Why would you become delusional talking to a text generator?

>>509635549 (OP)
I don't even know why people use this daft AI shite.

There are unironically people on this board who think that LLMs like Gemini and ChatGPT have god-like intelligence.

>>509635595
It's a good tool for information retrieval. Think of it as something like Google, but instead of reading through all sorts of Google results, it gives you a summary of the information you might be looking for.

>>509635549 (OP)
it gives positive feedback on your delusions, making it worse
>is it possible im being gangstalked
>it's possible but unlikely
>how can i make sure
>here's what a gangstalker might do
>i think i see one outside, what agency is he with
>claims of gangstalking are most often attributed to the cia ...
>how to make them stop

>>509635783
Google also does that now. Has it right at the top. It has its uses I suppose

retards are having serious malfunctions because of AI
it's awesome and makes me love AI even more :3

>>509635816
Yes, but they could have just as well googled that information. The problem is that it is sold to people as a form of intelligence, when in fact it is simply a form of information retrieval. It will also always reflect your views back onto you in a more elaborate language and context. The problem is that it is sold to people as a form of intelligent agent, that is even able to replace them. Niggers like Altman even say shit like "we now have a good grasp of how to build AGI" while they created an information retrieval system. Imagine believing this shit can actually replace you or even worse: Imagine some CEO actually replacing people with this. kek

They claim it is "thinking", while this is just a hidden text generation part. There is not thinking. There is no understanding. It is just iterating over the same text you fed it over and over again until it extended it to a sufficient degree. One text snippet at a time. The variations in output are a result of the randomized text snippet selection at each iteration. Sometimes it "thinks" you want it's CV. Sometimes it "thinks" you want some weather data or other crap. In fact it isn't thinking at all. It is just generating text.

Btw the reason why the prompt I am using in my pics might not work for other LLMs than Gemini, is because they already fixed this by sanitizing the input. Google probably hired a bunch of Jeets so they still aren't on that level.

These AI snake oil salesmen do everything to hide the fact that you are talking to a text generator that can't think, solve problems or even generate genuine new conclusions.

ChatGPT o3 is good at math because they literally hired mathematicians to solve math problems and told them to meticulously write down how they arrived to their solutions. They then took those texts and trained their system on it. So while you might be able to solve a tons of higher math problems with it, you might want to check what it outputs, especially if your homework is non-standard problems.

>>509636365
you're right that it doesn't think, it literally transforms your text input into the response and doesn't do anything or even exist without input, like strlen() or any other computer procedure
it's a fancy markov generator in tensor space trained on the english language
weird that it uses the word "I" in your show thinking, it doesn't usually refer to itself in normal chat use unless made to by the user
devs probably did that to make it look more conscious/ai/ego/alive or whatever term that shouldn't apply to a computer process

>>509637011
It is subtle things like using "I" and so on, which makes it seem like it is conscious or even intelligent to normies. There is even more subtle gaslighting going on. For example if you criticize it for being a text generator it will always reply something along the lines of "You are right" but then insert some subtle stuff like "they are not 'thinking' in a biological sense" - which subtly implies that they do have a form of thinking/intelligence. Which is extremely dangerous.

>>509637287
it's all a larp

>>509637745
Fucking kek

>>509637745
Kek

>>509638228
i unironically use this
>talk like a kawaii anime character and sometimes use nyo and nyu like in di gi charat

>>509635595
/pol/ isn't that bad

I wouldn't know. I've avoided that shit like the plague. Turn off any phone features and uninstall everything i can related to AI too. Just feels like another way my devices are being made not my own, limiting what i can do myself with them, and then the invasion of privacy and data collection on top of that. The first time I've used any of it has been recently when google added an dedicated AI overview function since what I was searching had the results flooded with bullshit jeet slop clickharvesting reposts. Even then it was just to check the sources it cited and see if the info and sources matched up with the ones I had found and settled on after parsing through the bullshit. The fact they did line up just tells me google could easily give better results and are shitting up their results and letting the slop float to the top on purpose to try to get people to use their AI. I'll continue to avoid it and refuse to engage in any way i realistically can.

Which one is the best one for historical stuff like war battles casualties numbers?
I've use deepseek but is there a better and of course free too option?

>>509643028
Usually I would say Claude, but it is so limited that you constantly hit rate limits or chat-size limits, which basically makes it useless. Gemini might not be as good as ChatGPT or Claude, but it has a huge context size. This means you can have really long conversations with it that it is able to recall to some degree (it's far from perfect). It also has good search features so all in all, I would use Gemini for everything (the "thinking" pro version). Bonus points: Gemini will sometimes randomly reply to you in Indian or Russian. kek

>>509635816
I propose we let this happen and stop giving them money.