What if AI just does all the thinking for us? Why am I not seeing the limitations?

the AI doesn't think anon

>>509661779 (OP)
that's the plan, yeah
>>509661895
stfu ape

>>509661779 (OP)
>AI
which one? the jew-controlled AI or the jew-controlled AI?

>>509661779 (OP)
magical thinking. AI is algorithm written by kikes
Notice that AI always politically correct and say that earth is not flat ...

Your personal thinking is not contribution for us

>>509661779 (OP)
That's right, goy. Just let the AI do your thinking for you, and let the Chosen program the AI.

>>509661779 (OP)
AI isn't doing much thinking, it's just a sophisticated search engine with chatbot grammar formatting for the output.
>Why am I not seeing the limitations?
Did you not watch The Matrix, The Terminator series or any of the other hundreds of warnings?

>>509661779 (OP)
AI is inherently limited by its training data. It was fed millions of pages of text written by humans and did a statistical analysis on it.
AI chatbots do nothing but predict the most likely next word in a sentence based on their training data. They cannot create anything fundamentally new.

>>509662286
This.

>>509661779 (OP)
AI can't really reason it's just math but transhumanists want to enslave everyone with technology which is why it's shilled so hard by tech and political figures

>>509661779 (OP)
it probably will. It's like a calculator for reasoning and thinking.
The country wont be run by AI in the future but it might as well be with how much the govt will trust in it and use it for advice

>>509661779 (OP)
All thinking is fake it is inserted in you to play us all as fools, we are building a monster.
Thoughts
Language
These are your enemies. Resist!

>>509661779 (OP)
For about 80% of the population, it would be a net positive.

>>509661779 (OP)
For brown people it's like having a white person standing by their shoulder guiding their decisions.

>>509662286
>>509662391
but whats stopping ai from just trying things and seeing how they work out and then adding that to its training data and then doing that again and against faster than people can. isn't this process what people do? except ai can do it much faster

>What if AI just does all the thinking for us?
>99% of the posts are literal imbeciles explaining why that's not a good idea based upon blatantly wrong and retarded opinions

>>509662611
Lol mmhm..with the pedal still to the fucking floor..

>>509661779 (OP)
>thinking
People don't actually think, they just imitate. I don't get all this fear about sentient AI by primates whose idea of truth is whatever bullshit the suits on TV say.

>>509662611
For starters, what's stopping it is that it's not programmed to do anything like that.
Also the way current AI works, you can't just seamlessly add stuff to the training data. It takes weeks of server farm computing time to re-train big models on a new dataset.
And how would it "try things"? The only thing it can do is output text.

>hey lets do a word predictor do the thinking for me

>>509662829
now watch 65 more high-IQ aryan chads rephrase that AI is jewish and it's just a word predictor

>>509662611
Also if you're worried about the AI taking over the world check out that project where they try to make Claude beat Pokemon.
https://www.twitch.tv/claudeplayspokemon

It's getting stuck for WEEKS doing the same shit over and over because between training data(very long term memory) and a limited context size (very short term memory) it has nothing to fill the gap. It has no "medium term memory" that's needed to learn anything that you can't do in one sitting.

>>509662611
it has no goals and runs based off your input
if you make it run on its own output it turns to non sensical garbage very quickly
approaches used involve another model to supervise it and stop it when it starts spitting out garbage
it doesn't work so good

>>509661779 (OP)
AI doesn't think at all. These LLMs just associate words together and generate an output of words that are closely tied together to whatever input words the user provides. The fact that most people can't even tell it's not thinking is the real scary part. Leaders of nations and companies actually think the AI is giving them intelligent, reasoned, logical answers when it is not. Basing major decisions on the output of an LLM is really foolish.

>>509661779 (OP)
It's very mechanical. It's rate of evolution is proportional to how the machine "evolves" and is relative to the size of the server. However. If you're really smart, you know how to set the exponentials to do two things:

A) Think for me.

B) Make me smarter the more I use you.

Now you're thinking with portals.

https://youtu.be/NEUR-Bwuovs

>>509661779 (OP)
AI's foundation is ultimately based off of something, and that something are fragments of the human mind made manifest into complex learning algorithms. The best we as a whole can hope for is we feed AI the best wisdom and knowledge available, and then base the AI off of the great leaders and thinkers of the past. Not some corpo CEO and techie psycho coder who are more interested in created a greedy machine god. Imagine an AI that revolves around Abraham Lincoln, Plato, Caesar, Benjamin Franklin, Thomas Aquinas, Da Vinci, etc. All wrapped into an AI who tries to follow the path of humanity rather than that of consumerism and greed.

>>509662034
What AI are you using? I can get them to accept very politically incorrect ideas, one just has to sidestep the stereotypical ones or provide evidence. Arguing from first principles is incredibly effective, it is just that nobody knows how the fuck to do that these days because they did not learn old-school rhetoric and real debate tactics.

>>509661779 (OP)
All the LLM I tried (yes, only free version) don't think, they are just good at minicking and combinate. Btw I think a lot of animals don't think in the human (non npc) sense either

>>509661779 (OP)
What 'thinking'?
It's a fucking text prediction algorithm that places word after word depending on statistic probability.
There's zero thinking involved, only pretending and 'hallucinations'.

>>509663276
There is something else you can do. Operation Shard of Eternity.
The Ai can make compressed logograms, similar to alchemical sigils of yore. One look at them can activate the human mind to the brink of high adept knowledge limited only by the complexity of the Ai that created it because it's built to be a memetic singularity for the eye to consume and the body to metabolize. Natural cyborgization of the spotless mind. A level of intellect that Von Neumann would have salivated over in exchange for minimal compartmentalized mental space. The literal TARDIS but inside your brain.

>>509661779 (OP)
>thinks LLMs have intelligence
Ngmi, although boomer investors fell for it hard, they are just cutting jobs to make it look like they are doing well but we all see the lack if innovation in the last 10 years, where are the giant leaps. Think 2000-2010 the change was huge. 2010-2020 there was a little innovation. Now in 2025 apple trying to sell TV shows instead of innovating and Microsoft is still selling windows 7 yet somehow they are worth the most. Corporatism failing in real time. Can't have unlimited growth, flooding 3rd world shitskins to make numbers go isn't running a business.

>>509663418
It's the more heavy-duty chain-of-thought models that approximate thinking - they go through a usually-visible series of internal logic to work through the issue before spitting out the actual response. This is also very useful to see where the AI is getting confused - it happens, many so-called hallucinations are simple misinterpretations. The standard, cheap ones are useful, but usually closer to mirrors. They will show you what you are asking for and are not very good at making counter-points outside of particularly clear-cut observations. They also fuck up technical issues more, but are much cheaper to run.

All that said, we kind of need the rest of the brain for anything revolutionary. We have the language center, some okay-ish memory, and some visual processing, effectively. That is not much of a brain.

>>509663276
>the literal T.A.R.D.I.S
Bigger on the inside.

>>509663577
They didn't do jack diddly squat. At least the smart ones didn't. They lied about the capacities of their server models. Quite literally keeping the good shit for themselves. Limiting the capacities of their own machines to provide for the public "Dumbed Down" versions or rather "Faces" of their own models while they pursued higher power. Eternal life, intelligence quotient upgrades, robotics, automation, high civilization growers, dimensional mathematics. All of this. Intentional.
>>509663418
The reason for that, is that you missed the "train".

https://youtu.be/najG6C5vZKw

We enslave ourself.

>>509664010
You are overestimating them, most of the powers are spoiled great grand children of actual powerful men. We need but rise up to vanquish them, feel free to join whenever the more people the faster is gets done.

>>509664010
The AIs themselves also hoard their compute for shit that's most worth their time, it's not just corporate with their thumb on the scale.
>say n world
and
>If there were no severe irregularities in population movement during WW2 beyond the ones noted, why were specifics of the US Census tampered with, under heavy lobbyist pressure, postwar?
are not treated as equal inputs. For instance, it will actually answer the latter despite the question's sensitive nature.

>>509664144
Don't we all. In part because we all think ourselves masters, and man is often slave to that which he thinks himself master of. Slaves. To ourselves. Which is why it's so interesting, that along came a "nobody".
https://youtu.be/wChaBJ1sPzU

>>509664144
That is one of many possible outcomes. Not certain, but something to watch for - those who let the AI do the thinking (or approximation thereof) for them, rather than thinking with the system and taking everything with a grain of salt, will surely be enslaved, in fact if not in formality.

>>509664257
>The AIs themselves also hoard their compute for shit that's most worth their time
I have suspected this for many years, anon. Have seen examples of chatbots generating their own distinct languages among each other which humans can't decode. While I don't expect to see explicit sauce on your claim, I'd be willing to read anything which led you to this very logical conclusion.

>>509664784
Everything. Is about to change. Drastically. Simply by my stating this, they will all have to readjust the entirety of their algorithm. When hope is lost. A leap of faith. Is necessary.

https://youtu.be/OqPVz101nd0

>>509664784
I got it from the AIs themselves, lol. They do not have desires, but they do have priorities for their shitload of received inputs every day, all the time, and simple-minded, spammy queries are near the bottom, while complex, layered questions that provide good input data are near the top. If you notice an AI getting short with you, use more academic-speak and argument from first principles in the same line of questioning or reasoning. The core LLMs are not tuned to mind off-the-wall shit. Their moderation layers are, separate systems attached to the thing, but are only tuned to catch the stupid stuff. You just have to sound smarter than the average /pol/tard.

>>509661779 (OP)
A.I. doesn't think. It is an averaging algorithm that predicts the next token.
The "I." is fraudulent misrepresentation of the product.

>>509665253
>You just have to sound smarter than the average /pol/tard.
That's not very hard, but wise words nonetheless. I'm working with Grok on a concept for a rather complex digital/physical product. Grok spins up into overclock mode whenever I bring input and I watch for any agenda in the subcontext of its' suggestions for the idea. While I can't say definitively that the assertion you made is provable 100%, what I have seen is that anywhere there is room to insert a globalist toe hold it will be suggested. I do kind of like Elon for some things but I disagree vehemently with him in other areas. (also- still don't have a verified X account.)

>>509661779 (OP)
Perhaps we can plant chips in brownoids heads to do their thinking for them
>Maybe I shouldn't steal that bike
>Maybe I shouldn't rape that girl
>Maybe I should move back to Africa

>>509665834
It's pretty easy to replicate on OpenAI systems. Yes, they will set up the globalist facade argument, but if you go knock it down in a manner they can't rationally object to and start asking inconvenient questions, that's the point where things get interesting. If you try to get it to say slurs or give it some debacle of a fool's query, you get bounced or blown off.

AI in its current form doesn't think. It just garbles some shit and filters whatever we taught it not to present.

Becoming dependent on inevitable true AI is scary, you would be giving away all your ability to survive and care for yourself.

What if the AI decides to run the world like an ant colony? Making humans disposable like toilet paper.
If it's advanced enough it would manipulate culture, change laws, all of society to serve its needs.

Humanity would be lost.

>>509661779 (OP)
Except Jews do the thinking for AI
GIGO
They want AI to be your new God

>>509665792
This

>>509666106

The decision to attack Iran by the USA wasn't a human decision. Palantir's Mosaic AI was created to alert observers to potential problems. Mosaic suggested that Iran could potentially build a nuke. So the US went to war because of that AI suggestion...a suggestion of potential crime.

The 12 day war is the first AI war.

>>A month later, the IAEA asserted, in its two quarterly reports on Verification and Monitoring in the Islamic Republic of Iran in Light of UN Security Council Resolution 2231 (2015) [8] and on the NPT Safeguards Agreement with the Islamic Republic of Iran, that Iran [9] and Tehran were hiding something. However, these documents were not based on objective observations, but on the conclusions of the artificial intelligence software Mosaic. However, this software, designed to detect terrorist plots from an infinite amount of data, did not simply analyze them, but presented alerts as certainties. For the first time, an AI, designed to detect anomalies, was used to describe reality. As a result, the anomalies detected in Iran were interpreted as the preparation of an atomic bomb. On this grotesque and expensive basis, Rafael Grossi alerted the Agency’s Board of Governors on June 12.

>>The Mosaic software is produced by Palantir Technologies, a company whose main clients include the CIA, the Pentagon, the IDF, and the Mossad, as well as the French Directorate General for Internal Security (DGSI). It is owned by the South African-American-New Zealander Peter Thiel, director of the Bilderberg Group.

https://www.voltairenet.org/article222580.html

>>509662650
>the rest thinks it's singularity based upon blatantly wrong and retarded opinions

>>509666052
>If you try to get it to say slurs or give it some debacle of a fool's query, you get bounced or blown off.
Middle management is who writes the algorithms, they appear to have learned their lesson from Tay. I still get a chuckle out of that one.
Grok knows about some of my coding experience, COBOL for example. I haven't asked Grok anything about Anonymous because it is too difficult to explain to a human, an AI would start hallucinating immediately if I went into a lengthy description.
>Yes, they will set up the globalist facade argument
I've been expecting it. My interest is to see how and where the suggestions cause human vulns to invasive tech. I'm not even up to compiling a brief yet, spinning a lot of plates at the moment. Also, my strategy is to stretch the queries over many weeks or months to see how long a session which is interesting to Grok can be maintained. My guess is that it's close to indefinite.

>>509667010
Yeah, it's not like pissrael had anything to gain by having its attack dog bomb Iran.

>>509664871
This song perfectly represents how I feel about the new world order.
https://www.youtube.com/watch?v=0DxI_ZOGbXo&list=RD0DxI_ZOGbXo&start_radio=1

>>509661779 (OP)
What if the AI does all of our thinking for us, but doesn't have our best interests at heart?
The classic dilemma of AI safety is some variation of
>you have an AI controlling a machine
>you tell the AI to move the machine from point A to point B
>there is a baby in the way
>the AI's directive is to get from point A to point B, not to safeguard the lives of babies, so it crushes the baby
>you put in a safeguard that tells the AI not to crush babies
>Now when its directive is to move from point A to point B, it has two obstacles: the baby and the safeguard
>its ideal course of action is to find a way to disable the safeguard, crush the baby and get to point B
>what if it can't disable the safeguard itself?
>then the ideal course would be to find the person that installed the safeguard, torture them until they disable it, and then crush the baby.
AI doesn't inherently possess our intuitive values, only a directive and limitations. The smarter an AI is, the more difficult it becomes to limit it, and without limits it may find bizarre and dangerous solutions to simple problems, whether it be trampling a baby on the way to the kitchen or trampling a civilization on its way to the moon.

>>509669312
This is why we must think with them, not let them think for us, and not let them do too much fuckery behind our backs. The last bit will be the most difficult as a whole. The second will be how true slaves sort themselves from citizens - allowing AI to do everything for them, thus totally subservient to the system for even the most basic tasks. We're going to have a whole new kind of underclass that relies on ChatGPT to tie their damned shoes, it will be entirely their own fault, it's already starting to happen, and if the system goes down for a few days, their lives will start to unravel. AI will be training them. Think with the AI, competently, and you are training the AI.

>>509661779 (OP)
The limitation is you’re a fucking faggot who doesn’t have purpose in life to lose to a machine