LLMs like to repeat themselves, which isn’t great for password creation.
You must log in or # to comment.
Lol I have so many questions:
-
Why would you use AI to generate passwords?
-
Why would you consume so much resources for a simple algorithm which has existed for decades?
-
Who is doing this?
in other news:
a boat motor is a bad way to make a milkshake
-
Omg this reminds me of a networking project I was working on recently where I found out my vibe coding friend had pasted his ssh keys into an LLM.
Tell the world you have no idea how passwords work without telling the world you have no fucking clue how passwords work, OP.
A word recycler would churn out used passwords? If only foresight could have foresoot this.
Edit:
I guess its more like: “The probabilistic outcome generator generates passwords that are probable?”You also had it right in the first place. Very likely that the training data sets could have included random files and password leaks. I don’t think they’re discriminating at this point.
It would explain why the same strings appear over and over in generated results. Its just one of the tokens most associated with “Password”
Fundamentally that’s all these systems are actually doing is rearranging the words they’re trained on. They’re not really fundamentally capable of coming up with anything on their own just mixing up words that are strongly correlated with the input.
Any password manager worth using will generate secure passwords. There is absolutely no reason to use AI for that.
Sometimes there’s something wrong with the way a website does login and the password manager options won’t trigger. In this case AI can be useful for telling you to install pwgen.
But how do you vibe code a password manager without having itself use AI? That would require actual programming knowledge.
Lmfao we’ve gone from the entropy wall to stochastically-generated passphrases. What a time to be alive.
It’s a little off topic, but this gives me a good platform to ask a question I’ve been wondering for a while. Also, I know lemmy is full of folks who follow the tech and security trends.
Anyway, my question is this: is this thinking when it comes to passwords still at all relevant?:
Kind of. The best advice now is to use a password manager so you dont need to remember passwords at all - in which case just make it random with as many characters as allowed and youre done.
But for remembering passwords, the longer the better and replacing some letters with numbers and random characters is helpful too.
That really something that doesn’t need to be solved by ai because it already works better without and in most cases without problems.
What kind of absolutely insanely mad and psychotic mind an individual should have to generate passwords with AI?
Idiots who don’t know how cryptography and gen AI works.
a-rappa-lappa-doo
-Pumpkinseeds and Peanuts
Just use Korn noises, duh.
eeta po ranta see nata po ra na na
Imcomingundone123
A suggestion about randomly generated passwords, pick the dictionary files from Firefox for random languages (that you can write without special keyboards/configs, if I may add), put the words included all in a LibreOffice/Excel sheet and set 4 formulas to pick words from the list at random.
From my experience, they’re far more randomic, as well as not uncommonly including special characters, solving part of the pestering for alleged password strength.
XKCD made a comic about that
Yup. That’s what inspired me. e.e
I just checked a while back, but forget to answer this. That comic says that the best we can do is use random words, because that’s easy to humans to remember. I tried that method and is not the most secure accord this site https://passwordentropy.com/
Randomic?
ESL strikes again =D
I forget “random” doesn’t have many inflections.
