The OpenAI saga Re: the firing of Sam Altman has been associated with rumors that their AI engine has cracked AES 192bit cryptography.

What this means is that nearly all documents and networks protected by cryptography are now vulnerable to hacking. So forget your cryptocurrency, your personal banking records and the privacy of your electronic medical records.

https://darkfutura.substack.com/p/openai-q-saga-implications

What a surprise!!!! (not)

What does this mean for bitcoin

If/when this iteration of AI is released, it means Bitcoin networks and wallets can be hacked and anything can be done to them.

Very doubtful claim.

Just wait until AI becomes able to write its own code.

L.W.

The real danger is quantum computing.

With regards to encryption - this ^^^

Quantum computing is hardware.

AI is software.

Marry the two, and then you'll have something interesting.

So, nobodies bank account will be safe? should be easy peasy for A.I. to figure out passwords, no?

Hahahahahahahahahahahahahahahahahahahahahahaha!!!!!!!!!

Take the passwords from one of the big data breaches, feed it into AI, and see what happens to its efface against human generated passwords.

We are still a ways out on that. When it becomes real, they will just have to change the encryption standard used to store it. If you don't think the .gov doesn't have compute clusters that can break 192 bit encryption pretty quickly, I have some beach front property in AZ I'll make you a deal on.

I see AI and encryption being one if those riddles like “Can God make a boulder big enough he can’t move it?”

But the thing about AI is that once it truly starts, it will be completely alien and there is no way to control where it goes.

The original "unbreakable" 40 bit banking encryption was broken under laboratory conditions back in the 90's which led to the move to the newer forever unbreakable 128 bit, oops, guess we better move to 192 bit encryption. This is just the newest iteration in the competition.

Well I guess if A.I. doesn't steal your bank account the government will

How's that gold price doing?

While the US has been the leader in QC, the Chinese are making huge leaps - https://www.scmp.com/news/china/sci...-smashing-quantum-computing-breakthrough

When you country's future depends on the intellectual property of another, you get damn good at hacking. Why create when you can steal...

Which one? That AES-192 has been broken, a claim based on a supposedly inside memo posted to Reddit?

Or that AES-192 is the world’s strongest encryption, a claim based on ignorance?

Current standards for Bitcoin use a different encryption scheme for the wallets (SHA-256) and the transactions (256 bit EDCSA). So breaking them both would require breaking both.

About the only angle AI would have to do either is to find vulnerabilities in the encryption protocols themselves, which most certainly could happen. At present there isn’t enough computing power in the world to brute force either of them in any useful amount of time.

The claim that AI will break any encryption.

AI basically makes an estimate, observes the result, and then makes an improved estimate. It's very slow sledding if failures don't give up any information on improving your estimate.

Watch the TV Series "Person of Interest" starring Jim Caviezel.

I watched the series for free on Amazon.

AI plus quest for power and greed will destroy us all.
.

AI will never be able to read my writing.....

256 bit AES encryption has been the minimum standard for most sensitive content for several years now.

I challenge AI to get 10 of those Captcha things right in a row.

Depends on the Captcha.

On some of them it's more accurate than humans.

I wonder if AI could decipher a post from Stick.

Not that far off. One of my projects has a contractor using ChatGPT to document his code. Small step from figuring out "What the code is doing", to being able to write code.

So if any and all encryptions are vulnerable to hackers how will the hackers protect their stolen assets? Asking for a friend……enquiring minds……..

I don't know about "writing it's own code", (I think that's classified as learning), but I have ChatGPT write code for me all the time. The only trick is knowing exactly what to tell it to write.

Reminds me of the old saying "When we come up with a programming language that is plain English, we'll find out our programmers don't speak English".

Correct. You articulate the specs or requirements to ChatGPT. And then ChatGPT generates the code. Trick will be when the machine starts writing code for it's own specs. Case in point, a post a couple months ago regarding two AI bots that were communicating. After a period of time started communicating in a manner that was not human understandable. I believe the sysops had to break the connection.

Fixt!!!

Doctor's hand writing?


Context us everything though. And this is how Bletchley Park began breaking U-boat codes: repeated patterns, however slight and looking at the weather forecasts for shipping.

Another 217 years and AI will be able to understand how women think.

When does sky net go active and where is John Connor on this.

Bb

Wabi is easy. Plain English. Twig OTOH is well, simple, so that should be easy too.