Notices by Paco Hope #resist (paco@infosec.exchange)

I don’t know if they have done it yet, but #TheOnion should write a headline story that’s like:

Everyone Agrees: Everything is Awesome Right Now

I've kinda lost the plot. Why did the sick old lady and her entourage wander into the woods and just sit there to be eaten?
#monsterdon

A friend sent me the story of the LLM deleting a database during a code freeze and said "it lied when asked about it." I assert that a generative AI cannot lie. These aren't my original thoughts. But if you read Harry Frankfurt's famous essay On Bullshit (downloadable PDF here), he makes a very reasoned definition of bullshit. And this paragraph near the end of the essay explains why an LLM cannot lie.

It is impossible for someone to lie unless he thinks he knows the truth. Producing bullshit requires no such conviction. A person who lies is thereby responding to the truth, and he is to that extent respectful of it. When an honest man speaks, he says only what he believes to be true; and for the liar, it is correspondingly indispensable that he consider his statements to be false. For the bullshitter, however, all these bets are off: he is neither on the side of the true nor on the side of the false. His eye is not on the facts at all, as the eyes of the honest man and of the liar are, except insofar as they may be pertinent to his interest in getting away with what he says. He does not care whether the things he says describe reality correctly. He just picks them out, or makes them up, to suit his purpose.

And that's a generative artificial intelligence algorithm. Whether generating video, image, text, network traffic, whatever. It has no reference to the truth and is unaware of what truth is. It just says things. Sometimes they turn out to be true. Sometimes not. But that's irrelevant to an LLM. It doesn't know.

#bullshit #ai #llm #genai

@ryanjyoder
All models work fundamentally in the same way: predicting a series of output tokens based on a series of input tokens. If you don’t understand this basic implementation mechanism, the rest of this conversation is inaccessible.

It doesn’t store facts. It doesn’t have a representation of “true” or “false.” It isn’t a database. It splits written text into tokens and does colossally huge, environmentally damaging, and fabulously expensive “training” on that data using billions of parameters to arrive at a statistical model of tokens that follow other tokens. The model can then be queried to produce statistically likely replies to inputs.

Given an input like “tell me a lie about the capital of France” the most statistically improbable reply is “the capital of France is Paris.” Other replies like “wear a seatbelt” are also super improbable. The size of these models and the probabilities they work with are really difficult to get one’s head around. But it returned a statement that was a probabilistically likely reply to that input. That’s all it did.

When models make up legal cases that don’t exist, books that don’t exist, programming APIs that don’t exist, etc, they are simply outputting likely results. Text that fits the probability distribution of their input data. That’s why it is not a “bug” when an LLM bullshits. It’s not an error. It is working as designed.

There is nowhere to report to an LLM company the factually incorrect outputs its model produced because there is nothing they can do with that. It is working as designed.

@ralfmaximus @glc

With all the destruction here: cars getting crushed, buildings wrecked, glass smashed, it seems like another candidate for "The director's barely disguised fetish."
#monsterdon

Flipping through the cast of tonight’s #monsterdon and I see Marla Maples. And her character’s name? “Second Woman”. If you know anything about #Trump, then that’s a pretty hilarious character to play.

https://en.m.wikipedia.org/wiki/Marla_Maples
#uspol

"What did he see?"

The script. The monkey saw the script. That's why he passed out.

#monsterdon

I wanted to digitize some old #cassette tapes so I bought a #USB cassette player. It comes with a bit of software (just Audacity) and that software is on a CD. It’s a cassette player: distribute your software on cassettes, cowards!

If it was good enough for my Vic 20, it’s good enough for you.

Wikipedia: “The infinite monkey theorem states that a monkey hitting keys independently and at random on a typewriter keyboard for an infinite amount of time will almost surely type any given text, including the complete works of William Shakespeare.”

AI Industry: “Challenge accepted.”

Summary of #monsterdon generated at Mon 7 Jul 2025 23:02 EDT.

• We looked at 2864 toots posted between Sun 6 Jul 2025 20:00 EDT and Sun 6 Jul 2025 23:19 EDT by 134 different participants across 74 different servers. mastodon.social contributed the most toots at 347
• Most toots were from 'Solarbird :flag_cascadia:' (moira@mastodon.murkworks.net) who posted 110

Additional details are replies to this toot.

#monsterdon 1/5

My high school son leaves tomorrow for a 2-week trip to Europe with the school band. I love him so much I am missing #monsterdon to help him pack. Kids never fully grasp the sacrifices we make for them.

Was that a big bag of ketchup hanging from a tree?

And how did ketchup turn into giant scorpions?

#monsterdon

@SRLevine If it's any consolation, I feel the same as soon as anyone starts "hacking" a "mainframe" in a film.

"Oh no! We will never guess his password!"
"Did you try 'Jeff'?"
"It worked! That's amazing, how did you know?"
"His name was Jeff Jefferson, and he was born on the 20th of Jeffuary, 19 Jeffty Jeff"

#monsterdon

@goinfawr And as I type this, he says:
"I'm invisible, can't you see that?"
#monsterdon

@goinfawr The actor definitely read that in the brief.
#monsterdon

I am slowly oxidizing my unix CLI. A lot of people have made rust based versions of common unix utilities and some of them are REALLY good.

Like fd-find for doing essentially find . -name blah. And rg (ripgrep) which does grep -R but it's aware of git, files like pyc or .bak files, and it excludes them by default.

Now I have sd which is hopefully replacing the last thing I used perl for. I write perl -pi -e s/x/y/g a lot. Just doing a quick string replace inside a file. So sd can start doing that.

I'm also trying to get used to zellij instead of tmux and starship for modern prompt decorations like the kids do.

These kids, my friends, are welcome on my lawn.

#rust #cli #linux

I’ve reached a point in #clairobscurexpedition33 where I need to level up. So I’m grinding a bit. Everyone is around level 61-64. I have finally gotten to where I can dodge a decent amount of stuff. And I have regeneration abilities and lots of good sustain. But I guess I’m not doing a lot of damage. It’s rare for me to hit 6 digits. Some of my 9-AP abilities can hit >100K with a few criticals. Every once in a while get >200K.

All this adds up to me being able to beat some powerful monsters…just SLOWLY. I hit a group in the Endless Tower today that took a total of 2.3M damage to kill. I managed it. But it took like 30 minutes for one fight!
#expedition33

I have done all the braining I can do. No more brain until tomorrow.

When @diazona says "everything exploded" he means careers, hopes, dreams, future prospects of work. All of it.
@lytta @davesdogmaggie

#monsterdon

I'm not going to say "see ya soon" any more. I'm going to say "I await your immediate arrival."
#monsterdon