Re: Encryption and advancement of AI and even more tools to come.

On 2026-05-29, Dale wrote:

> Howdy,
>
> I was reading a article where people "think" about things to come but
> that are coming as a general rule.  It's a matter more of when than if
> I think.  I have to say, I've read this in other places as well in the
> past.  I do think it is getting closer.  Since most of us are
> concerned about the security of our info, I'm wondering what others
> think and what the corrective measures will be, if they are known.
>
> The topic, encryption.  Whether it is file system or the encryption
> used between say a user and their bank or other secure data methods
> both for personal info as well as Govt data like for militarys and
> such.  Will AI, or future advances in computers not to far off,
> eventually make encryption really hard to keep secure and as a bonus,
> will data encrypted but stored somewhere, be crackable later with
> those advancements?  One of the things the article pointed out is that
> some Govts are storing encrypted data with the hope of decrypting it
> later, either for intel or for some sort of investigation they are
> doing.
>
> This is one of those topics that there is likely no wrong answer,
> except for it will never happen.  I'm sure opinions on this will
> vary.  Heck, mine varies.  With the coming AI build up, what effect
> will that have if any?  Will future computer systems, that use MUCH
> faster technology than what we have today, be the breaking point? Will
> it be a combination of both?

Speed may have its impact in what is crackable in a given time, and
AFAIK that has been taken into account for years now, with
recommendations increasing for e.g. key sizes.

Also AFAIK what could change the landscape is if new computing methods
("quantum computing" comes to mind) significantly change the ways in
which such cracking can happen, but wasn't some of that already being
taken into consideration?

GenAI in itself can do nothing, it's actually a quite dumb technology
for this, you could perhaps use it to guide brute-force attacks if
somebody did use predictable items, like passwords, but encryption
itself relies on mathematical properties that go a bit further than
"choose a good password".

> I'm mostly curious because this affects us all.  I access my bank
> online and pay bills online.  I also type in passwords and such as
> well that I would not want just anyone to be able to see and use.  I
> also have encrypted data stored locally here.  What do some of you who
> keep up with this topic more than most think or are hearing through
> the grapevine so to speak?  Is this getting closer real fast?  Is it
> going to be a long time yet?

Would it even be realistic to conceive LLMs producing ground-breaking
results in the P v. NP question?

> Just looking for some thoughts on this.  Again, no wrong answer
> really.  It's just a thread that some may want to read and then sort
> of add up the thoughts and ideas.
>
> Dale
>
> :-)  :-)
>
>

-- 
Nuno Silva