I have not confirmed this. Biden signed two executive orders in regard to cyber warfare and AI, AI when used in cyberwarfare can aparantly wreak havoc in a major way.. It would seem they are expecting cyber warfare on a scale not seen before that would result in infrastructure failures and supply chain failures that would not have easy fixes. Most cyber warfare is done for blackmail so it has fixes when the victim pays. AI cyber warfare would be designed so no fixes could be done even by the hostile party,
- Thread starter AK264
- Start date
So a piece of paper signed by a dementia patient will stop it?
No. Thats the point. power can get locxked up. food delivery can get locked up. gaas pumps can get locked up. If AI does it the entire system ay not be salvageable. Acording to the vid AI can write in code of its own creation that cant be troubleshot.
You would think that someone would have thought of this and warned us years ago
Kyle Reese tried to warn us about skynet.
*And* pedophile! Never forget that "speshul" qualification!
I find it ironic that the fix for this digital world....is analog tech. Critical systems that aren't connected to the world, can't be attacked by AI.....unless someone were to break/sneak in and do something, but that's not what we're talking about here.
I wouldn't be surprised if attacking nations wouldn't start fucking with our power grid just because. I'm almost certain they are likely already mapping routes of attack. Figuring out a way in, probing our systems if you will. If a foreign nation is going to do something, they are going to do something very, very big that's going to really hurt us. I'm sure some dipship in the upper level of government saw terminator, ignored all the premise for the movie and thought "Skynet....that's what we need!".
Branden
I wouldn't be surprised if attacking nations wouldn't start fucking with our power grid just because. I'm almost certain they are likely already mapping routes of attack. Figuring out a way in, probing our systems if you will. If a foreign nation is going to do something, they are going to do something very, very big that's going to really hurt us. I'm sure some dipship in the upper level of government saw terminator, ignored all the premise for the movie and thought "Skynet....that's what we need!".
Branden
Dumb fucks and ai. These retards can’t figure out it’s going to be the end of mankind. Why they can’t leave it alone I don’t know. But the genie is out and its not going back in.
Not sure on the cyberwarfare but the White House is working on creating guidelines for the "responsible use of AI" by the commercial side of the house. Expectation within the industry is that it will just be a way to ensure the control of AI by the big connected players. Government lost control of social media and isn't interested in having AI controlled or used improperly by the "wrong" people.
Also, Military and Gov AI won't be subject to those same rules so if something breaks loose and we get Skynet it will be the Government that did it to the rest of us.
Also, Military and Gov AI won't be subject to those same rules so if something breaks loose and we get Skynet it will be the Government that did it to the rest of us.
Battlestar Galaticia, what?
Which is exactly why the last POTUS had an EO to get certain countries, like those that gave Billions to Pedopeter & Family, OUT of our power grid. Pretty early in office Pedopeter reversed it & let Chy-na back in.
I guess one of these days I'll have to fire up my 'nix box again and the port sniffing tools.....and become a l337 haxor huh ?
With the amount of holes in commonly used software these days it should be easy.
Maybe I can start with Billary's email logs ?
With the amount of holes in commonly used software these days it should be easy.
Maybe I can start with Billary's email logs ?
It won't even be necessary to seek out routes of attack. We've purchased, and continue to to purchase, critical grid transformers from the Chinese that we've already found contain malware for shutting down our systems. We did the work for our enemy.
Chinese transformers in critical electric sector confirmed by two US administrations - Industrial Cyber
Chinese transformers in the critical electric sector confirmed by two US administrations, heightening security concerns in the electric grid.
industrialcyber.co Last edited:
It won't even be necessary to seek out routes of attack. We've purchased, and continue to to purchase, critical grid transformers from the Chinese that we've already found contain mallard for shutting down our systems. We did the work for our enemy.
Chinese transformers in critical electric sector confirmed by two US administrations - Industrial Cyber
Chinese transformers in the critical electric sector confirmed by two US administrations, heightening security concerns in the electric grid.
Ducks? Ducks in our infrastructure? Where’s my shotgun?
Malware OTOH, is a lot harder to take out….
On a seriois note:
We all know its coming, their excuse for martial law and seizre of weapons and what ever else they deem important. Glad I live in Oklahoma and have a months supply of everything. We have tons on natural gas but cooling from 111 could be a bitch.
Don't bet the farm on that...
I work on oil/gas infrastructure. The amount of digital control is staggering. If a line is disabled by cyber warfare we're going to have problems.
Cushing, OK is pretty much the center of the world as far as petro product movement. Booster stations are typically dozens of miles apart... And usually in out-of-the-way remote areas. Manually overriding controls to open/close valves is possible. But restoring flow could take a while.
As to the OP and topic... Sounds like a gambit to nationalize our oil/gas industry... Turn the border into a sieve so anyone/everyone with beef can enter the country and drum up hostilities with foreign antagonists who have the means to do real harm to critical infrastructure without firing a shot. Government swoops in and takes over to "save the day."
Create the problem. Implement a solution that is advantageous to coalescing power and control.
This has the appearance of incompetent bumbling... But it's not.
This is the perfect time of year to start screwing with oil/gas infrastructure too... Cold, hungry people will be easy to manipulate.
Mike
AI is the new global warming/climate change boogeyman BS. It's just another way for the corrupt to funnell taxpayer money into their bank accounts under the guise of your "safety and security" It's just the new "Carbon Tax"
Ok so we're fucked too. At least i've got a couple tanks of propane....
And a couple months supply of food and water. No guns or ammo though, all lost in a boating mishap.
I work in the cybersecurity industry and after my EOD days in the mil did coolguy signals & EW stuff, still do it as a hobby, but more work on the data science, data engineering & cyber security side. More private sector nowadays but I think it "qualifies" me a bit to talk to this stupidity.
Firstly, the Executive Order that diaperboy signed is about "Ethics in AI" this was the outcome of that meeting the White House had with NVIDIA, Meta (Facebook), Alphabet (Google), Microsoft, OpenAI, and a bunch of other nerds. Elon wasn't invited lol. So it's a feel good thing about equity, and access, but also laying out certain stipulations. Not too dissimilar to having cryptography, direct energy, EW (well, EP/ESM), and other similar non-kinetic stuff on ITAR. Except this comes with big helping of socialist feelgood talking points. Notably, their major complaint is the training data that various AI (and subsequent ML & DNN products) are trained on for whatever they purpose is is biased or racist and what not.
There is some truth to it, but none of these systems are perfect, and none of them do the same thing. Natural Language Processing (NLP) is a machine learning family of algorithms to help machines understand speech, translation, cross-map dialects, recognize writing, and are used for everything from "Translate This Page" on Google or LinkedIn to helping retarded ass doctors annotate patient files, recognize handwriting, and what not. Another part of NLP is the conversational element for these "chatbots" which have been around for several decades in various forms, that is where we also enter into Large Language Models (LLMs) and other broad-domain corpus that these NLP systems can be trained on.
When it comes to "cyberwarfare" or "Cyber Electromagnetic Activities (CEMA)" within the broader Electromagnetic Spectrum Operations (ESO, once called Information Operations [IO]) and the application of AI, well it's not new either. There have been several AI- and ML-backed cybersecurity tools that do everything from anomaly detection, automating recon and discovery of digital targets, craft payloads, obfuscate payloads, and more. The Chatbots and LLMs, by virtue of:
1) Being easy to use - thank you NLP!
2) Being trained on a ton of "interesting" data like cybersecurity!
It's easy to use them to craft your own hacking tools all along the "cyber kill chain" - be it for "hey ChatGPT write me a tool like masscan that finds Industrial Control Systems" or "hey ChatGPT write me an encryption program in Python or Golang that uses XSalsa20 and is obfuscated that can run on MacOS Ventura" -- so going back to the ethics piece, it's also the Government, who has no real clue what they are doing making sure plebs cannot weaponize these things.
That said, they already have. There are plenty of other LLMs besides the ChatGPT-* series of them, Amazon has Titan, Meta has LLaMa, and there are dozens more. Being that the training and gathering of data and the compute power is what makes these things hard to operationalize, this tech has already proliferated, you can find versions and wrappers all over that can get around the protections that the AI companies are forced to rollout to stop fucking morons from asking on how to synthesize high explosives or SuperAIDS.
That also leads to the other flaws they want to get ahead of, the so-called "hallucinations" due to inaccurate training data, or maliciously inaccurate training data that is also called "model poisoning". Because the fact is the Gov & Mil are big consumers of this stuff, check out the DIU's solicitations or anything coming from the IC asking about AI-driven systems. DARPA has their own AI models, notably in the C-UAS space around taking over or meaconing for DJI drones they use in Ukraine. All that to say, now that adversaries know that these companies are scraping for huge bits of data, it's not too hard to setup a "trap" and get the model jacked up. It takes A LOT OF BAD DATA to poison a LLM, but it can happen, at least academically.
So, that brings me back to the whole "muh AI cyberwarfare". It's not new, it can be done, but it's not some Skynet fully automated thing, yet. AutoGPT and some botnets with LangChain and internet plugins for ChatGPT have been shown to be limited autonomous but cyber isnt as simple as "lol chatgpt plz hack my gf facebook" -- there is still a lot of ISR-T that goes into it. Cannot exactly hack a system if it's not internet facing or you cannot get a human in the loop to do something stupid like click a link.
And to top this all off - IT DONT APPLY TO THE GOV. They can use it to build SuperAIDS and SuperC4 and SuperMetasploit...not you, dirty fucking pleb, and also AI is racist reeeeee wahhhhh.
We are in dangerous times though. Now the normies know about AI, all sorts of stupid shit will continue to happen.
Firstly, the Executive Order that diaperboy signed is about "Ethics in AI" this was the outcome of that meeting the White House had with NVIDIA, Meta (Facebook), Alphabet (Google), Microsoft, OpenAI, and a bunch of other nerds. Elon wasn't invited lol. So it's a feel good thing about equity, and access, but also laying out certain stipulations. Not too dissimilar to having cryptography, direct energy, EW (well, EP/ESM), and other similar non-kinetic stuff on ITAR. Except this comes with big helping of socialist feelgood talking points. Notably, their major complaint is the training data that various AI (and subsequent ML & DNN products) are trained on for whatever they purpose is is biased or racist and what not.
There is some truth to it, but none of these systems are perfect, and none of them do the same thing. Natural Language Processing (NLP) is a machine learning family of algorithms to help machines understand speech, translation, cross-map dialects, recognize writing, and are used for everything from "Translate This Page" on Google or LinkedIn to helping retarded ass doctors annotate patient files, recognize handwriting, and what not. Another part of NLP is the conversational element for these "chatbots" which have been around for several decades in various forms, that is where we also enter into Large Language Models (LLMs) and other broad-domain corpus that these NLP systems can be trained on.
When it comes to "cyberwarfare" or "Cyber Electromagnetic Activities (CEMA)" within the broader Electromagnetic Spectrum Operations (ESO, once called Information Operations [IO]) and the application of AI, well it's not new either. There have been several AI- and ML-backed cybersecurity tools that do everything from anomaly detection, automating recon and discovery of digital targets, craft payloads, obfuscate payloads, and more. The Chatbots and LLMs, by virtue of:
1) Being easy to use - thank you NLP!
2) Being trained on a ton of "interesting" data like cybersecurity!
It's easy to use them to craft your own hacking tools all along the "cyber kill chain" - be it for "hey ChatGPT write me a tool like masscan that finds Industrial Control Systems" or "hey ChatGPT write me an encryption program in Python or Golang that uses XSalsa20 and is obfuscated that can run on MacOS Ventura" -- so going back to the ethics piece, it's also the Government, who has no real clue what they are doing making sure plebs cannot weaponize these things.
That said, they already have. There are plenty of other LLMs besides the ChatGPT-* series of them, Amazon has Titan, Meta has LLaMa, and there are dozens more. Being that the training and gathering of data and the compute power is what makes these things hard to operationalize, this tech has already proliferated, you can find versions and wrappers all over that can get around the protections that the AI companies are forced to rollout to stop fucking morons from asking on how to synthesize high explosives or SuperAIDS.
That also leads to the other flaws they want to get ahead of, the so-called "hallucinations" due to inaccurate training data, or maliciously inaccurate training data that is also called "model poisoning". Because the fact is the Gov & Mil are big consumers of this stuff, check out the DIU's solicitations or anything coming from the IC asking about AI-driven systems. DARPA has their own AI models, notably in the C-UAS space around taking over or meaconing for DJI drones they use in Ukraine. All that to say, now that adversaries know that these companies are scraping for huge bits of data, it's not too hard to setup a "trap" and get the model jacked up. It takes A LOT OF BAD DATA to poison a LLM, but it can happen, at least academically.
So, that brings me back to the whole "muh AI cyberwarfare". It's not new, it can be done, but it's not some Skynet fully automated thing, yet. AutoGPT and some botnets with LangChain and internet plugins for ChatGPT have been shown to be limited autonomous but cyber isnt as simple as "lol chatgpt plz hack my gf facebook" -- there is still a lot of ISR-T that goes into it. Cannot exactly hack a system if it's not internet facing or you cannot get a human in the loop to do something stupid like click a link.
And to top this all off - IT DONT APPLY TO THE GOV. They can use it to build SuperAIDS and SuperC4 and SuperMetasploit...not you, dirty fucking pleb, and also AI is racist reeeeee wahhhhh.
We are in dangerous times though. Now the normies know about AI, all sorts of stupid shit will continue to happen.
Ok so we're fucked too. At least i've got a couple tanks of propane....
You've done more than most.
Shit will come unglued quick if access to fuel for transportation falls apart.
Mike
Isn't great that our government has fucked up how it's supposed to work. Giving power to the president to make rules instead of passing legislation?
Totally unacceptable for what is happening and using it to blame someone else when running for another election
It worked so well in combatting poverty and drugs and terrorism, so why not run it back again?
Is this still available?This is the future........
Yes, and it is coming at no cost to you.....
I'll take two
Can I get it with the phased plasma rifle in the 40 watt range?This is the future........
Mike
How did trusting AI work out for Dave Bowman?
Similar threads
