> I think USB 4 is finally going to the right direction.
USB 4 is actually going into an even worse direction. USB 4 = Thunderbolt 4, except everything is optional. e.g. USB 4 might not even support DP Alt mode. Thunderbolt 4 always will.
There were some M-series chips with 8 gigs, iirc. There was a whole debate going on about that on the net when they were released. Not the M5 though, as it seems.
And yes, absolutely. All you need is a bootchain exploit. However unlike in the old jailbreaking days when people found and publicized them for fun, these days they are worth millions. Apple will pay you $500k for sandbox escape into the kernel. If you nail the bootchain, it'll be in the millions. From Apple. And god knows how much such a thing would go for in the black market.
I'm trying to figure out what it is. Is it matte, but just a different matte from the rest of the case? I kind of want to see it in person. It looks very tasteful.
How do you determine to whom an IP is even registered to? They get sub-leased all the time.
The best you can do is check who has administrative control over the prefixes RIR info, but that doesn’t mean that anyone with control is the factual user of the IPs.
You could check the IRR for the ASN and base it on that, but still.
There's also no way to actually know _where_ an IP actually originates from. Only its AS path.
The DFZ contains all prefixes announced everywhere, for the internet is completely decentralized.
8 port-channel13.core4.dal1.he.net (184.104.196.170) 1.830 ms 1.969 ms *
9 ns1.he.net (216.218.130.2) 1.539 ms 1.560 ms 1.555 ms
Virginia (measured from Maryland):
11 port-channel2.core1.ash1.he.net (184.105.222.174) 19.666 ms 24.395 ms *
12 ns1.he.net (216.218.130.2) 16.748 ms 17.268 ms 20.507 ms
And California (measured from California):
8 port-channel13.core1.fmt2.he.net (184.104.188.144) 3.830 ms be7.core1.sjc1.he.net (72.52.92.132) 5.197 ms port-channel13.core1.fmt2.he.net (184.104.188.144) 3.901 ms
9 ns1.he.net (216.218.130.2) 2.600 ms 2.435 ms 2.728 ms
The speed of light doesn't lie, IP addresses don't have any sort of physicality.
> we don't want IP's from Frankfurt showing up somewhere in Dubai
what do you mean, IPs from Frankfurt?
IP addresses are just IP addresses, they know no geographical boundaries. In RIR DBs you can geolocate them to wherever you want. Which is the entire reason why Geo IP DBs even exist - they triangulate.
wasn't a lack of networking what made it a temple, untouched by the influences of the corrupt internet or something like that? idk I'm not like a Terry Davis scholar by any means but I always figured he did that limitation with some kind of reason in mind
> Absolutely everyone should be allowed to access AI models without any restraints/safety mitigations.
You recon?
Ok, so now every random lone wolf attacker can ask for help with designing and performing whatever attack with whatever DIY weapon system the AI is competent to help with.
Right now, what keeps us safe from serious threats is limited competence of both humans and AI, including for removing alignment from open models, plus any safeties in specifically ChatGPT models and how ChatGPT is synonymous with LLMs for 90% of the population.
Yes IMO the talk of safety and alignment has nothing at all to do with what is ethical for a computer program to produce as its output, and everything to do with what service a corporation is willing to provide. Anthropic doesn’t want the smoke from providing DoD with a model aligned to DoD reasoning.
the line of ego, where seeing less "deserving" people (say ones controlling Russian bots to push quality propaganda on big scale or scam groups using AI to call and scam people w/o personnel being the limiting factor on how many calls you can make) makes you feel like it's unfair for them to posses same technology for bad things giving them "edge" in their en-devours.
The cat is out of the bag and there’s no defense against that.
There are several open source models with no built in (or trivial to ecape) safeguards. Of course they can afford that because they are non-commercial.
Anthorpic can’t afford a headline like “Claude helped a terrorist build a bomb”.
And this whataboutism is completely meaningless. See: P. A. Luty’s Expedient Homemade Firearms (https://en.wikipedia.org/wiki/Philip_Luty), or FGC-9 when 3D printing.
It’s trivial to build guns or bombs, and there’s a strong inverse correlation between people wanting to cause mass harm and those willing to learn how to do so.
I’m certain that _everyone_ looking for AI assistance even with your example would be learning about it for academic reasons, sheer curiosity, or would kill themselves in the process.
“What saveguards should LLMs have” is the wrong question. “When aren’t they going to have any?” is an inevitability. Perhaps not in widespread commercial products, but definitely widely-accessible ones.
> There are several open source models with no built in (or trivial to ecape) safeguards.
You are underestimating this. It's almost trivial to remove the safeguards for any open-weight model currently available. I myself (a random nobody) did it a few weeks ago on a recently released model as a weekend side-project. And the tools/techniques to do this are only getting better and easier to use!
Sounds like you're betting everyone's future on that remaing true, and not flipping.
Perhaps it won't flip. Perhaps LLMs will always be worse at this than humans. Perhaps all that code I just got was secretly outsourced to a secret cabal in India who can type faster than I can read.
I would prefer not to make the bet that universities continue to be better at solving problems than LLMs. And not just LLMs: AI have been busy finding new dangerous chemicals since before most people had heard of LLMs.
chances of them surviving the process is zero, same with explosives. If you have to ask you are most likely to kill yourself in the process or achieve something harmless.
Think of it that way. The hard part for nuclear device is enriching thr uranium. If you have it a chimp could build the bomb.
I’d argue that with explosives it’s significantly above zero.
But with bioweapons, yeah, that should be a solid zero. The ones actually doing it off an AI prompt aren't going to have access to a BSL-3 lab (or more importantly, probably know nothing about cross-contamination), and just about everyone who has access to a BSL-3 lab, should already have all the theoretical knowledge they would need for it.
USB 4 is actually going into an even worse direction. USB 4 = Thunderbolt 4, except everything is optional. e.g. USB 4 might not even support DP Alt mode. Thunderbolt 4 always will.
reply