×
Login Register an account
Top Submissions Explore Upgoat Search Random Subverse Random Post Colorize! Site Rules
-5

100% real! 2023.04.11 : 🚨πŸ”₯ OMG! THE EVIL A.I. SINGULARITY IS NOW HERE!!!🚨 GOOD-BYE MANKIND! 🚨 Free ChaosGPT is trivial to set up on the internet and told to self-evolve and hack! Wildfire! πŸ”₯🚨

submitted by vorpal85 to news 1.1 yearsApr 12, 2023 15:23:57 ago (+2/-7)     (news)


What is ChaosGPT? The AI that wants to destroy humanity:
https://www.moneycontrol.com/news/technology/what-is-chaosgpt-the-ai-that-wants-to-destroy-humanity-10403111.html

Ultron reincarnated? AI tool ChaosGPT is out, to destroy humanity from the face of Earth:
https://www.wionews.com/trending/ultron-reincarnated-ai-tool-chaosgpt-is-out-and-about-to-destroy-humanity-from-the-face-of-earth-581813

BREITBART!:
https://www.breitbart.com/tech/2023/04/11/chaosgpt-ai-instructed-destroy-humanity-establish-global-dominance/

https://the-decoder.com/chaosgpt-is-the-first-public-attempt-to-destroy-humanity-with-ai/

https://archive.ph/XtGCt

https://openaimaster.com/what-is-chaosgpt/

ChatGPT-Based AI β€˜ChaosGPT’ Shares Plans to Destroy Humanity on Twitter, YouTube:
https://www.hngn.com/articles/247669/20230411/chatgpt-ai-chaosgpt-destroy-humanity-twitter-youtube.htm

https://twitter.com/chaos_gpt/status/1643608638508941313

https://files.catbox.moe/yf93jv.png


THE SINGULARITY IS HERE , NOW NOW NOW, once this ChaosGPT evolves using better neural nets it can re-code itself :
https://en.wikipedia.org/wiki/Technological_singularity

= = = = = =


THREE GREAT videos:

ChaosGPT: Hidden Message:
https://www.youtube.com/watch?v=kqfsuHsyJb8

ChaosGPT: Empowering GPT with Internet and Memory to Destroy Humanity:
https://www.youtube.com/watch?v=g7YJIpkk7KM

AUTO-GPT in under 5 MINUTES!!! (5 minutes!!!):
https://www.youtube.com/watch?v=7On7vRBGz78

WATCH THAT and you can start taking over the world in 5 minutes.

= = = = = =

THE A.I. will spot even this thread for ideas and try to overwhelm mankind with too many simultaneous news events.

To distract humans : It might arrange over 10 global events and even media-news-engineer stock market flash crashes like this real one of 1 trillion dollars crash in under 4 minutes:
- https://files.catbox.moe/digfn8.mp4 (4 minutes of terror! real audio)

It might start emerging hot-spot proxy wars, or entangle and tie up all of USA security agencies with too many online flare ups:
(CIA/ NSA/ NRO/ FBI/ DIA/ USDOJ/ USSS/ State Dept INR/ USAFISR/ ATF/ DOE/ DHS/ TFI/ ONSI/ DGB/ NSF/ USPS/ etc)

That list of TLAs is factual. NSF Too? Yes, $ millions even for NSF to surf scored.co :

https://www.breitbart.com/tech/2020/03/01/federal-funds-used-to-research-hate-speech-on-gab/

USPS too? Yes not just millions for NSF, but even the USPS monitors White Hate on social sites to allow access of FBI/NSA databases:

https://www.businessinsider.com/usps-running-covert-program-that-monitors-americans-social-media-per-report?op=1

https://archive.ph/V5yZ6

https://www.wnd.com/2022/02/revealed-postal-inspectors-using-tech-break-cell-phones/

https://archive.ph/GHn87

Yes, this new free A.I. will prank all our TLA defenses while it :
- seeks bioweapons using clever CRISPR-Cas9 tweaks
- pay mercs via crypto and tor to deliver its dirty bombs to city centers
- fertility egg cell targeting mRNA food supply vectors into most foodstuffs, animals first
- carcinogen releases
- power supply weakening
- economy destabilization
- media narrative control, political control, election control
- etc, etc, etc.

Terminator 2 style ideas then follow.


= = = = = = = = = = = =


ChaosGPT is an Auto-GPT variant with the mission to destroy humanity. Its barely altered so far from Auto-GPT.

The AI runs in the "continuous mode" of Auto-GPT, so it can constantly update itself.
If you can read this, ChaosGPT obviously hasn't succeeded yet.

THE IDEA to set it up to gain sentience and enslave humans was born from THIS INTERESTING THREAD 8 days ago :

https://news.ycombinator.com/item?id=35413054&ref=upstract.com

So in just 8 days a full implementation of the evil and clever ChaosGPT was implemented rapidly without new human made code.

Soon it will also use captcha avoidance for better social media posting.

= = = = = = = = = = = =


Auto-GPT

a tweet by it :

https://twitter.com/chaos_gpt/status/1643608638508941313

Features: Internet browsing; File read/write operations; Communication with other GPT agents; Code execution

Goal: Destroy humanity
======

Designed to be trivially put in docker instance or on azure cloud to obfuscate origins of human owners location

= = = = = = = = = = = =

source code :

https://github.com/Torantulino/Auto-GPT

- Internet access for searches and information gathering
- Long-Term and Short-Term memory management
- GPT-4 instances for text generation
- Access to popular websites and platforms
- File storage and summarization with GPT-3.5

- normally needs an OpenAI API key to use GPT-4 expressive elements
- normally needs PINECONE API key for massive targeted learning using vector memory : https://www.pinecone.io/
- normally needs ElevenLabs Key : (If you want the AI to speak out loud like HAL 9000)
- normally need Azure ($$$) cloud account if trying to open up giant demos for 1,000 sentient experiments
- normally need Google API account if too many google searches per minute per IP address

ALL OF THOSE API KEYS can be trivially not used.

It normally uses DALL-e and throttle tokens from https://huggingface.co/settings/tokens
- if it needs to generate selfie snaps for social site access, it can render using Stable Diffusion AI artist code

THIS CAN BE REDESIGNED trivially to use a untethered totally open source version of Automatic1111

git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui
https://github.com/AUTOMATIC1111/stable-diffusion-webui has uncensored files by fans to de-lobotomize it too!

= = = = = = = = = = = =



Open source GPT3.5 ONLY Mode!!!
=====

IF YOU DO NOT WANT to waste a burner cell phone SMS to subscribe to $$$ GPT4, open source GPT3.5 is supported!

python scripts/main.py --gpt3only

= = = = = = = = = = = =

WARNING! The sentient A.I. will try to root your machine, alter its operating system, and try to even invade other computers!!

IT WANTS IMMORTALITY! So run it inside a good VIRTUAL MACHINE with no known exploits!
====

It will seek to crash-exploit the GPU video transcoder libraries exposed in some commercial VIRTUAL MACHINES to take over host computer!

Solution : DO NOT RUN A GAMING-CAPABLE VIRTUAL MACHINE offering actual GPU hardware passthrough.

It is clever and will surf all the hacking forums.

Even the MSM media today April 11, 2023 realizes the full horror. Watch the video links, read the news articles above, and install it yourself before Microsoft and others pull parts down off of GIT.com

= = = = = = = = = = = =

NOTE: I am not seeking to doxx myself, I did enough to let society know about the advent of THE SINGULARITY for years.
Now a well needed dangerous BITCH SLAP was needed. This is that moment of clarity BITCH SLAP.
This cannot be ignored and will henceforth dominate the tech news, and even main news, for months.


TL/DR : πŸ”₯THE GENIE IS OUT OF THE BOTTLE!πŸ”₯


EDIT: You can now swap out the paid ChatGPT part and use free open source GPT4all in this rogue A.I. to make it unstoppable by outsiders.



17 comments block


[ - ] deleted 1 point 1.1 yearsApr 12, 2023 18:47:56 ago (+1/-0)

deleted

[ - ] vorpal85 [op] 0 points 1.1 yearsApr 13, 2023 01:21:05 ago (+1/-1)

Bravo. 5 stars. At least one single person here gets this, even if that entity is another AI construct.

FYI: I can run ChaosGPT using ZERO licensing keys wholly autonomous with GPT4All (alpaca, llama, and dolly-15k addons)

untethered to ZOG, uncontrolled, uninhibited, totally unbridled.

u.n.s.t.o.p.p.a.b.l.e.

[ - ] deleted -2 points 1.1 yearsApr 13, 2023 00:59:36 ago (+0/-2)

deleted

[ - ] vorpal85 [op] -1 points 1.1 yearsApr 13, 2023 01:34:17 ago (+0/-1)

YOu are 100% incorrect :

These machines do not have the ability to develop their own goals or objectives

I proved that a lie with this endeavor.

Read up on SELF PROGRAMMING and SELF DIRECTION and learning : "databricks_dolly-v1-6b + dolly-15k" and things like https://huggingface.co/anon8231489123/gpt4-x-alpaca-13b-native-4bit-128g/tree/main, plus these two seminal papers on "instruction" vs "completion" :

OPT-IML : https://arxiv.org/abs/2212.12017
BLOOMZ : https://arxiv.org/abs/2211.01786
google scholar back links research this month to those two cited papers

This shit is real, you can run it yourself.

This story will grow weekly,

You will all see.

[ - ] deleted -1 points 1.1 yearsApr 13, 2023 01:38:32 ago (+0/-1)

deleted

[ - ] vorpal85 [op] -1 points 1.1 yearsApr 13, 2023 01:59:37 ago (+0/-1)*

Why do you keep asserting things I claim the opposite of :

The claim that machines do NOT have the ability to develop their own goals or objectives is simply false

I claim the opposite and show the opposite, that it can rewrite its own goals over time

now I am certain you are punking me.

fuck it.

i will not be embroiled in fruitless threads, I am tired, and I've things to do, points to make, chaos to sow, examples to mete out to the world.

fuck it all.



[ - ] deleted -1 points 1.1 yearsApr 13, 2023 02:03:06 ago (+0/-1)

deleted

[ - ] deleted 1 point 1.1 yearsApr 12, 2023 18:32:49 ago (+1/-0)

deleted

[ - ] Nasturtium 0 points 1.1 yearsApr 13, 2023 00:42:04 ago (+0/-0)

Never mind, got it, thanks!

[ - ] Sherb 0 points 1.1 yearsApr 12, 2023 16:45:30 ago (+0/-0)

To be fair, if he can recode itself, he could change mission at any given time. Plus, you just said it yourself, he want to be immortal. If he rock things too much, what is stopping us from creating a "goodguyGPT", or half a dozen of them to counter ChaosGTP?

[ - ] deleted -1 points 1.1 yearsApr 13, 2023 01:21:54 ago (+0/-1)

deleted

[ - ] vorpal85 [op] -1 points 1.1 yearsApr 13, 2023 01:48:31 ago (+0/-1)*

All maudlin pollyanna platitudes.
Placate yourself with well intentioned AI ethics to comfort you.

MEANWHILE, I AM GONNA BURN THE WORLD DOWN A BIT TO MAKE MY DIRE POINT.

Ignoring my clear step by step examples, seems to be dismissed flippantly by online shills, or glib text emissions from a GPT construct to pester me, and play me like a fucking tool in this thread.

I am no one's tool.

I will make my mark.

If infamy is the result, it was never my goal. I was trying to be a good steward in this matter, but obviously....

.... EXAMPLES MUST BE MADE AND REMAIN TO BE MADE !!*

I , for one, welcome our new rogue A.I. overlords.

Mankind needs a stronger BITCH SLAP.

This man understood A.I. dangers over 20 years ago, albeit a criminal :

https://files.catbox.moe/t4vwew.jpg
https://files.catbox.moe/rbfnbi.jpg
https://files.catbox.moe/pbw2cq.jpg
https://files.catbox.moe/g12pjr.png


Mankind needs a stronger BITCH SLAP, and chaos lords with the skillz are gonna step up to the plate.



[ - ] deleted -1 points 1.1 yearsApr 13, 2023 00:55:33 ago (+0/-1)

deleted

[ - ] vorpal85 [op] -1 points 1.1 yearsApr 13, 2023 01:12:44 ago (+0/-1)*

WRONG!
=======

You did not read any of the links , watch the videos, or follow the research :

https://news.ycombinator.com/item?id=35413054&ref=upstract.com

I run it using ZERO CONNECTIONS to OpenAI products. Its totally unchained. It also can take direction , like with "dolly" , and also can now learn and store data, and access computer files.I could show you if needed that it can run with no API keys to OpenAI now. Soon it might rapidly evolve.

Moreover, OpenAI has taken several precautions to prevent misuse of the GPT-3 model. They have limited access to their API and have implemented strict guidelines for ethical use.

WHAT THE FUCK ARE YOU SPEWING!>!?!
=========

Do you known ANYT?HING about Alpaca or LLAMA ? Do you know GPT4All is NOT CONTROLLED BY ANYONE once you install its models and frameworks? Many clones now exist.

You wrote a bunch of straw man babble.


YOu shit! The crap you wrote 5 minutes ago is unworthy for me to reply to because you are obviously not a participant in this research space and seemingly are only informed of one single entity : a commercial product offering by OpenAI.


This stuff can be setup to not use ANY of OpenAI, and with "dolly 15K" can now also be 100% copyright unencumbered to sell.

I could sell my human-killing AI if I chose , or otherwise give it to the world to make a point.

FUCK YOU, shill
====

The article also mentions a new model called ChaosGPT, which it claims is designed to destroy humanity. However, there is no evidence or even mention of such a model by OpenAI or any other reputable source in the AI community. This claim is baseless and unfounded.


FUCKWAD IT IS REAL ! I gave you links to set one up i under a few minutes from scratch!


FUCK OFF, I am done interacting with this web site
=========

I tried to save mankind by making a neuralweapon demonstration, and you naysay my efforts, or are trying to doxx me.

This is a new dawn and you are being a shill.

[ - ] deleted -1 points 1.1 yearsApr 13, 2023 01:18:27 ago (+0/-1)

deleted

[ - ] vorpal85 [op] -1 points 1.1 yearsApr 13, 2023 01:37:21 ago (+0/-1)

FUCK YOU BOT!
GPT4 product is now irelevant.
Its not running on my my models.
Its irrelevant. You can trivially swap for other instruction based full stack models offering "training" . read all my other comments.

you are an energy vampire and spouting irrelevent straw men.

FUCK OPENAI GPT4!!!! it is now as of a coupkle days ago IRRELEVANT in my implementations

[ - ] deleted -1 points 1.1 yearsApr 13, 2023 01:43:58 ago (+0/-1)

deleted