Rant -

If you are an illiterate fuck who can barely speak the English language, and you pepper every other sentence with "like" and "ya know", don't send me a multi-paragraph email that appears to be written by a professional negotiator with a perfect command of the written word.

It's exhausting and I'd gamble that you don't even understand what you sent me.

It’s not something I have or probably will ever have any experience with, but in thinking about it, a question:

It occurs to me that continued and more common reliance on AI for tasks like that could result in two things:

1. Users of that aid might see the products and become more educated and motivated to become better at communicating themselves without that assistance.

2. They might become increasingly dependent on the aid to the point that they become functionally incapable of communicating complex information without that reliance.

What’s your opinion? Is one or the other more likely? Does it depend on the individual?

Would restrictions on how AI products are used promote number 1? I.e., people would still use is for ideas and assistance in how to express them, but would have to reword the products to present them more like they would if they didn’t have the aid?

^ I could see both in equal measure.
I've used it a couple times to refine a message, but not to generate the whole thing with just a prompt.

Could be worse. I teach college courses and students use AI to complete assignments, learning nothing.

AI will do for critical thinking what calculators did for math skills.

quote:

Originally posted by sigfreund:
It’s not something I have or probably will ever have any experience with, but in thinking about it, a question:

It occurs to me that continued and more common reliance on AI for tasks like that could result in two things:

1. Users of that aid might see the products and become more educated and motivated to become better at communicating themselves without that assistance.

2. They might become increasingly dependent on the aid to the point that they become functionally incapable of communicating complex information without that reliance.

What’s your opinion? Is one or the other more likely? Does it depend on the individual?

Would restrictions on how AI products are used promote number 1? I.e., people would still use is for ideas and assistance in how to express them, but would have to reword the products to present them more like they would if they didn’t have the aid?

AI is making people retarded. There is no benefit to it for normal people. For advanced stuff, sure but all it does now is let kids write bullshit papers and give people stupid enough to trust it garbage answers when they ask it things.

quote:

Originally posted by sigfreund:
It’s not something I have or probably will ever have any experience with, but in thinking about it, a question:

It occurs to me that continued and more common reliance on AI for tasks like that could result in two things:

1. Users of that aid might see the products and become more educated and motivated to become better at communicating themselves without that assistance.

2. They might become increasingly dependent on the aid to the point that they become functionally incapable of communicating complex information without that reliance.

What’s your opinion? Is one or the other more likely? Does it depend on the individual?

Would restrictions on how AI products are used promote number 1? I.e., people would still use is for ideas and assistance in how to express them, but would have to reword the products to present them more like they would if they didn’t have the aid?

I think that scenario one is possible if someone is truly aware that they are deficient in their communication skills and is willing to review the AI generated content, learn from it, and then make the effort to put it into practice themselves.

Realistically though, probably 99+% of people who are using these tools are simply doing it out of laziness, and the natural outcome of this type of behavior will be the atrophy of any limited skills that they might have had to begin with.

We're going to see an overwhelmingly greater percentage of #2 than of #1.

quote:

Originally posted by tbgb50:
Could be worse. I teach college courses and students use AI to complete assignments, learning nothing.

...And then they will be completely surprised when AI replaces their job.

Could someone use AI to improve their own proficiency with a language? Yes.

Does Dunning-Kruger suggest that those who most need such training in their own language understand they are desperately deficient? No.

After a complaint to a business, I received an email which I have no doubt was AI-generated, after having bits of data input before generation: my name and the issue at hand, among other small details.

Yes, it's off-putting. It smacks of laziness and a general disregard.

But, the value of Large Language Models cannot be denied. It's a matter of proper application of the tool, and selection of the correct tool.

I wouldn't use ChatGPT or any other LLM besides Grok.

GIGO: Garbage in, garbage out

If the programming is shit, then the output will be shit. In this respect, only Grok is to be trusted, in my opinion.

I use Grok as a research assistant and when I do so, I make it clear when I utilize that data.

There's nothing wrong with using a reliable LLM to aggregate data for you. You can spend an entire day or more, plugging searches into search engines, visiting the web pages you're directed to, manually searching those pages, making notes along the way.

You can spend all that time, or you can utilize an LLM, which will perform those same searches and will then filter out the data you're seeking, and out put it to you in plain English.

Time. That's the chief difference.

Here are a couple examples of me starting with a personal belief and employing Grok in the capacity of research assistant to save a great deal of time.

Confirmation Bias part 1 and Confirmation Bias part 2

First Amendment and Public Prayer

That first example- the answer confirming my belief would have come about fairly quickly. but the second example would likely have required more time than I would be willing to invest, and even if I did find the answers confirming my position, what I would have posted wouldn't have been so well-organized and comprehensive an answer.

The key to using LLMs in this forum is to clearly state that the information was acquired via an LLM. As I've stated several times now, any attempt to pass off text which came from an LLM as your own, will be viewed dimly, to put it mildly.

quote:

Originally posted by Ronin1069:
Rant

I have actually watched a “young professional” type up a message, copy-and-paste it into GPT or such, then copy that into an outgoing email and send it.

I thought it was disgusting.

I use DoW AI to write my monthly TPS report. I doubt anyone ever reads it. I read one once and had no idea what it said. I try to make my work speak and not rely on speaking about my work.

Why can’t you have Chat GPT respond to the emails?

LLM ChatGPT reponds as such:

The whole point of using ChatGPT to formulate an email is to communicate more clearly and effectively. Not everyone is naturally good at writing, and being able to express yourself professionally in writing doesn't mean you have to speak that way in everyday conversation.

People use tools to spell-check, proofread, and improve their writing all the time. ChatGPT is simply another tool that helps organize thoughts and get a point across without poor grammar or awkward wording distracting from the message.

The important thing is whether the email accurately represents what the person is trying to say, not whether they could have written it perfectly without assistance. By that logic, we should also be questioning anyone who uses a calculator because they might not understand the math behind the answer.

My boss uses AI for damn near everything. It’s friggin’ disgusting. He ask 3 different AI sources, “Claude” and “Chat Gpt” seem to be his favorites, a question and the he sends that shit to me.
As an example he asked them a question or 2 about proper maintenance on the engines on the boat. I printed it out and it was 10 pages!

I have a guy who works for me, a native Spanish speaker whose English is very VERY limited. We communicate alright considering, with a few occasional issues.
But every once in a while I will get a text from him...
"Sir, an unforeseen situation has arisen and unfortunately I will be unable to participate the overtime project scheduled for this evening. Please forgive my inability to complete this project, I assure you there is no need for concern and that this task will be completed at the first available opportunity."

Wow Andy, your English is improving!

quote:

Originally posted by HRK:
By that logic, we should also be questioning anyone who uses a calculator because they might not understand the math behind the answer.

Hey, Chat GPT got one right!

I owed $99.87, hand the guy a $100, and he gets out a calculator to figure out he owes me 13 cents.

Do these ai responses get influenced by how the question is phrased and what type of answer it thinks you expect? For example, one might ask a question that sounds like one is supportive or against the 2A. Like ‘why would people ever want to eliminate the 2A?’ vs ‘why can’t we get rid of the 2A?’ Both question should hopefully resolve to the important foresight of the founders for a free people. But would ai respond with support of the perceived position of the question?

quote:

Originally posted by konata88:

Do these ai responses get influenced by how the question is phrased and what type of answer it thinks you expect?

Have you asked AI?

Use Grok. It's as close as you're going to get to no bias.

The others? I wouldn't piss on them if they were on fire.

Come to think of it, it's a bad idea to urinate on any electrical fire, but you get the idea.

We got this dipshit in the Army who does it. It’s especially bad because he’s a native Spanish speaker and very obvious it’s not him who is writing it. Idiot even writes his counseling statements with it and uses terms such as the four pillars of leadership. How dumb is this guy? If breathing weren’t automatic, he’d be dead.

The LTC tried to make an excuse for him by saying he’s ESL. She stopped when I told her I’m ESL.

You’d think someone with a college education would have some basic grasp of writing like an adult.