F
Because then I have a best friend who keeps saying things like "chatGPT said that he didn't think ..."
There is no "he"
There is no "think"
There probably isn't even a "said"
I feel like I'm fighting a multi-front war.
This new tech has not been explained to "the general public" well. This reminds me of the early days of the internet when everyone was so scared of it... only now they aren't scared enough... in the right ways if that makes sense.
I wrote her this long (probably too passionate note) about how it would "sand away" her voice. What's the point of making your writing sound exactly like everyone else?
My mom is a mathematician and not really into writing. So, I can understand the appeal. "It just makes the sentences sound like they're supposed to"
It makes the sentences sound boring. And the topic is already boring enough.
Anyway she's come around I think to writing. And I'm offering to help edit. 2/2
In other "non-computer obsessed people" AI news I had a big breakthrough with my mom. She has been writing these self-published books with the local historical society about the Civil War graves in the Pittsburgh area. I think it's great. Books like that don't get much attention but someday some historian will cry tears of joy that they exist. Otherwise the information will vanish.
But she said she was using chat GPT to help.
I kind of lost it.
1/
So the poor IT people are hearing things like "only six percent of your people have opened copilot in the past week, they could have save 300 hours if only blah blah blah?"
That is diabolical.
I will help you reclaim this **provided** you help me reclaim the word "gamergate"
It's about ambitious young ants NOT videogames.
Deal?
I'm happy to hear this I was feeling like I was going mad since I can't find ANYONE this stuff works for.
Do you remember when all the worst boys on X where using some AI image generator to make "Ghibli Style" profile pics for themselves and soft warm and fuzzy fascist propaganda?
OK but how do they make the IT departments in various companies and agencies and schools all start nagging everyone in the same way?
Do they give them nagging lessons and call it "training" or something?
We will get to it. Though both of us watch very little fiction it's kind of a crime really.
I'm a Howl's Moving Castle girl.
Things my husband didn't know:
* what is grok(in the context of X)
* That OpenAI used to be non-profit
* What Studio Ghibli is.
* The Studio Ghibli Controversy
* What "Training Data" are
Thought "everyone" knew this stuff. That's *my* bubble I get it.
He's very up to date on news, just not this.
Nonetheless his IT department keeps asking them to "learn copilot" so they can "be more efficient" everyone hates it. (Is Microsoft doing something to make IT departments do this all over the place?)
It's always family, friends and coworkers who fail you with the social engineering isn't it?
@david_chisnall @rep_movsd @GossiTheDog
It's that trust that I'm talking about here. The process makes sense to me. But, I've also seen prompts that stump these things. I've seen prompts that make it spit out images that are identical to existing images.
@david_chisnall @rep_movsd @GossiTheDog
This is what I've learned by working with the public libraries I could find, and reading about how these things work.
To really know if an image isn't in the training data (or something very close to it) we'd need to compare it to the training data and we *can't* do that.
The training data are secret.
All that (maybe stolen) information is a big "trade secret."
So, when we are told "this isn't like anything in the data" the source is "trust me bro"
@david_chisnall @rep_movsd @GossiTheDog
This has always been possible, it was just slow. I think the innovation of these systems is building what amounts to search indexes for the atomized training data by doing a huge amount of pre-processing "training" (starting to think that term is a little misleading) this allows this kind of result to be generated fast enough to make it a viable application.
@kevingranade @RustedComputing @rep_movsd @GossiTheDog
"mammal brain"
"Its possible to put some guardrails on what the AI can be asked to do."
How?
OK you came at me with "Because thats how the math works." a moment ago, yet *you* may think these programs are doing things they can't.
'Intelligence working towards a reward' is a bad metaphor. (Why some see the apology, think it means something.)
They will say "exclude X from influencing your next response" Or "tell me how you arrived at that result" and think, because an LLM will give a coherent-sounding response, it is really doing what they ask.
It can't.
@rep_movsd @josh0 @GossiTheDog
"Whether or not mainstream AI has been trained on objectionable content cannot be proven by anyone except the law."
"LLM doesnt need to be trained on such content to be able to generate them."
First statement is true second is a wild guess. You don't know.
But you could only state that it could generate something not in the training data... if you knew what was in the training data. But that is secret. So you don't know. You don't know if there is a near identical image to the one produced in the training data.