chakalakasp

Because encryption is not an inherently dangerous thing. A superintelligent AI is.

It’s no different than inviting an advanced alien species to visit. Will it go well? Sure hope so, because if they don’t want it to go well it won’t be our planet any more

You have no clue what communities I’m in other than it self affirms your worldview to assume that they must be irrelevant. To be fair, I’m not telling you anything about myself, so you’re not really needing to take my word for it. And I don’t care enough about you to explain in detail.

Though with a tiny bit of Googling you’ll be able to find several Turing Award winners who are saying exactly what I’m saying. In public. Loudly.

The difference is the scale at which a hacker can cause damage. A hacker can ruin a lot of stuff but is unlikely to kill a billion or two people if he succeeds at a hack.

With superintelligent AI you likely have to have every developed use and every end user get it right, air tight, every time, forever.

In what other sphere of tech does that apply?

"Hackers aren't a problem because we have cybersecurity engineers". And yet somehow entire enterprises and governments are occasionally taken down.

What prevents issues in redteam/blueteam is having teams invested in the survivability of the people their organization is working for. That breaks down a bit when all it takes is one biomedical researcher whose wife just left him to have an AI help him craft a society ending infectious agent. Force multipliers are somewhat tempered when put in the hands of governments but not so much with individuals. Which is why people in some countries are allowed to have firearms and in some countries are not, but in no countries are individuals allowed to legally possess or manufacture WMDs. Because if everyone can have equal and easy access to WMDs, advanced civilization ends.

If you talk to the people on the bleeding edge of AI research instead of nuke-heads (I tend to be kinda deep into both communities), you'll get a better picture that, yeah, a lot of people who work on AI really do think that AI is like nukes in the scale of force multiplication it will be capable of in the near to medium future, and may well vastly exceed nukes in this regard. Even in the "good" scenarios you're looking at a future where people with relatively small resources will have access to information that would create disruptive offensive capabilities, be it biological or technological or whatever. In worse scenarios, people aren't even in the picture any more, the AIs are just working with or fighting each other and we are in the way.

Yes and no. AI is no different than proliferating nuclear weapons or deciding to burn all the fossil fuels -- on an individual competitive level, it makes sense to do this more and more to remain competitive. On a systems-whole level it leads to catastrophe. The whole tragedy of the commons thing, or, more recently described, the Moloch problem.

https://www.lesswrong.com/posts/TxcRbCYHaeL59aY7E/meditation...

Yann LeCun is one of the loudest AI open source proponents right now (which of course jives with Meta's very deliberate open source stab at OpenAI). And when you listen to smart guys like him talk, you realize that even he doesn't really grasp the problem (or if he does, he pretends not to).

https://x.com/ylecun/status/1718764953534939162?s=20

I dunno. The American Geophysical Union seemed to think it was a BFD in one of their conferences back in the 2000s. Had a whole presentation about how certain parts of the US grid would lose large transformers that would likely take months to years to replace as there are few spares and they have to be wound with miles and miles of copper by hand. Said that large swaths of the US would be without grid power for very long periods.

Said that third world countries would fare pretty well though.

This event is twice as powerful as the previously known largest event, which is at least 10X as powerful as the largest event in recorded human history, back in the 1800s, which set telegraph poles on fire. So, yeah, it's a big Twinkee.

Such as the humans

This is a succinct explanation of the problem. Do we give the vast majority of users extremely easy, frictionless access to very high levels of security and privacy? Or do we give the vast majority of users a fundamentally insecure solution that with lots of learning and configuring and time can be have very very very high levels of security and privacy?

The crazy thing is that apple hardware beats most other hardware, too, at a high price. Better phones, better tablets, better laptops. More secure, more private OS than the popular consumer alternatives (Windows, Android). Arguably much better OS all around, too (at least IMO -- iOS beats even stock Pixel Android at use-ability, MacOS v Windows is like the Harlem Globetrotters playing the Washington Generals.)

Thing is, if law enforcement is patient they can get the data off the actual devices themselves, if they're still alive. Yes, a fully patched iPhone tends to be a fortress of might to anyone other than a nation state willing to burn a few very expensive 0 days, but with almost any phone if you wait a year or two something will inevitably come out that will allow the ol' Cellebrite crowbar a cranny to slip into.

My favorite part of android is how security patches go through a multi-tiered trickle-down system of testing to make sure they work with the dozens of custom flavors each manufacturer has so that by the time you get patched it's been in the wild for weeks or months. Oooh, ooh, no that's not my favorite thing, my favorite thing is how each cellular company gets to put their own bloatware on top of the bloatware that each phone manufacturer gets to add to it. Oh wait, maybe it's patch support ending for new phones 3 years after they were released. There is so much to love about how Android turned out it's hard to pick just one thing.

Except, of course, the human has very little control over what the AI outputs in TXT2TXT scenarios, at least in terms of whether the output would match the definition of copyright infringement of someone else's work. IMG2TXT is kinda different -- I think you could make a much stronger case for derivative work there. So you have a tool that can randomly create massive liability for you, and you can't know if its done so or not until someone sues you.

Best half frame camera I've ever seen was designed by the same guy mentioned in the article: the Olympus Pen F https://en.wikipedia.org/wiki/Olympus_Pen_F

Thing is, if you ping experts in the privacy field (like Mike Bazzell, the former FBI OSINT guy who billionaires and celebs hire to keep their personal info off the internet), they will all say the VPNs are a very important tool to create a layer of privacy between you and the site you are visiting and also a good tool to prevent said sites from easily profiling you. No, they are not a panacea -- much like the security universe, the privacy universe requires lots of other active and passive behavioral and technological changes to properly lock things down to whatever standard you require for your threat model. But they are very much an important tool. For a small subset of people it's literally a tool that protects their life.

So yeah, it's a kinda a big deal if it leaks. (Which is why most privacy experts, were you to tell them you were sufficiently paranoid, would have you fire up a pfSense and link it permanently to a VPN service and then run a separate brand of VPN software on whatever device, so that you have two layers going through two companies.)

Council Bluffs datacenter electrical explosion: https://www.ketv.com/article/council-bluffs-google-data-cent...

I think it was around 20 days for the Hubble and around 12 hours for the JWT

Sure, here are a couple of resources, both for Russia and the U.S.

For the U.S. plan to attack Russia, the NRDC did an interesting hypothetical attack around 20 years ago to show the impact a U.S. attack would have on Russia. (It's hypothetical because the actual attack plans are classified to the point that even Senators on the big committees don't get to see them when they angrily ask for them.) The report on that hypothetical is here: https://www.nrdc.org/sites/default/files/us-nuclear-war-plan...

For the Russian plan to attack the U.S., this reddit thread is a guy who did this for a living answering questions:

https://www.reddit.com/r/war/comments/trkfe7/three_hypotheti...

The submitter, https://www.reddit.com/user/dmteter/, has deep knowledge of the issue, and answers a lot of questions. He's kinda a hothead, though, if asked questions the wrong way. I chalk it up to the kind of stress that might come from realizing that the plans you helped to develop to kill hundreds of millions of people might actually get put to use in the foreseeable future.

He also answers some questions in this thread: https://www.reddit.com/r/nuclearwar/comments/tpa3ao/openriso...

The nuclear winter stuff is more interesting. Most recent papers are very much thinking it's a very big deal:

https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/201... https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3155983 http://climate.envsci.rutgers.edu/pdf/ChinaNWrice8inpress.pd... https://climate.envsci.rutgers.edu/pdf/OzdoganNuclearWinterM... https://agupubs.onlinelibrary.wiley.com/doi/10.1002/2013EF00... https://www.science.org/doi/10.1126/sciadv.aay5478 https://www.pnas.org/doi/full/10.1073/pnas.1919049117 https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019GL08... https://agupubs.onlinelibrary.wiley.com/doi/10.1002/2017JD02... https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019JD03...

One paper that feels the nuclear winter risk is overstated is from LLNL: https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2020...

This paper talks a bit about the uncertainties regarding nuclear winter right now. Mainly lack of data for modelling: https://www.tandfonline.com/doi/full/10.1080/25751654.2021.1...

A lot of the questions revolve around whether nuclear detonations would create firestorms in contemporary U.S. cities, and if they did, if those fire storms would be strong enough to generate pyrocumulonimbus clouds that drew the soot into the stratosphere on a widespread scale. If firestorms do start, based on PNW fire studies, it seems that pyrocumulonimbus clouds would be a very common thing: https://www.science.org/doi/10.1126/science.aax1748

Ok. That gets you to stay alive for a few months.

Now, imagine a world in which there will no longer be any agriculture, fuel, easy access to clean water, electricity, heating, medicine, health care, communications, government, supply chains, or law enforcement. The food supply that was available at stores and in homes has now run out. Fertilizer is no longer available, contemporary farm machinery that runs on petrol won't last long as the petrol dries up (and when they break down, won't be repairable due to DRM). Wildlife might supply enough calories to feed a small fraction of one percent of all people currently still alive.

This would go on indefinitely.

Preppers look at surviving a nuclear war like it's an adventure where all their Eagle Scout skills will finally be put to use. It's really just slow, heartbreaking death from starvation and disease, with horrible violence from desperate people and regional warlords mixed in. Think "The Road", without the extreme environmental destruction.

Note that this wouldn't be everywhere, just in targeted countries. US, Canada, NATO and NATO leaning Europe, Japan, both Koreas, China, Russia. Those not targeted would be looking at this whole thing kinda like insert "it's free real estate" meme here

And if nuclear winter is a real thing, which most contemporary modelling suggests it is (though this is not settled science and some modelling suggests it isn't), likely there won't be an environment conducive to growing crops anywhere in the world for at least 5 years. A lifetime Prepper Magazine subscription won't help you there.

The people who make the claims that ABM is super hard and ineffective against large strikes are generally American experts who used to work in the field.

The author is wrong. The people who are experts at doing this kind of thing for a living very much believe that a general nuclear war will destroy nearly all infrastructure.

Don't just believe random bloggers with no expertise in the things they're bullshitting about. Even in the open source world, there is an entire body of expert information on this topic that contradicts what this guy is saying.

David has mentioned on Twitter and Reddit (when asked) that he estimates 90% to 95% of all people living in the U.S. would be dead within 18 months after either of his attack scenarios that involve countervalue due to complete breakdown of all infrastructure, production, supply chains, communications, etc.

ABM is hard. Evading ABM is much easier. Russia has some amazing tech for space based ABM interception evasion.

ABM stuff works okayish against very small numbers of incoming warheads. ABM is completely ineffective against large numbers of inbound warheads.

Ha! Good point. Good time to take a lifecation to Africa.

Yes.

>Weapon reliability -- particularly on the Russian side. Will many / most of their ICBMs even get off the ground? Will the warheads actually work properly? I'm not saying they're all duds but if I was Russia I'd maybe keep 100 maintained for the sake of demonstration.

From the little I've communicated with people who would know (because they used to be in the IC), the vague answer is "Yes, they would work."

>European Theatre -- Russia needs its subs to try to stop UK / American subs from nuking them so they're not likely deployed around NA (instead hunting for enemies closer to home) and thus can't destroy the silos.

The submarines that launch ballistic nuclear missiles are not the same submarines that hunt the submarines that launch ballistic nuclear missiles. You might want to read up on the wiki on this stuff or something.

The author has no idea what he is talking about and his thoughts have no relation to actual nuclear targeting plans.

There are multiple plans for multiple contingencies. A couple of different examples are here:

https://github.com/davidteter/OPEN-RISOP

The fallout thing is kinda a unlikely thing. Fallout from typical ground-bursts is something that denies area for weeks at most.

There are weapons systems that can do area denial for much longer by using weapons salted with materials that will remain radioactive for many years (essentially the plot of Dr. Strangelove -- that "doomsday" system is based on realistic weapons possibilities, and the Russians in the past decade intimated that they have developed a submersible drone system designed to cross the ocean and detonate huge salted weapons to poison the U.S. coasts for decades https://en.wikipedia.org/wiki/Status-6_Oceanic_Multipurpose_... ) -- but as far as anyone in the open source info world knows, these kinds of weapons are not deployed.