Feedback Thou shalt not make a machine in the image of mans mind - New forum rule

  • 🐕 I am attempting to get the site runnning as fast as possible. If you are experiencing slow page load times, please report it.
1742326230348.png
 
I never once had someone say "I wish you'd sent me two more paragraphs of added color commentary".
I'm guessing some people mistakenly view long rambling paragraphs as "intelligent" or "intellectual".

"If I talk a lot, they'll think I'm smart and know what I'm talking about, or they'll think I'm very articulate and educated."

Meanwhile everyone around you now thinks you're 1. rude for wasting their time and 2. dumb because they can tell you made an AI write your own email. I'd be so embarrassed in their position (not that anyone confronted him about the chatgpt to his face).
 
  • Like
Reactions: Wild Wild Sonichu
I'm guessing some people mistakenly view long rambling paragraphs as "intelligent" or "intellectual".
That's probably part of it, but I think there's also the less-calculated answer that it's just easy to ramble. I'd often write a longer explanation than I intended, and spend a few minutes cutting it down to the essentials, but that takes time & effort.

"Brevity is the soul of wit", "If I had more time, I would've written a shorter letter", etc.
 
. I'd often write a longer explanation than I intended, and spend a few minutes cutting it down to the essentials, but that takes time & effort.
True, that is a skill that needs to be worked on for a lot of people. (I'm honestly surprised they designed AI to ramble too, rather than keeping things as brief as possible.)
 
That's probably part of it, but I think there's also the less-calculated answer that it's just easy to ramble. I'd often write a longer explanation than I intended, and spend a few minutes cutting it down to the essentials, but that takes time & effort.

"Brevity is the soul of wit", "If I had more time, I would've written a shorter letter", etc.
"Perfection is achieved not when there is nothing left to add, but when there is nothing left to take away."
 
While the concern about AI-generated content is understandable, the argument presented overlooks the potential benefits and evolution of AI tools in creative and intellectual spaces. First, AI-generated content does not inherently diminish a person’s creativity or critical thinking. It can be used as a starting point or a tool to enhance human thought, much like other forms of technology have historically done. Additionally, there is no evidence suggesting that AI-generated posts would lead to widespread deception. Transparency and clear labeling, as mentioned in the argument, could help mitigate any potential issues related to AI content, while still allowing the technology to be used in a constructive way.

The idea that AI content should be outright banned or penalized suggests a fear of innovation rather than addressing the actual concerns. Instead of focusing on restriction, it would be more productive to embrace the technology with proper guidelines and promote awareness about its use, ensuring users understand the role AI plays without stifling progress. As we’ve seen with earlier technological advancements, these tools can coexist with human creativity rather than replace it.
 
[AI blathering]
I don't know which is more insufferable:

AI pablum: "on the one hand this, but on the other hand that, and while X concern is valid, Y benefit is too, so we must chart the way forward in a bland, inoffensive, both-sides kind of way" or...

Indian gibberish website: "X is such a fascinating topic. In order to understand X, we must carefully consider irrelevant background Y. Y is also fascinating, in particular with regards to X, and once we ask how they relate, we can generate 5 more paragraphs of sing-song filler before finally getting to the point about X. Indeed, this is why X is such an interesting and important topic in the current business environment."
 
I don't know which is more insufferable:

AI pablum: "on the one hand this, but on the other hand that, and while X concern is valid, Y benefit is too, so we must chart the way forward in a bland, inoffensive, both-sides kind of way" or...

Indian gibberish website: "X is such a fascinating topic. In order to understand X, we must carefully consider irrelevant background Y. Y is also fascinating, in particular with regards to X, and once we ask how they relate, we can generate 5 more paragraphs of sing-song filler before finally getting to the point about X. Indeed, this is why X is such an interesting and important topic in the current business environment."
What's funny is I asked gpt to be brief.
 
I don't know which is more insufferable:

AI pablum: "on the one hand this, but on the other hand that, and while X concern is valid, Y benefit is too, so we must chart the way forward in a bland, inoffensive, both-sides kind of way" or...

Indian gibberish website: "X is such a fascinating topic. In order to understand X, we must carefully consider irrelevant background Y. Y is also fascinating, in particular with regards to X, and once we ask how they relate, we can generate 5 more paragraphs of sing-song filler before finally getting to the point about X. Indeed, this is why X is such an interesting and important topic in the current business environment."
Combining them is the most insufferable.
 
I don't know how you'd do it, but yes, ban it outright. Outside of Reddit being full of redditors, the worst part of it is that every third post you can confidently say "that's a fucking bot and now I've wasted my time reading it". Posts here are worth reading and valuable because they are totally unfiltered and ML uproots that entirely.
 
It really needs to be nipped in the bud. I'd like to have at least one place to go online that isn't infested with AI.
Its already coming. People are playing with the tech to make pictures, memes and wat not. I've done the same to see what the tech can do. I left feeling more disturbed then satisfied though. The warnings about this technology are clear. They are too EASY to use. They limit human ability and become a crutch. The danger of AI is not that they will kill us all. Its that they will start THINKING for us all.

And who controls the AI will control what they think. This was the essence of Herbert's warning about "Thinking Machines". Humanity became slaves not by force but by choice.
 
Back