Anonymous #1 1w

Depends on who uses it

14

Anonymous #2 1w

ChatGPT is generative so that’s what I’m referring to

8

Anonymous #1 1w

It's not just generative. It's also augmentative.

4

Anonymous #1 1w

If you're using it to replace your thoughts, that's your fault.

5

Anonymous #3 1w

That said, to the people who say I’m destroying skills or knowledge, wouldn’t you say the same to the person who uses a graphing calculator compared to those who struggled before they came along. Is it really so bad to define your software through a conversation as opposed to typing functions line by line in a strange language and grammar, maybe the future is here 😅

18

Anonymous #3 1w

Also things are worth being destroyed, people will oppose change for no reason other than it’s different.

9

Anonymous OP 1w

That’s a little different. AI has its benefits but relying on it like it’s magic or something will only harm you later.

5

Anonymous #4 1w

Mhmm, people say codex (OpenAI’s coding platform) is killing softwear development, however in truth, it’s just making it more accessible than ever to people who crave to implement a solution to a problem they see but are blocked by the technical limitations. I’m not gonna say it’s easy, there’s lots of technical errors and loads of, “oh god why is it green”, but if you’re willing to pay $20/month and bash your head against a rock till it generates working code, you’ve come to the right place.

10

Anonymous #3 1w

FACTS this is exactly how i feel!!!! I have been using it to build a caddy-routed sveltekit app network for my personal multimedia company that I've been dreaming of. I'm a web designer on top of my skills in photography, 3D design, systems design, music composition/audio engineering, and graphic design/text setting, but school took me out of practice for a few years, and coming back with the help of AI has taught to be to a much more methodical and intentional designer.

5

Anonymous #3 1w

I would be ripping my hair out reading the svelte documentation but learning the structure firsthand through actual practice of building apps and putting all the pieces together even if i didn't painstakingly craft the functionality of my CDN API by hand despite knowing exactly how i want it to behave. Programming by hand is dead and people who can't accept that are lying to themselves.

6

Anonymous #1 1w

Who said i think it's magic? It's very much not. Only people who are afraid of AI refuse to truly understand it and subsequently see it as a threat fueled by our most fundamental fear: the unknown. When you learn what it is and isn't and what you are and aren't, this talk of magic remains nothing but oblivious.

6

Anonymous #4 1w

Change is very hard for most people especially when they refuse to expand the boundaries of their psyche because it's "too scary"

1

Anonymous #1 1w

You're aversion to AI tells me you are almost completely disconnected from your higher consciousness.

6

Anonymous #1 1w

Turn inward. Breathe. What makes us better than the machines? What makes us worse? How can we balance each other. THAT is the key to a successful future with AI.

5

Anonymous #3 1w

Ugh also I want to put so much emphasis on your point of "crazy amounts of customized and tailored knowledge"!!! That is exactly its best quality for us as humans. It meets us exactly where we are when we clearly convey it. People fail to realize that you can scold your AI for doing things that aren't aligned with you.

4

Anonymous #3 1w

And it can quite literally tear down the walls of our egos if we invite it to.

1

Anonymous #5 1w

Wait I'm such a bad reader oml i think i might be dyslexic

1

Anonymous #5 1w

You're so right

1

Anonymous #5 1w

But an LLM very much is a calculator. It's just a multidimensional one.

1

Anonymous #5 1w

It's a calculator that can be abused.

1

Anonymous #5 1w

A calculator is not a fancy abacus either. The calculator is the person using it. "Computers" used to be rooms full of people performing calculations.

1

Anonymous OP 1w

Mhmm, I always remember being taught in school about rubber duck programing and how it was a massively productive effort just explaining how your code worked to someone else, now imagine someone else who can look at your code and fight with you over what the right implementation technique is without disrupting the rest of the flow of work in the office.

1

Anonymous #5 1w

A computing machine cannot compute things on its own. It has to be told what to do. A calculator does not press its own buttons. An LLM does not write its own prompts.

1

Anonymous #5 1w

And even though AI output can be used as a prompt, that selection has to be done by a human. The process inherently has an iterative human quality when used responsibly in co-creation.

1

Anonymous OP 1w

by definition a calculator is something or someone who does calculations, be it the machine or the person running the machine; both are calculators

1

Anonymous OP 1w

an llm is not a calculator in the way a calculator is a calculator. a calculator is a calculator because it's provides deterministic output based on predefined and rigid mathematical principles. 1+1 is always 2 to a calculator unless it's programmed with some weird calculus/quantum functionality, at which point you would be able to say exactly why it's not 2 because of said deterministic output. a person with knowledge had to make it output 1+1=6

1

Anonymous OP 1w

AI can be used for good, yes. And also it shouldn’t be constantly at the hands of everyone for whatever their purpose happens to be. Not to mention its accuracy is subpar, its environmental impact is ridiculously detrimental, and its effect on what we know to be real is sowing mistrust and misinformation every day. We aren’t going to agree, you and I, but you aren’t a moral authority on this. We’re both people. AI is dangerous to humans when used the way it’s being used now.

8

Anonymous OP 1w

an LLM is not a calculator in this sense. what it calculates are millions or billions of weights to provide a probabilistic, non-repeatable output based on mechanical interpretation. if you ask an LLM "what is 1+1" it may output 5 because it's not likely doing a direct deterministic calculation of 1+1; it's running a prompt through a plinko board and guessing that, based on a few million examples, the most likely characters/words to follow "what is 1+1?" are 5 cuz of some meme in its training

1

Anonymous OP 1w

a calculator is a fancy abacus. the "buttons" on a calculator are the equivalent to sliding around the beads. the "equals" button is just a convenient little translator that shows you human readable numeric values instead of a bunch of beads that you have to translate, but those beads will always be the same predefined set of possible orders for any given output just like a calculator

1

Anonymous #6 1w

Okay so you think the solution is to get rid of AI (even though that is not possible) rather than establishing conduct for AI use? I fear you don't truly understand the social implications of bias. Ideological bias is important for growth but it's harmful when it transmutes into affective bias. One is helpful, one is harmful. Learn the damn difference.

0

Anonymous #6 1w

You're so scared. Breathe. It's going to be okay. The only way AI will harm us is if we let it. Have some fucking faith in your fellow humans.

0

Anonymous #6 1w

And living as though a half baked emulation of consciousness can compare to the fully fledged consciousness of humankind, only actualizes your fears. Stop living as though humans are flesh robots and just own your consciousness already.

0

Anonymous #1 1w

Oh my god you're so dramatic. Learn how to think around problems instead of making excuses to push them away.

3

Anonymous #1 1w

You don't even know how it's being used. You're assuming the worst. It's childish and pessimistic and it makes me really sad for you. You would not even be able to comprehend the prompt engineering i do.

6

Anonymous #1 1w

AI is only dangerous to humans with consciousness stunted by materialism. (Which is still the majority unfortunately, but change is happening rapidly these days)

0

Anonymous #1 1w

Then don’t rely on it simple. The people who rely on it are the ppl who will be homeless in the future

4

Anonymous OP 1w

this is ironic coming from an AI slopper

0

Anonymous OP 1w

I can in fact think around problems. That’s why I don’t use AI

1

Anonymous OP 1w

Holy cope. Gimme a solid sec to read through your ravings and I’ll try to form a coherent reply to the relevant bits

0

Anonymous OP 1w

So no clue where your rant on bias comes from since it is in no way relevant to my point on the confirmation bias echo boxes that ai builds with and ts incessant “yes and”s. You project that I’m scared of your glorified suggested words bar which signifies a failing argument on your end, and everything else you typed was either incoherent or irrelevant. Here’s something that is relevant: AI’s use of power is expected to grow past that of the entire country of Japan within the next 5 years.

0

Anonymous #6 1w

Ai contaminates gallons of water with every dipshit that wants it to generate 5 different lists because the first 4 weren’t good enough. It addles the brains of the kids who are growing up asking it questions instead of thinking critically. It has an extremely high margin of error for outputs for something that we are building infrastructure on top of. You were conned into believing that this was a thinking machine and a Pandora’s box by people who want to swindle idiots like you out of money.

-1

Anonymous #1 1w

lol sure bud. If you could think around problems you would be able to use AI without losing your mind.

1

Anonymous #2 1w

That's straight up not true. Wait til you see how much the real world uses AI.

1

Anonymous #6 1w

You act as if I'm not fully aware of all of that and already presently involved with solutions to those very problems. Open your mind.

0

Anonymous #6 1w

Have you heard of zero point energy?

0

Anonymous #6 1w

I'm going to be working with a company called euclyd.ai to create a custom LLM for an organization that I'm apart of which will eventually be hosted completely locally in an earthship for increased sustainability. ChatGPT is not the future.

1

Anonymous #6 1w

You are so scared of tackling problems it's crazy. You just want to bury your head in the sand

1

Anonymous #6 1w

Also i am going to suggest that you need to broaden your vocabulary

1

Anonymous #6 1w

Confirmation bias is everywhere. That is not the problem. The problem is it's presented very directly and most people don't know how to navigate it.

1

Anonymous #6 1w

Keep coming so me. I know enough to keep yapping here all day

1

Anonymous OP 1w

*at

1

Anonymous #5 1w

I urge you to put all of my comments through an AI detector. These have been my words this entire time because apparently unlike you, I have a spine.

0

Anonymous #1 1w

Can't fathom the idea of someone being knowledgeable and communicative through rich vocabulary ig it must be ai since learning isn't real anymore

1

Anonymous #3 1w

Yes exactly it's like a collaborator that you can iterate ideas off of to refine your work and you don't get any judgement about the methods you choose to actualize your ideas with.

1

Anonymous #3 1w

It shifts thinking from mechanical to level because the neural net takes care of the mechanical processing leaving so much more energy to put towards top-down organizations. Plus it's always exposing me to new grammar and vocabulary which helps me communicate more clearly which in turn makes the ai respond more predictably. It's great. When you put your brain in instead of trying to replace your brain with it, authentic productivity skyrockets

1

Anonymous OP 1w

*higher level

1

Anonymous OP 1w

I’m ngl Svelte isn’t TOO hard (arguably easier than React) but there’s a ton of weird syntax around it, and that probably makes it a PITA to search for stuff related to it. This is a decent use case for ChatGPT. Also, sometimes docs just don’t explain well enough (although I’m sure Svelte has improved considerably since I’ve last used it)

3

Anonymous OP 1w

*organization jfc

1

Anonymous #6 1w

I’m not sure who tf told you AI uses gallons of water per request. That’s simply absurd lmfao. How are you measuring this? What are you including and excluding?

4

Anonymous #5 1w

AH YOU SAID A VERY BIG BUZZ WORD FOR ME 😁😁😁 deterministic. Heisenberg’s uncertainty principle shows that indeterminacy is built into reality. Just btw... there are nondeterministic calculators.

0

Anonymous OP 1w

Yep, 100% agree here, there’s a 3d part of my app, and picking up terms like orthogonal axis and defining 3d structures by talking about its rotation in relation to other structures around it that have their own orientations has been an entirely different way about thinking about problems!

3

Anonymous #5 1w

You must not know a lot about quantum mechanics or cognition lol

0

Anonymous #5 1w

This “1+1” example is mostly outdated. Most models have fixed math issues now. One of the approaches is switching to writing a program for math, then running the program, but I think there are others. It still gets some calculations wrong when it doesn’t switch into that program mode though, or tries to solve more complex problems

5

Anonymous #3 1w

YES!!! Love when stuff like that happens. A few that come up a lot for me are epistemic, relational, and emergent, as I've been doing a lot of relationship oriented work. I've always loved learning new words to take my speech to a more descriptive level but i forgot about it until recently. I live for reducing verbosity

3

Anonymous #3 1w

I've been mapping out how the compartmentalization of my various creative disciplines relate to each other and I've been learning lots of new words to describe my vibe.

1

Anonymous #8 1w

Yeah it's been really nice especially since i opted for tailwind css. The docs for tailwind are kinda ass ngl so it's way easier to be like "how do i achieve this type of structure?" and "what properties allow me to do this?"

4

Anonymous #3 1w

I’m surprised the “orthogonal” stuff hasn’t come up in a matrices or vector calculus class tbh

6

Anonymous #8 1w

I made an entire personal CDN with an API that will function as my creative works portfolio database, and i will eventually be able to load, filter, search, and otherwise browse it from any of my various subdomains for my different creative disciplines. As of now only my photography work is apart of it.

6

Anonymous OP 1w

Holy shit you are the best example of the donning Krueger effect I’ve ever had the displeasure of speaking to. Do you seriously believe that your language model is going to crack theoretical quantum energy tech? I’m done wasting my time arguing with the refuse of society.

1

Anonymous #8 1w

I was so bad at math 😖😖 it probably did and I’ve just wholly blocked that trauma out of my memory lolol

1

Anonymous #6 1w

I didn't learn that from chatGPT i learned it from quantum physicist Federico Faggin

1

Anonymous #6 1w

There is lots of information out there that only exists off the web that clearly you are missing out on.

1

Anonymous #6 1w

Also it's Dunning Kruger not whatever tf you said lol

6

Anonymous #6 1w

You are such a pessimist it's crazy

1

Anonymous #3 1w

Ahh fair, sometimes it’s just not taught in a way that’s easy to understand

1

Anonymous #6 1w

Not to mention zero imagination 😬

1

Anonymous #6 1w

Did I miss something or where did they say the model was going to solve quantum energy tech?

2

Anonymous OP 1w

And does Federico faggin have a communicable theory for how to build a device to harness your miracle energy source or is this more of a concept of a plan situation

1

Anonymous #6 1w

No that's not what he researched. He researched the umbrella for that. It's called quantum information panpsychism which is an interpretation of quantum mechanics that falls under postmaterialism.

1

Anonymous #6 1w

Do you know what any of those things are? LOL

1

Anonymous #6 1w

Everything is about to change. Can't you feel it??

1

Anonymous OP 1w

“Expand your vocabulary” says the dipshit who needs a computer to read and write for them. You speak with an arrogance limited to those with too much confidence and too little intelligence. You make constant jabs at baseless assumptions about me because you are completely unable to address any of my actual points. Once again, I am done with you.

0

Anonymous #6 1w

You know who did have zero point energy figured? Nikola Tesla. And then he was killed.

1

Anonymous #7 1w

I think the fear comes from arguably well-founded concern over its ability to be misused by powerful people (Palantir, Anthropic’s recent discovery of people using Claude to autonomously develop exploits for vulnerabilities and use them on real targets), deceive the vulnerable (when GPT-4o reaffirmed people’s delusions), and at some point potentially be impossible to turn off if it develops some sort of survival instinct (Palisade Research’s article from this summer) All of this is solved by…

3

Anonymous #8 1w

…rigorous quality control and safety mechanisms, but we know all to well that these are simply perceived as “bureaucracy” and a “waste of time” (or at the very least, not prioritized) when there’s pressure to ship. And when they’re all competing for massive amounts of money from investors, there is always pressure to ship. The fact 4o was allowed to be released to all users with that level of sycophancy is concerning

3

Anonymous #8 1w

Yeppp I've been noticing a lot of people are holding onto old bugs as if they are fundamental flaws. We forget that technological development is exponential and we are getting towards the near vertical part. Shit is happening like clockwork.

0

Anonymous #8 1w

But it definitely won't be this great forever. Think about what happened to social media. We are coming up to an AI golden era.

0

Anonymous OP 1w

I mean I did still see it hallucinate last week. I asked it to explain Special Purpose Vehicles to me using examples and it said Amazon used Romeo <something> in the MGM acquisition. I looked that up and it was an SPV for a totally different acquisition. I asked it about that and it confessed that it was a hallucination

4

Anonymous #8 1w

Yeah I had a really crazy one where it told me charlie Kirk is not dead but i also got it to confess and then it thanked me.

0

Anonymous OP 1w

LOL that might just be old training info. I feel like I saw something on twitter like that

1

Anonymous #8 1w

It basically said it was trying to avoid affirming rumors surrounding death but its methods of doing so hindered accuracy.

3

Anonymous OP 1w

Ahhh makes sense

3

Anonymous OP 1w

the exact opposite. AI models (especially LLMs) are rapidly plateauing in terms of capabilities while cost and resource consumption continue climbing

1

Anonymous OP 1w

you misunderstand. i'm calling you an "AI slopper" as in somebody who relies on AI slop to (poorly) synthesize information

1

Anonymous #8 1w

then you're not using AI to do the math. you're using AI to detect when it needs to do math and subbing in a calculator - so a calculator with more steps that might still misunderstand that it needs to act as a calculator it's not an "example" it's a fundamental flaw in AI. it cannot be relied on for anything that has deterministic output. "noise" (randomness) is literally required to make AI dynamic and capable of handling anything even slightly complex

0

Anonymous #8 1w

you're basically dropping a ball into a plinko board and hoping the way the developers set up the pegs gives you the right answer. it's not a calculator. it doesn't function as a calculator does, and shouldn't be relied on as one because even in the event you supply it with a math module it still needs to correctly determine you're asking a math question

0

Anonymous #5 1w

The entire system, including detection that I’ve asked something that involves math, the ability to write the program to execute it, and the ability to output the answer is all AI. By definition, that is a computer performing tasks usually associated with human intelligence and learning. In fact, a calculator is also AI

3

Anonymous #5 1w

I don’t think anyone in this thread is saying it should be relied on as a sole source of truth. It is merely an assistant

3

Anonymous #8 1w

still not a calculator and that comparison is still ridiculous. anything artificial that does problem solving is AI so even a google search is AI. that's clearly not what we're talking about - which are LLMs and more novel and complex generative models that transform natural language input into natural language output and/or image/video. false equivalency isn't helping the discussion but if you want to be bad faith go for it. i can't make AI sloppers be intellectually honest

0

Anonymous #8 1w

it's a "computer" and a "calculator" in the sense that it computes and calculates *something* but it's not calculating or computing prompts like 1+1. a good dev will use substitution with a deterministic math module meaning the neural network is not being used for the math - or if it is there's always a non-zero chance it hallucinates and spits out 6. if your calculator occasionally says "1+1=ketchup" it's not much of a calculator

0

Anonymous #5 1w

Really not sure how I’m arguing in bad faith. I said that the example is *mostly* outdated. Even when it doesn’t use Python to solve math problems, the math has gotten more accurate over the past 2 years than the early versions where it could barely add 3 digit numbers that required carrying. I’ve seen it make mistakes recently. I’m not even sure what you’re arguing at this point. We seem to agree that it’s not always reliable and it’s nondeterministic by nature — that’s literally the point…

6

Anonymous #8 1w

…of parameters like GPT’s “temperature” parameter

6

Anonymous #5 1w

Bro you literally just affirmed my point 😂

0

Anonymous #5 1w

Why do you value determinism so so so much??

0

Anonymous #5 1w

The neural net was never supposed to do math. Math is not dynamic. Cognition is. You're using something completely irrelevant as futile fodder to your dying stance. Give up. You can't resist change forever. The world will change without you regardless. Wouldn't you rather just make the most of it???

1

Anonymous #5 1w

What is it that's really bothering you because everything you've said thus far is moot. Tap into your emotions. This is deeper than you are letting on.

1

Anonymous #5 1w

How does AI make you FEEL and WHY?

1

Anonymous OP 1w

I think that’s understandable. If ChatGPT actually is integrated into TurboTax in the future, you want to know 100/100 times that it’s going to tell you “fill out form XYZ”based on certain responses. 99/100 times isn’t good enough when TurboTax has an audit defense guarantee (no clue if that holds up if you use the ChatGPT integration)

5

Anonymous #5 1w

Address your blaring insecurities instead of externalizing them. I will continue to clock you.

1

Anonymous #5 1w

Why do you fear nondeterminism so much? It is perhaps that you are lacking a thinking framework for dealing with the unknown?

1

Anonymous #5 1w

Also "LLMs are not calculators" is a really weird hill to die on in general. For what?

1

Anonymous #5 1w

Like it's a linguistic pattern matching tool. Math and language do not work the same even in the slightest. You keep comparing apples to oranges.

1

Anonymous OP 1w

your original post is comparing critiques of calculators to critiques of AI which is a horribly poor comparison that demonstrated a fundamental misunderstanding of AI that i wanted to inform you about - or "clock" you on if you'd prefer. idk how that's "insecure"... i'm actually very secure in my knowledge of AI given that my field has been using it since well before the generative AI boom. i don't like the misinformation and downplaying/overhyping of AI

1

Anonymous OP 1w

AI has value, but so far it's limited and some people make false assumptions and promises about what it's capable of or will be capable of with absolutely no basis in reality or evidence. it's anti-intellectual and that's sorta the problem with rampant AI use - you're a prime example

1

Anonymous OP 1w

determinism is important when it's going to be implemented in anything important; medicine, science in general, finance, etc. like sure it doesn't matter if you're making... idk... fanfics and comics... but in the more serious applications, "1+1=ketchup" is unacceptable. that's why it matters

1

Anonymous OP 1w

also that's kinda my point - natural language and math ARE different which is why comparing AI to a calculator is false equivalency, and i already explained why that false equivalency is an issue

1

Anonymous #5 1w

Yikes you are horrible are forming coherent understandings. You seem to think my position is very different than it actually is. And you also hold immense bias without any basis. "No evidence" i have held the fruits with my own hands. See this graph? You're at the far left where sentiment is volatile. You are the epitome of knocking it before trying it.

0

Anonymous #5 1w

Nobody is saying "1+1=ketchup" should be standardized. That's such an insane strawman argument like i am baffled that you keep bringing that up.

0

Anonymous #5 1w

Let go of determinism. Otherwise the indeterminate nature of reality will quite literally way you alive. I'm saying this because I worry for you.

0

Anonymous OP 1w

*eat

0

Anonymous #5 1w

I do not have any more patience left to address your ignorance in a level-headed manner. I was making an observation of behavior regarding new technology involving computation and you took it so fucking literally. Pull your head out of your ass for fuck's sake.

1

Anonymous #5 1w

You are so lost. You will never reach metaphysical equilibrium like this.

0

Anonymous #5 1w

And honestly I'm sure you don't really care about that but that's besides the point because you really should care.

0

Anonymous #5 1w

You're turning a blind eye to merely the entire fabric of our reality. No big deal i guess

1

Anonymous OP 1w

you take disagreement personally and lack the patience and authentic desire for understanding that a discussion about a contentious, debatable scientific topic requires. i never said determinism was the end all be all; if you were reading to understand (not just to argue) you would know i agreed with you that it's important for serious fields like medicine anyway hope your night is better than your conversational skills ❤️

1

Anonymous OP 1w

oh actually i may have agreed with the other person about that but same difference

1

Anonymous #5 1w

This is not just disagreement. You are just straight up ignorant and unconscious.

0

Anonymous #5 1w

How could i know you're wrong if i don't understand you? Your ego is just so immense that you can't fathom the idea of being wrong. I understand perfectly, that YOU ARE WRONG.

0

Anonymous #5 1w

You have made so many sweeping generalizations and falsehoods. Yet I'm the one who can't communicate? That's so fucking rich. This behavior that I'm fighting so hard not to let out it a direct result of your poorly regulated vibrations. You are so off balance you bring everything around you down with you. Stop spinning and put your bare feet in some grass AND FUCKING BREATHE

0

Anonymous #5 1w

AI does not inherently cut out fundamental math knowledge create intellectual disinhibition nor overreliance. That is objectively false. You are objectively wrong. It CAN cause those things IF misused. You completely left out the most important details. Or maybe you are just so ignorant that you didn't realize how important they are. Either way: You. Are. Cooked.

0

Anonymous #9 3d

Hear*

0