banana across the room :)

ok i mean they’re information in regards to physics but in regards to the statement, any information they were trained on was accessible somewhere else, so the access wasn’t revoked

no one else "liked" this comment but afaict your view is just the correct one, viruses are also in a sense "just information"† but that's not really germane to anything, it's "is this a natural kind that can be carved out by policy under a cost/benefit analysis" (Yes)

19th century intuitive liberal mores about censoring criticism of parliament in newspapers are not really "in sample" in this situation †premoving the objection that viruses aren't just information because they need to be physically instantiated: software needs to be physically run on computers

Two possible approaches here: 1) Agree with your premise, substitute "thought" for "information," then note that this edit is not v reassuring. 2) Note that this would also permit censoring encyclopedias, since they are just summaries of information in principle available elsewhere.