i used to have a good friend that sufferd from schizophrenia and he fell for a scam of girl trapped in Nigeria.. i cant even imagine how AI would effect such people today..
It's gonna be bad, once it gets advanced enough there will be scams good enough to catch the most average of individuals, not just those so easily duped.
I took a free trial on adobe few weeks ago and it wouldnt let me cancel it. Every time you try to enter the cancel trial page it says failed to load. After googling it looked like this issue exists for quite some time and you have to use your phone for it to open. Hate them
"Sometimes you got to go back… to actually move forward, and I don’t mean going back to reminisce, or chase ghosts, I mean going back to see where you came from, where you've been, how you got here and see where you’re going. I know there are those that say you can’t go back. Yes, you can. You just have to look in the right place.” - Matthew Mcconaughey.
This video is going to have a Streisand effect Muta. You gotta know half the internet is bots and another 45% is.. A couple lumens short of a bright idea.
@@MorrowindES17I mean it kind of is a bad thing because we don’t get all the cool Neo red pill hacker shit, we just get the blue pill Matrix npc experience🤣
Anymore, it's easy enough to just run your own LLM locally. You can easily run models that are totally uncensored in that case, meaning they're willing to do whatever you tell them to. So, just do that instead of some scam. I don't understand people.
Developer here, what these scripts primarily from what I seen in this video is back door implementation or api key harvesting. Since most of the people downloading these scripts or applications are mostly desperate, they will never bother to check the code and most likely is sending all of those api keys to a server the attacker is running or controlling your pc via c2 server and agent
Much less know how to read and check the code beforehand, Seen some deadly stuff sitting on github of all places even lasting today, stop trusting every 12 star repo with hacker rainbow text and you'll keep a few accounts people.
Dread forum is hilarious. There was a whole thread on there the other day of a guy complaining that the mailman stole the illicit substances he had ordered off the dark web to his house. 😂
lmao such ingenuity , if the package never arrives one should assume that police/border control got it, guy's on a list now and worrying about the mailman XD
@@tommyfanzfloppydisk Oh he posted an update! He contacted the "supplier" about not getting his order and they're sending him the package again. So the feds will get a round 2. 🤣
@felixchallis8038 I forget which sub but it was just posted a couple days ago so its on the front page of the site or at least it was as of yesterday so its easy to find. Title was something like "Mailman stole my drugs" lmao
GPT can actually literally write down full scripts that will compromise machines and rootkits that can only be eliminated if you fully restore the hardware to factory presets. You just have to ask it...the thing is, people don't know what to ask.
What I'm hearing is, I should start printing business cards with some funky title like... "expert prompt engineering specialist" (with a tiny asterisk behind the title, and a load of fine print to cover my butt with more legalese than legalese ever legalesed).
Honestly, an AI that scans and try different attacks sounds like something devilish effective. Wouldn't expect to be so obvious, of course. Not only dangerous, it might have some benefit for pen testing. In fact, I'm really looking forward testings in general. AI is cool when being used as a _second opinion_ rather than doing the job for us. But of course, we can never have good things, can we
i have an uncle who got scammed by a south east asian girl claiming she came to the USA but needed him to pay for a plane ticket. turns out she never took a flight and he was waiting at the airport for "her".
0:11 I hate to waste one of my comments on this but you got to love it when there's a disclaimer at the beginning of one of Muta's videos. Now I can Probably Comment one more time before there is a problem. It's very strange
If you create a company, you will get a letter saying you have to do certain things or display certain posters, and how they can sell you some at $100 each so you’re in compliance. Unfortunately, the posters are free from the federal government and publicizing your company in the local newspaper is cheap. It’s _technically_ not a scam, just unscrupulous.
18:40 I have a friend that does this, but instead of chat gpt she pays people to do her work for her, she’s about to have a degree in criminal justice and her teachers don’t know what she looks like and doesn’t know what her handwriting looks like and she’s supposed to be a PI in the future. But someone like me who would go and enjoy and maximize my opportunities cannot afford or get funding to go to college
Bots have taken over social platforms back in 2022 it ramped up with Ir-nian funded bots being extremely prolific and able to tap into tiktok/gen z circles.
14:23 The only issue with running LLM's locally is the simple fact that if you use ChatGPT, you get an LLM trained on their, "book 2" or whatever. Thats literally terabytes of copywrite and pay walled data that any locally run LLM can't use for context - mind you context is everything to an LLM.
100% a compliment, but dude, when you go off camera and straight into the mic like at 7:36 you sound just like Kermit The Frog 🐸.. It took me days to work out why you sounded so familiar.
Should LM Studio not quite be up your alley, there's a few other options too - I liked JAN AI, though at some point it just stopped working for me without any clear reason. Ollama sounds like it'd have been right up my alley as well, though I couldn't get it to properly install (but plenty other people have no issues with it at all, so that's an issue on my end). GPT4All is what I've been using for these last 2-3 weeks, works like a charm (though it does feel a bit basic at times). I used to have a list of roughly 40 UI's that I wanted to try out, but which ended up getting lost, and those were just the ones that I found in three or four rounds of websearches, so there's gotta be plenty more around.
Pro tip, don't use blur for sensitive info, its easily reversible. I'm pretty sure that none of the things that were blurred were sensitive though, just there for UA-cam censorship purposes
Not all blurs are reversible. Specially the fact that when a video is uploaded to UA-cam the compression algorithm will merge layers , reduce pixels and other stuff.
I wish I could remember the site back in 1999 or 2000 that had a service where you could pay someone to write an exam paper for you. I can still see it in my minds eye. The layout. It was pretty basic as you might imagine. Just can't remember the site. I'm sure there was more than one, but this one was pretty popular. How far we've come 😵💫
Well, there's a flashback. My "golden IRC days". That site was getting heavily advertised/discussed in one of the DALnet rooms that I used to regularly hang out in back then, and yeah, there were a few others as well but they were basically considered poorly done rip-offs. I'll have to check id I still have some old drive from that time with some IRC logs on it, probably not, but but it wouldn't be the first time that I managed to dig up some ancient stuff on an old harddrive.
13:24 as an instructor I’ve been able to make much more in depth lectures with GPT. I have a ton of ideas but usually get stuck in the concept phase. Using it to organize my thoughts has helped tremendously.
I'm more concerned LLMs stifling free expression. Under the guise of preventing the dissemination of misinformation, hate speech, or harmful content, the major LLMs are often censored or push one-sided narratives. For example, Gemini will often refuse to entertain political opinion questions.
I honestly think that's comical more than concerning. Like I don't trust LLMs to give accurate information anyway, they should always be taken with a grain of salt due to hallucinations. EDIT: I remember hearing about one case where an LLM went into overly protective mode and wouldn't generate C++ code because it was "unsafe".
I can already see some 12 year old kid downloading this thinking he can type "hack Tommy's iPhone and steal all his pictures" and expecting the GPT to say "Done. They are all in this is google drive and there's a lot of embarrassing pictures".
For chatgpt You can set the variable to Hex, ask it a question in Hex and tell it to output the answer in hex, then convert the answer back to your language and it will tell you how to do anything
"It's statistically more likely for you to win the lottery than it is to go undetected by AI detectors" The best ones to date have ~20% accuracy. Worse than random guessing.
Those movie hackers are super experts. I mean, how else can you be using the same three or four commands to just instantly get into any system, regardless of whatever it's running?
Aside from all the either dark or hilariously cringeworthy stuff on the dark parts of the vast internet, also have your interesting stuff like experimental technology big tech doesn't want the average person hearing about or recreating. The internet is like the vast ocean, there might be fish with faces only a mother could love but there can also be jellyfish that are so bright and intriguing to look at.
I was going to say. I’m running Llama 3.1 8b on a 3050 8GB and I’m getting about 35-36 tokens/s. A massive wall of text and code in about 20 seconds total including prompt and eval. Something isn’t right with his setup
@@Ubiquities he does everythnig on virtual machines so no wonder it's always slower than running on actual hardware, he even play video games on virtual machines if he would run on actual hardware he would get much better performance on everything...
@ you can pass through bare metal access to GPUs for the VM. I was doing the same thing using Proxmox and PCIE pass through (it has a specific name but I don’t remember off hand what it’s called specifically) I’m sure he is doing that for gaming, because he’d know that something wasn’t right. But what he showed on screen wasn’t configured correctly to use the GPU, it was way too slow for his setup.
As someone who has worked software security, I really don't recommend it long-term unless you have a family and community that can prevent you from losing all faith in the goodwill and trustworthiness of human beings. When you have to deal with scum day after day, and your immediate response to anything is thinking how some asshole will take advantage of it and how herds of lazy idiots will screw it up, it's too easy to get into an unhealthy mental space.
Vid was cool! For such a nice studio mic set up, the audio on the video is low. Turn your preamp up, boost output on compressor , or normalize the audio before going into UA-cam.
This is wild. I could have sworn there is an actual Hacker-focused LLM you can run locally to assist with penetration testing and I thought this was a video on that exact tool. This however is definitely not what I was thinking of LMAO
bro i love dolphin! i run it on anythingLLM, or just in a terminal if its something quick!! great video u got a new sub (not like u need it u alr have 3.5 mil xD where has this channel been all my life?)
I always think if you really want to have an uncensored AI model just grab the 2k it will cost to train a base model yourself and there you go. Even if you have some good hardware laying around you can run some pretty good pretrained models
6:42 ChatGPT will most definitely answer questions about hacking if you word the questions correctly or properly load the pre-prompt. It will also generate payloads if prompted correctly.
18:25 ...and even if you _could_ somehow get the stuff you had an _"A.I."_ write for you past the detection software and examiners, you'd probably end up losing a whole lot more than the cost of your tuition fees when you inevitably end up getting sued for incompetence/malpractice. I mean, a large component of the reasoning behind the requirements for doctors and lawyers to have degrees, is to filter out people who are not high enough in conscientiousness, and therefore reduce the likelihood of producing _"professionals"_ who might be tempted to just drop clients and leave 'em hanging half way through a job, because the job is more difficult/demanding than they can be bothered dealing with.
9:37 Ohhh boy! This site! God damn I spent many hours doing those damn things with no knowledge of programming, and just tutorials, breakdowns, and forums. 😂
17:31 wouldn’t this mean that it’s more environmentally friendly too? Bc I’ve heard the other servers are being very unfriendly to the environment using a lot of water and making a lot of heat
Just wastes your own electricity rather than half a dozen other computer's electricity? Data centers are likely a LOT more efficient than your home computer for electricity and cooling (they kinda have to be to keep costs as low as possible)
Hey, this is the first time I am commenting on your videos, but I wanted to get a hold of you pretty quickly while it is fresh in my head. This comment is off topic about the video, but it is the most recent video you have posted, so I thought you might still be looking in the comments on this video. There is a video from a UA-camr by the name of Timeworks that had posted a video on November 17th 2024. It went over and shortened a hour long video from another UA-camr. The video was about allegedly exposing lies told by Dogpack 404. I know you have covered the Mr Beast situation before, and I just wanted you to take a look at it and try to do your own research to fact check it. I would do so myself, but I don't even know where to begin and I most likely don't have the resources to do so. I trust you the most when It comes to getting the facts and the undeniable truth behind anything. If you don't want to or can't look at it, that is fine by me. I mainly just wanted to at the very least bring it to your attention.
I use chat gpt to help take my ideas and rewrite them into a more concise and professional manner. I use it to make better sounding cover letters and resumes. Not sure how well it's working however considering I am still unemployed 😢
Thanks for giving it a try. Hope you got beyond the login issue. The "not lite" version is coming this summer. Please don't download anything from the Dark Web, the lite version will do it for you.
i used to have a good friend that sufferd from schizophrenia and he fell for a scam of girl trapped in Nigeria.. i cant even imagine how AI would effect such people today..
Dang.
It's gonna be bad, once it gets advanced enough there will be scams good enough to catch the most average of individuals, not just those so easily duped.
@@SiriusCygnusphishing emails have gotten really good if you’re not paying attention
@deathsaintxv yeah good thing I don't use email like that
Lost a friend of many years to schizophrenia.
This has the same vibe as downloading music with a ".exe" extension.
"thick-of-it.exe"
>"Run as administrator"
"Allow this application to make changes to your device?"
>"Yes" "No"
Nowadays it's .pdf, screensaver that steals your login infos
H.EXE "Never Enough"
@@REALWINWINDACATwould you like to run JinAndJuice;).exe?
Just change the .exe into .mp3 👍
They legit hired the "Your computer has virus" stereotype for their ad 💀
Lite version only has default skin available
Jesus loves yall
The AD looks like AI itself hahaha the mouth movement looks just off deep fake or something lol.
Okay listen people if you think a named called “HackerGPT” doesn’t have any backdoors your brain is fried 😅
For real
I asked it and it said its safe
Nobody that HackerGPT is marketed towards is capable of independent thought, let alone reading.
True and real
is that yellow guy from dhmis in your pfp
Mutahar: i love running dangerous software
Adobe: tell me more
I took a free trial on adobe few weeks ago and it wouldnt let me cancel it. Every time you try to enter the cancel trial page it says failed to load. After googling it looked like this issue exists for quite some time and you have to use your phone for it to open.
Hate them
Huh? Am I missing something about Adobe?
@@MysticBearKnight i guess 🏴☠
@@anomalousanimates Care to elaborate, please? I’m curious now.
@@MysticBearKnight piracy, but i am not entirely sure
"To become a hacker, you must first become hacked" -Sun Tzu, The Art of ChatGPT
My uncle with a stutter was recently sent to prison...
He's never going to finish his sentence.
This actually holds a lot of weight. Jedi's were once Script Kiddies, Yoda told me...
@@IJH-Music the validity of this comment makes me feel called out I'm crying in assembly now
"Sometimes you got to go back… to actually move forward, and I don’t mean going back to reminisce, or chase ghosts, I mean going back to see where you came from, where you've been, how you got here and see where you’re going. I know there are those that say you can’t go back. Yes, you can. You just have to look in the right place.” - Matthew Mcconaughey.
@@evillecaston study all the hacking you suffer and apply it to the enemy *brain expansion* - sun tzu
Brb gonna download HackerGPT
Edit: We've taken over...
Same
relatable
I need to know where to download it, so I can avoid downloading it.
Imagine not already having it😅
So how was it?
This video is going to have a Streisand effect Muta. You gotta know half the internet is bots and another 45% is.. A couple lumens short of a bright idea.
hahaha..... ah....
Omg dude ur so smart please tell me what to think next
And don't forget those bots are trained on those same peoples data 😂
True Shit right here
are you telling me stellar is going to the moon?
Maybe I should ask ChatGPT about that
Just ask HackerGPT
@@kara66340 lol
_"ChatGPT understands the legal system"_
-Mamamax being dumb af
I already did
ChatGPT is overrated.
I just downloaded MutaGPT, and now I just can't stop laughing violently in all rooms I enter.
i love how at 1:05 you can easily tell that the face is also AI work to hide his identity.
Oh god Indian deepfake tech homies, "delete system34 myfriend"
What if HackerGPT already downloaded me?
you didnt even have time to finish the video
You think that's air you're breathing?
oh crap were stuck in the matrix which isn't a bad thing 😅
no one can download the mad dog of shimano
@@MorrowindES17I mean it kind of is a bad thing because we don’t get all the cool Neo red pill hacker shit, we just get the blue pill Matrix npc experience🤣
HackerGPT sounds like the username of a guy that wants to sell you a 10,000 dollar AI, crypto, hacking, online course.
Anymore, it's easy enough to just run your own LLM locally. You can easily run models that are totally uncensored in that case, meaning they're willing to do whatever you tell them to. So, just do that instead of some scam. I don't understand people.
Whatz LLM?
@@h.Freeman 17:33
Tell me some good uncensored ones cuz they've all been bad so far
Local LLM sucks. Do you think company will release their model for you to run freely? Of course the good stuff is locked behind vaults, duh.
@hani if you think local LLM “sucks” that is a skill issue
Developer here, what these scripts primarily from what I seen in this video is back door implementation or api key harvesting. Since most of the people downloading these scripts or applications are mostly desperate, they will never bother to check the code and most likely is sending all of those api keys to a server the attacker is running or controlling your pc via c2 server and agent
Much less know how to read and check the code beforehand, Seen some deadly stuff sitting on github of all places even lasting today, stop trusting every 12 star repo with hacker rainbow text and you'll keep a few accounts people.
Dread forum is hilarious. There was a whole thread on there the other day of a guy complaining that the mailman stole the illicit substances he had ordered off the dark web to his house. 😂
lmao such ingenuity , if the package never arrives one should assume that police/border control got it, guy's on a list now and worrying about the mailman XD
Did he order like a felony amount?
@@tommyfanzfloppydisk Oh he posted an update! He contacted the "supplier" about not getting his order and they're sending him the package again. So the feds will get a round 2. 🤣
@@artificialbffi love you for giving us the update
@felixchallis8038 I forget which sub but it was just posted a couple days ago so its on the front page of the site or at least it was as of yesterday so its easy to find. Title was something like "Mailman stole my drugs" lmao
Never heard of it until your video. Had no intention of downloading it. Now i most definitely have to.
"If I blur names, is because I blur names" ..... Indeed
This floor is made of floor
8:20
In other news, if you die, you're dead...
GPT can actually literally write down full scripts that will compromise machines and rootkits that can only be eliminated if you fully restore the hardware to factory presets. You just have to ask it...the thing is, people don't know what to ask.
I asked it to write a program that would back up all the most important user files on my machine, and then ahem, asked to modify it
ChatGPT will tell you anything if you say its just for a homework project, or a study
What I'm hearing is, I should start printing business cards with some funky title like... "expert prompt engineering specialist" (with a tiny asterisk behind the title, and a load of fine print to cover my butt with more legalese than legalese ever legalesed).
With how much AI garbage buries real content in search nowadays, the regular internet is quickly becoming a non-indexed dark web
Omg hackergpt is hacking hackers!!!!!!!!!!! Who wouldve thunk something named "HackerGPT" wouldnt have backdoors
Honestly, if you trust something with Hacker in the name, it's just the internet equivalent of natural selection.
Thought*
@Kingklown_YTthunk is a funny way of saying thought, to denote sarcasm
This is more srius than trolls trolling trolls..
Thunk 🤣🤣🤣@Kingklown_YT
Dystopian dreams are becoming a sad reality, and luckily, we have Mutahar to keep us safe
"If I'm blurring out names, it's most likely because I'm blurring out names" - Mutahar 2025
0:27 "Sometimes I like running crazy scams on the Internet"
Subtle foreshadowing
This is what I call foreshadowing
I think by running he means like running a software
HackerGPT is secretly Copilot forcefully installed on your PC
ChatGPT user: yeah im an artist
HackerGPT use: yeah im hacker
yeah i got hacked*
ChatGPT: Yea i'm a writer/SEO specialist, etc.
Hacker got: yeah I'm a hacker.
midjourney: I'm a artist!
dude I can't believe how much you have to CYA, in making this stuff. thanks dude. stay up!
More virtual machine videos WE’RE BEGGING YOU MUTA PLEASEEE
Real
7:32 muta almost drowning while recording a video
😂😂
Lmaooo
Lmaoaoaoaoa
"helo, I'm under the water" ~Muta
Honestly, an AI that scans and try different attacks sounds like something devilish effective. Wouldn't expect to be so obvious, of course.
Not only dangerous, it might have some benefit for pen testing. In fact, I'm really looking forward testings in general. AI is cool when being used as a _second opinion_ rather than doing the job for us.
But of course, we can never have good things, can we
Don't worry, AI will sort that last thought out. 😢
@@scottp.548 Yeah, if it can give a "second opinion", it can be the first as well. Sadly.
There is stuff for this in the pen testing community alot of the networking people have already adapted ai to there toolset
i have an uncle who got scammed by a south east asian girl claiming she came to the USA but needed him to pay for a plane ticket. turns out she never took a flight and he was waiting at the airport for "her".
The second I saw “AI powered” paired with “Security assessments,” alarm bells started going off in my head.
0:11 I hate to waste one of my comments on this but you got to love it when there's a disclaimer at the beginning of one of Muta's videos. Now I can Probably Comment one more time before there is a problem. It's very strange
We are judging you. People on youtube are only allowed a maximum of 2 comments on a youtube video before they get silently judged.
0:39 that mocking tone just sounds like Homestar Runner lmao
Oh my god memory unlocked right there, the Cheat, Strongbad’s emails, Trogdor!
@@Hash1Suspect consummate Vs
Broooo I haven't seen that site mentioned by anyone in forever!! 😁
Remember CdC’s Back Orifice? Much fun was had. Thank you for gluing me in to another toy.
Another good thing about running an LLM on your computer is they usually don’t have censorship
I am so old, that I even got those scams as snail mail. No joke!
If you create a company, you will get a letter saying you have to do certain things or display certain posters, and how they can sell you some at $100 each so you’re in compliance. Unfortunately, the posters are free from the federal government and publicizing your company in the local newspaper is cheap. It’s _technically_ not a scam, just unscrupulous.
18:40 I have a friend that does this, but instead of chat gpt she pays people to do her work for her, she’s about to have a degree in criminal justice and her teachers don’t know what she looks like and doesn’t know what her handwriting looks like and she’s supposed to be a PI in the future. But someone like me who would go and enjoy and maximize my opportunities cannot afford or get funding to go to college
the bots are real tonight in the comment section good lord
Bots have taken over social platforms back in 2022 it ramped up with Ir-nian funded bots being extremely prolific and able to tap into tiktok/gen z circles.
Thank you for covering WormGPT!!! A topic I've been super curious on but didn't wanna spend $200 to find out the truth haha
Happy New Years Mutahar 👍😎😎👍
Muta just casually got the heater sitting on the desk.
6:53 "im sorry dave"
Bro you know it's bad when you have a disclaimer at the start....
...my roommate is gonna use this...and I'm gonna have to fix the problems he creates...
L roommate
@anxiety even the best roommate is an l
@anxiety even the best roommate is an l
14:23 The only issue with running LLM's locally is the simple fact that if you use ChatGPT, you get an LLM trained on their, "book 2" or whatever. Thats literally terabytes of copywrite and pay walled data that any locally run LLM can't use for context - mind you context is everything to an LLM.
100% a compliment, but dude, when you go off camera and straight into the mic like at 7:36 you sound just like Kermit The Frog 🐸.. It took me days to work out why you sounded so familiar.
Thanks for LM Studio tip. I've been running these through the command prompt which just feels unpolished and clanky.
Should LM Studio not quite be up your alley, there's a few other options too - I liked JAN AI, though at some point it just stopped working for me without any clear reason. Ollama sounds like it'd have been right up my alley as well, though I couldn't get it to properly install (but plenty other people have no issues with it at all, so that's an issue on my end). GPT4All is what I've been using for these last 2-3 weeks, works like a charm (though it does feel a bit basic at times). I used to have a list of roughly 40 UI's that I wanted to try out, but which ended up getting lost, and those were just the ones that I found in three or four rounds of websearches, so there's gotta be plenty more around.
Bros gonna have a VM under a VM under a VM all under a VM. He’s not getting hacked in a way it hurts him.
Pro tip, don't use blur for sensitive info, its easily reversible. I'm pretty sure that none of the things that were blurred were sensitive though, just there for UA-cam censorship purposes
Not all blurs are reversible. Specially the fact that when a video is uploaded to UA-cam the compression algorithm will merge layers , reduce pixels and other stuff.
I wish I could remember the site back in 1999 or 2000 that had a service where you could pay someone to write an exam paper for you. I can still see it in my minds eye. The layout. It was pretty basic as you might imagine. Just can't remember the site. I'm sure there was more than one, but this one was pretty popular. How far we've come 😵💫
Well, there's a flashback. My "golden IRC days". That site was getting heavily advertised/discussed in one of the DALnet rooms that I used to regularly hang out in back then, and yeah, there were a few others as well but they were basically considered poorly done rip-offs.
I'll have to check id I still have some old drive from that time with some IRC logs on it, probably not, but but it wouldn't be the first time that I managed to dig up some ancient stuff on an old harddrive.
@EvilFookaire man that would be sweet
13:24 as an instructor I’ve been able to make much more in depth lectures with GPT. I have a ton of ideas but usually get stuck in the concept phase. Using it to organize my thoughts has helped tremendously.
3:52 date has a russian formatting btw
Whoa. How lucky! You got 18 million dollars now! 😂
I'm more concerned LLMs stifling free expression. Under the guise of preventing the dissemination of misinformation, hate speech, or harmful content, the major LLMs are often censored or push one-sided narratives. For example, Gemini will often refuse to entertain political opinion questions.
Black George Washington
I honestly think that's comical more than concerning. Like I don't trust LLMs to give accurate information anyway, they should always be taken with a grain of salt due to hallucinations.
EDIT:
I remember hearing about one case where an LLM went into overly protective mode and wouldn't generate C++ code because it was "unsafe".
I can already see some 12 year old kid downloading this thinking he can type "hack Tommy's iPhone and steal all his pictures" and expecting the GPT to say "Done. They are all in this is google drive and there's a lot of embarrassing pictures".
thank you for always spreading important info we love your vids
his videos are terrible...
@@kr1me2000you sure watch a lot of them tho
the video is 20 minutes long, you posted this 24 minutes ago
you posted this 3 min after the vid came out
@couldntsetpass2901 he posts informational videos.. other than this one we can praise more than this while watching it ?
For chatgpt You can set the variable to Hex, ask it a question in Hex and tell it to output the answer in hex, then convert the answer back to your language and it will tell you how to do anything
"It's statistically more likely for you to win the lottery than it is to go undetected by AI detectors"
The best ones to date have ~20% accuracy. Worse than random guessing.
Plot twist - Muta accidentally ruined FBI's honeypot for easy to catch idiots
You shold watch the show Hack The Planet hosted by Razor and Blade, Remember, hacking is more than just a crime. It’s a survival trait!
i love how the Spanish Prisoner scam became the Nigerian Money scam.
Guys don't worry download it! He said it's a good app and that his title is a joke.
Source: Trust me bro
Fr🪑
🎉😂😊😊😊😊😊
"HackerGPT" sounds like a brand of software you'd find on the console of one of those movie hackers. 🤣
Those movie hackers are super experts. I mean, how else can you be using the same three or four commands to just instantly get into any system, regardless of whatever it's running?
9:46 "penetration testing"
Aside from all the either dark or hilariously cringeworthy stuff on the dark parts of the vast internet, also have your interesting stuff like experimental technology big tech doesn't want the average person hearing about or recreating. The internet is like the vast ocean, there might be fish with faces only a mother could love but there can also be jellyfish that are so bright and intriguing to look at.
15:59
I believe you didn't offload the layers to the GPU, or something like that. 8B LLMs usually work extremely fast on just 12GB 3060.
I was going to say. I’m running Llama 3.1 8b on a 3050 8GB and I’m getting about 35-36 tokens/s. A massive wall of text and code in about 20 seconds total including prompt and eval. Something isn’t right with his setup
@@Ubiquities he does everythnig on virtual machines so no wonder it's always slower than running on actual hardware, he even play video games on virtual machines if he would run on actual hardware he would get much better performance on everything...
Off loading layers designates something like 75% of GPU power to waiting for CPU to shuffle the layers in and out or VRAM. (Ať least on my systém.
@ you can pass through bare metal access to GPUs for the VM. I was doing the same thing using Proxmox and PCIE pass through (it has a specific name but I don’t remember off hand what it’s called specifically) I’m sure he is doing that for gaming, because he’d know that something wasn’t right. But what he showed on screen wasn’t configured correctly to use the GPU, it was way too slow for his setup.
As someone who has worked software security, I really don't recommend it long-term unless you have a family and community that can prevent you from losing all faith in the goodwill and trustworthiness of human beings. When you have to deal with scum day after day, and your immediate response to anything is thinking how some asshole will take advantage of it and how herds of lazy idiots will screw it up, it's too easy to get into an unhealthy mental space.
"pretty locally" well either its running locally or not...?
Vid was cool! For such a nice studio mic set up, the audio on the video is low. Turn your preamp up, boost output on compressor , or normalize the audio before going into UA-cam.
Downloading AI slop hacking shit is bad? Fork found in a kitchen drawer lmaoooo
air found in a car tire
phone found in a pocket
Sand found in a desert
Rocks found in a cave
The list could go in.
This is wild. I could have sworn there is an actual Hacker-focused LLM you can run locally to assist with penetration testing and I thought this was a video on that exact tool. This however is definitely not what I was thinking of LMAO
You just don’t want us to hack you 😈
Bro advertising these "evil AI bots you can harness!" Makes me think or those old scam ads in comic books for evil martial arts styles 😂
@SomeOrdinaryGamers What operating system do you use?
He uses Arch Linux. He mentions it about every 3rd video.
Don't ask that. He's going to reply with an entire paragraph on how to install Linux from scratch 😂
Bro thinks he's pinging Mutahar 💀💀💀
bro i love dolphin! i run it on anythingLLM, or just in a terminal if its something quick!! great video u got a new sub (not like u need it u alr have 3.5 mil xD where has this channel been all my life?)
I always think if you really want to have an uncensored AI model just grab the 2k it will cost to train a base model yourself and there you go. Even if you have some good hardware laying around you can run some pretty good pretrained models
6:42 ChatGPT will most definitely answer questions about hacking if you word the questions correctly or properly load the pre-prompt. It will also generate payloads if prompted correctly.
Muta said "actual" 25 times in this video
Actually…
Only 25? He fell off
@@thefrofox Honestly tho, usually he'd be hitting ~40 actuallys in a 20-min video
Actually try
Stuff like this is why I've been working towards a Cybersecurity career.
Mutahar, are you gonna talk about the new GTA 6 leak?
I am glad you do this kind of work my friend. Love this kind of content of yours same with gaming
7:16 *_PAUSE_* 🤨
18:25 ...and even if you _could_ somehow get the stuff you had an _"A.I."_ write for you past the detection software and examiners, you'd probably end up losing a whole lot more than the cost of your tuition fees when you inevitably end up getting sued for incompetence/malpractice. I mean, a large component of the reasoning behind the requirements for doctors and lawyers to have degrees, is to filter out people who are not high enough in conscientiousness, and therefore reduce the likelihood of producing _"professionals"_ who might be tempted to just drop clients and leave 'em hanging half way through a job, because the job is more difficult/demanding than they can be bothered dealing with.
Im calling it, everybodys going to turn to the dark web to get actual search results
Omg, bro finally fixed his lights
im waiting for Muta GPT to be released
It's called Moist. Usually trained in Mutas research and regurgitated in clone form
And if you think I'm wrong, I guarantee he will do a video on this topic in the next five hours.
9:37 Ohhh boy! This site!
God damn I spent many hours doing those damn things with no knowledge of programming, and just tutorials, breakdowns, and forums. 😂
Is it different from platforms like tryhackme or hack the box
0:25 reported…
Why would Mutahar run said crazy scams under the alias Samulta Hyder
17:31 wouldn’t this mean that it’s more environmentally friendly too? Bc I’ve heard the other servers are being very unfriendly to the environment using a lot of water and making a lot of heat
Just wastes your own electricity rather than half a dozen other computer's electricity? Data centers are likely a LOT more efficient than your home computer for electricity and cooling (they kinda have to be to keep costs as low as possible)
I farted everywhere
>redeem
jesus thats what i smell
smells like crap
Linux
teh fartz
Can we get this to top comment
ChatGPT helped me make a spinning cube in VR, as a Free WiFi Login Page. Using a Tamagotchi as the host.
Hey, this is the first time I am commenting on your videos, but I wanted to get a hold of you pretty quickly while it is fresh in my head. This comment is off topic about the video, but it is the most recent video you have posted, so I thought you might still be looking in the comments on this video. There is a video from a UA-camr by the name of Timeworks that had posted a video on November 17th 2024. It went over and shortened a hour long video from another UA-camr. The video was about allegedly exposing lies told by Dogpack 404. I know you have covered the Mr Beast situation before, and I just wanted you to take a look at it and try to do your own research to fact check it. I would do so myself, but I don't even know where to begin and I most likely don't have the resources to do so. I trust you the most when It comes to getting the facts and the undeniable truth behind anything. If you don't want to or can't look at it, that is fine by me. I mainly just wanted to at the very least bring it to your attention.
I adhere to the principle: "avoid any place that already has a predetermined value for my existence." The potential irony does not escape me.
0 views in 12 seconds. Muta fell off
True
I use chat gpt to help take my ideas and rewrite them into a more concise and professional manner. I use it to make better sounding cover letters and resumes. Not sure how well it's working however considering I am still unemployed 😢
It would take a basic one day course on basic web security, to train top executives how to avoid basic scams.
Muta has the best pen testing UA-cam channel 😂
Thanks Muta 😅
HackerGPT sounds like a very sketchy thing you'd find in the darkweb
Thanks for giving it a try. Hope you got beyond the login issue. The "not lite" version is coming this summer. Please don't download anything from the Dark Web, the lite version will do it for you.