“You study diligently, while I cheat and learn nothing” You’re goddamned right we’re not the same, and I hope I never slip to the level of quality this guy does. Chuds believing what this guy believes is job security for me.
Perfectly good approach if you know the subject well enough to know that the information you think you need is really what you need.
But if you were using a book in that scenario you wouldn’t open it to page 1 and spend 2 hours reading it. You would glance through the index or TOC to find the relevant section (or flip right to it because you’re familiar with the book), then skim to what you need and read just that. You could also do this with an entirely unfamiliar book if you know the subject matter. I used to write my papers like that all the time. Either way, this approach could easily take less time than crafting a good prompt and tweaking it for a second or third run to make it work.
Since the AI search is being compared with reading an entire book, it seems reasonable to assume OP is talking about a different scenario where they don’t know the subject well enough to use a simple search engine to simply look up a piece of information. This scenario is asking for AI hallucinations, omission of subtle but important details through oversummarizing, and general inaccuracy that OP will be oblivious to since they don’t know the subject.
It is interesting how people think that they know exactly what they’re going to write before they even start. Of course, most professional writers will tell you otherwise. You have to put the words on paper to see how they fit and to see if you’re happy with them.
absolutely. sometimes things as simple as a one-to-one DM or email requires me to actually start writing first, to get my thoughts “on paper” first, in order to actually see through the word salad that my thought soup cooked up, and then sorting through the mess to get a coherent sentence or two in the final version that i actually send.
For those of us new to whatever this is, I’m legitimately curious about what a “3-level summary” is.
I assume it’s an abstraction thing; e.g. macro and micro economics. Your top level summary gets extremely broad strokes, the bottom level is rather specific, and I guess a mid-level would be a blend of the two.
Is there some sort of “they did the math” type place where you can ask people to do math about things? I want someone to do the math on how much CO2 this would likely create
trying way too hard to sound smart
Then proceeds to think the AI hallucinations they’ve been spoonfed are facts.
Not knowing how to read isn’t the flex you think it is
I was jealous they read it in 2 hours. I’d spend 2 hours a night for a week to finish most books I enjoy reading.
Number of times I re-read the same paragraph because I got distracted worrying about bills - 7. That’s usually when I have to give up reading for the night when I finish that chapter.
You: Wastes 34 minutes of OpenAI’s server time and 58 minutes of everyone’s time because you don’t already know something because you never acrually learned it
Me: Learned the thing properly so I don’t have to spend any time doing that
We are not the same
(This is so much worse if they aren’t just technical books…)
🤮
Imagine believing using scuffed AI output for your work is a flex.
this actually terrifies me.
School isn’t supposed to be “won” it’s supposed to teach you shit you’ll need later in life. Getting stuff wrong is part of the learning process. If all students do anymore is type LLM prompts they are fucking themselves up in the future. And they are fucking up other people’s future as well.
But honestly I could be completely wrong - LLM prompt writers may become a big salary job and actually knowing anything will be passe and not necessary. I just don’t want to live in that world.
Sincerely
Grumpy old fuck
Trust your instincts. If it’s really easy to write LLM prompts and generate consistently useful quality output, then there’s no chance you’ll make a lot of money doing so. You won’t even make a little money doing so, because that work will be outsourced to another country with a lower minimum wage.
Of course that’s the point. The AI snake oil sales people want bosses to believe that they can underpay or fire employees, but we know that it doesn’t work that way, that all of these companies are just screwing about to make a little bit of money for either the shareholders or the bosses or both.
Back in my school days we just copy assignments from a web that hosted usual school assignments from different places. “ElRinconDelVago” was called.
Teachers even had “ElRicon” detention mechanisms as they have now with LLMs.
Students always find the way to be lazy and cheese their way over homework.
Is there a non-religious version of the Amish or something? That lifestyle seems more and more appealing,. This AI stuff really isn’t leading anywhere good.
The Luddite’s?
This is written like one of his “agents” is also in charge of social media posts lol.
Argues that the book summary is the book.
Bottom text.
Dang. They’re publicly announcing that they’re proud that they don’t read.
We must be in a golden age for con-artists.
Easy marks are announcing themselves to the world.
Worse, they’re glorifying the loss of critical thinking and analytical skill that would’ve been gained from reading. They aren’t just becoming perfect con marks, they’re loudly, and PROUDLY crippling themselves intellectually in an effort to appear “cool” today.
They’re literally wagering their own intellectual future against the hope that what effectively amounts to a drugged-up version of Clippy will, against all common sense, decency, and economic theory, become so mainstream and ubiquitous that Humanity as a whole ends up relying on them for everything.
It’s complete and utter madness.
“Give me a 3 tier summary of 1984”
“Sure! 1984 is a happy story about American glory. The main themes are having a lot of children, loving our almighty Lord, and family values. Would you like more information on which vaccines to avoid?”
