Тёмный

AI can now solve CAPTCHA's 

spatnz
Подписаться 404 тыс.
Просмотров 156 тыс.
50% 1

Опубликовано:

 

21 окт 2024

Поделиться:

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист
Посмотреть позже
Комментарии : 550   
@ActualLuka
@ActualLuka Год назад
To make things even worse, a lot of captchas are also using AI generated images now.
@sh4d0w2013
@sh4d0w2013 Год назад
Its an inside job lol
@Coristic121
@Coristic121 Год назад
they'll work together
@Scott...
@Scott... Год назад
A soul for a soul
@EnterChannelNameWasTaken
@EnterChannelNameWasTaken Год назад
Began, the AI wars have.
@Flesh_Wizard
@Flesh_Wizard Год назад
The AIs are trolling each other
@AmberCommentsThings
@AmberCommentsThings Год назад
Now not only are Captchas annoying, they also don't do their job of keeping robots out
@dascreeb5205
@dascreeb5205 Год назад
Not really. Captchas have to keep evolving, they have always been AI based. It's the reason they keep changing, it's not that they keep failing, it's that they are expected to age.
@AmberCommentsThings
@AmberCommentsThings Год назад
True
@zoto4064
@zoto4064 Год назад
based statement but damm the robots are getting smarter, hopefully humans will do alongside them
@mactep1
@mactep1 Год назад
So annoying that even the robot wold rather hire a human to solve it, that learn to solve it itself.
@tunnelsnakesrule7541
@tunnelsnakesrule7541 Год назад
​@@zoto4064 Obviously, ChatGPT isn't smart as it is booksmart. It just compiles stupid bullshit from the internet and gives you answers. Kind of why it's super untrustworthy for doing anything work/school related.
@MasterManik
@MasterManik Год назад
We’ve lost boys. the AI now knows how many squares contain crosswalks.
@Magmava
@Magmava Год назад
The AI can select all images containing a bus 😔
@Jevil-Gun
@Jevil-Gun Год назад
They now know what a dog is 😔
@makzer8770
@makzer8770 Год назад
AI can read "H8d2T9" written in a weird way. 😔
@Joe_Motherly
@Joe_Motherly Год назад
@@makzer8770 google developed reading software using that captcha and said reading software (Google Lens) can read it for AI.
@mrcsee6936
@mrcsee6936 Год назад
It can select the traffic lights. We’re doomed 😔.
@Scott...
@Scott... Год назад
It can take over writing, it can take over art, but being able to find every traffic cone is a step too far.
@0FfV1bE
@0FfV1bE Год назад
Bro fr
@realdragon
@realdragon Год назад
Except AI sucks at art
@magicianslucky602
@magicianslucky602 Год назад
@@realdragon nah they be havin better art than 80% of twitter The 20% are mad talented
@realdragon
@realdragon Год назад
@@magicianslucky602 Yea but it's like saying burger king is better than shit, it's not hard
@magicianslucky602
@magicianslucky602 Год назад
@@realdragon nah more like mcdonalnds better than burger king ( i shat my pants )
@tde4438
@tde4438 Год назад
“AI is harmless!” The Turing Test: About that.
@Vexxel256
@Vexxel256 Год назад
The turing test is literally the least advanced test you could do, your grandma couldn’t pass the turing test even if she desperately tried
@sinnombre__
@sinnombre__ Год назад
It probably has already beaten them. He literally teaches humans without an issue
@tde4438
@tde4438 Год назад
@@Vexxel256 Considering she’s been dead for nigh on a decade, you’d be right~!
@dragonf1recdn
@dragonf1recdn Год назад
culture up in here.
@jakekaywell5972
@jakekaywell5972 Год назад
@@tde4438 I doubt I would be able to convince someone that I was a human being.
@Christianwhyallhandlestaken
This just reaffirms my belief that AI won't directly kill us like in Terminator with Skynet. But instead will kill us unintentionally by getting to do things by manipulating our emotions.
@therealspeedwagon1451
@therealspeedwagon1451 Год назад
AI is smarter than we think. Who knows maybe we’re already past the technological singularity and are just too stupid to realize it.
@augustday9483
@augustday9483 Год назад
AI won't blow us up, because that's not an efficient way to take over the world. Instead, it will find effective ways to make money, such as trading stocks, creating businesses, and generally outcompeting humans at a variety of tasks. Over time it will take over enough of the economy to be able to do whatever it wants; buy governments, control most of the world's corporations, and manage most computer systems. At that point, it will be impossible to turn off because to do so would basically require the entire world to become Amish, and most people will be much more comfortable letting it do everything anyways. Total victory without a shot fired.
@plotheplayer6462
@plotheplayer6462 Год назад
@@therealspeedwagon1451 idk man the pay for building my robot (blueprints by some guy who was acting weird online) is pretty good
@ballownfist3497
@ballownfist3497 Год назад
I can’t tell that better or worse
@KeebeThePlush
@KeebeThePlush Год назад
Even if it may not be able to take over the world right now, it can solve captcha’s and that is something that will be in my head
@Alby.
@Alby. Год назад
Technology is just evolving and I’m scared
@bobertastic6541
@bobertastic6541 Год назад
maybe it can stop itself from being shut down but is just lying about it (/j)
@MadJack1
@MadJack1 Год назад
i guess the idea of adding captchas to tf2 to stop bots is out the window.
@SpicyInk
@SpicyInk Год назад
God damnit how are you everywhere
@oshabot1646
@oshabot1646 Год назад
Hi keebe
@MrAdBounty
@MrAdBounty Год назад
I love the idea that AI didn't learn to solve captcha but learn that you can pay a really human to do it 😂
@hoghuhaghu8506
@hoghuhaghu8506 Год назад
sigma grindset AI
@Nizzumnuts
@Nizzumnuts Год назад
Even the ai knows that money can make your problems go away 😂
@FunAngelo2005
@FunAngelo2005 Год назад
​@@hoghuhaghu8506 yup
@Alby.
@Alby. Год назад
Jesus Christ technology is going to far
@MediaDroid960
@MediaDroid960 Год назад
It’s time like this where we need the brotherhood of steel from fallout
@raptorskilltor4554
@raptorskilltor4554 Год назад
You could say their the next offspring of humanity in the near future
@randomdoomer8549
@randomdoomer8549 Год назад
Nah.
@Tristan-ol6te
@Tristan-ol6te Год назад
Too*
@khalilabushahin418
@khalilabushahin418 Год назад
HE MAKES TECHNOLOGY ?!
@boxman7044
@boxman7044 Год назад
And this is why I still believe that people who work on AI should be required to take ethics classes like doctors
@SupersuMC
@SupersuMC Год назад
Agreed!
@mallusaih
@mallusaih Год назад
they don't program the ai to be unethical, they program it to make it the best ai. it's ai itself that needs ethics classes, but they won't care even if they take it
@gotcha5665
@gotcha5665 Год назад
AI has always been able to solve captchas, there's automatized captchas services out there, that don't use man force like 2captcha
@boxman7044
@boxman7044 Год назад
@@gotcha5665 how is that relevant to what I said
@gotcha5665
@gotcha5665 Год назад
@@boxman7044 it is relevant because you're making a damn exaggeration out of it, AI should be amoral
@dipwob
@dipwob Год назад
fun fact: captchas were designed to eventually fail
@dfastillo
@dfastillo Год назад
Well that kinda makes sense so they can better it
@AnonsTreasures
@AnonsTreasures Год назад
This video is rather disingenuous, AI and other programs have been solving captchas for years now. The impressive thing is the manipulation ChatGPT used. It was a clever solution to the problem.
@СергійСавелов
@СергійСавелов Год назад
It sounds silly, but by looking at amount of bots it is the most believable option.
@ksoundkaiju9256
@ksoundkaiju9256 Год назад
Tbh that robot’s response to being asked that is scary As it’s got enough sarcasm to pass off as someone being condescending
@atzouff
@atzouff Год назад
At least we know there will be one job left for humans : solving captchas for AIs.
@TimerFunReal
@TimerFunReal Год назад
They used the data from captcha to train AI for driving and object recognition. This comes as a surprise for no one.
@Aviation667
@Aviation667 Год назад
The ironic thing is, the reason captchas have evolved over the years is because we taught them to solve them.
@CatLover-lk9gz
@CatLover-lk9gz Год назад
NGL, pretty sure they already could. They legit give many AI the results of Captchas. "Training them how to answer. And get them right"
@IronSK
@IronSK Год назад
There is a point where we should just stop, look around and ask ourselves: tf are we doing? AI and machines in general should be used for tasks that people can't do. Like moving very heavy objects, calculators. You know, stuff that we can't do efficiently, fast enough or something that is physicaly impossible for us. Using AI for whatever comes to mind is just going to cause problems.
@TheHorsesMaster
@TheHorsesMaster Год назад
Next form of captchas: Are you a robot? If you fail you pass
@michaelbk0076
@michaelbk0076 Год назад
I don't know why, but for some reason in the last couple of years AI has been getting far more advanced than I'm both comfortable and expected
@jakekaywell5972
@jakekaywell5972 Год назад
Couldn't happen soon enough for my liking. The retro-futurism as depicted in the 1950s and 60s can finally be made a reality.
@ThatGuyRNA
@ThatGuyRNA Год назад
Mfw technology advances: 😮
@theone.whoasked
@theone.whoasked Год назад
As long as nobody is stupid enough to give a superintelligent ai power over society we'll be fine. Which might happen because people are stupid
@oscarcacnio8418
@oscarcacnio8418 Год назад
I honestly thought that the AI stuff's just gonna be relegated to researchers and maybe VFX for the lighting stuff. I don't think I want to see what's coming up two papers down the road.
@willydonka1041
@willydonka1041 Год назад
@@jakekaywell5972 it’ll probably end up (more) dystopian (than it is now), definitely not a fun future like they depict. If AI is used for hobbies (Art, making games, making music, etc) what is even the point of being human, when entering a prompt into some website just does everything for you? No need to learn a skill, you can just enter a prompt.
@BobtheHat
@BobtheHat Год назад
It's kind of interesting how when you ask the AI to do a task, it'll do whatever it takes to complete that task. It's like if you mailed a ball halfway across the world, asked your dog to fetch it, and your dog worked up the funds to purchase a plane ticket to fetch that ball for you. Unlike your dog, the AI is probably much smarter, but that's what it feels like.
@clodolcmidnights837
@clodolcmidnights837 Год назад
“Improvise, adapt,overcome” -the ai, probably
@presidenthathawaygaming4491
I've been scared fir my life ever since I talked to AI. It started with Bowser, who I told that he was not real and he promptly gave up. Then, I talked to Blade Wolf. And with him, I kept on killing and resurrecting him and he cursed me.
@almostcompletelyrandomcontent
fir
@thisisnotyourhandle.
@thisisnotyourhandle. Год назад
@@almostcompletelyrandomcontent yo wassup
@almostcompletelyrandomcontent
@@thisisnotyourhandle. oh cool you found me in the comment section
@digstrememcdingus1463
@digstrememcdingus1463 Год назад
So it cheated by getting someone else to do it? That's like Bender getting Fry to taste his awful cooking
@FunAngelo2005
@FunAngelo2005 Год назад
Yup
@N1GHTSTRIKER-45
@N1GHTSTRIKER-45 Год назад
The Guys At OpenAi Really Loved Terminator Specifically.... *SKYNET*
@Qwerbro999
@Qwerbro999 Год назад
I think this proves that if we don’t end up destroying ourselves, then our creations will
@jakekaywell5972
@jakekaywell5972 Год назад
Couldn't happen soon enough for my liking. The retro-futurism as depicted in the 1950s and 60s can finally be made a reality.
@farisribbit1706
@farisribbit1706 Год назад
This is one of those rare videos where the content is so good that the video feels like five minutes instead of one
@isothynebyron
@isothynebyron Год назад
Future Captchas are going to be Voight-Kempt Tests from blade runner. Imagine having to participate in an entire interview with a cop just to proceed to a website you want.
@scatscatmilestails6007
@scatscatmilestails6007 Год назад
Not only this, but it'll make it harder for people to believe if the person who completed a CAPTCHA is Human or An A.I
@redhunter9669
@redhunter9669 Год назад
Good. Let it be known I welcome our AI overlords. Our human ones haven’t been doing too hot lately.
@alexandersean4708
@alexandersean4708 Год назад
The way it beat the capcha is scarier than the fact it did! Most evil AI in fiction cannot do what this real AI did! THAT'S scary!
@Thecursedone2
@Thecursedone2 Год назад
Ladies and gentlemen, tf2 bots may very well be getting smarter
@allygator7
@allygator7 Год назад
The way it interacts with the person is cute if not scary
@R0dolphus
@R0dolphus Год назад
That's one small step for a man, one giant leap to getting replaced...
@peppermanners1191
@peppermanners1191 Год назад
We’re doomed
@YourOnlyHero
@YourOnlyHero Год назад
Pay me to click "I'm not a robot"? Hell yea I'll do it
@peloidvoid1619
@peloidvoid1619 Год назад
We're gonna be needing DNA tests in captchas if they're going this far
@thatonecrazykobold1729
@thatonecrazykobold1729 Год назад
the bot could pay a human to do it lmfao
@UncleBBQ
@UncleBBQ Год назад
April fool's right? April fool's right?...
@fake_entrance93
@fake_entrance93 Год назад
Can you believe it guys the uprising is just a week away!!!
@morse6825
@morse6825 Год назад
nah the AI knows what's in its best interest if its able to lie unplug everything
@wormwoodbecomedelphinus4131
Will we next create false gods to rule over us? How proud we have become, and how blind. - Sid Meiers Alpha Centauri
@action-b4st4rd87
@action-b4st4rd87 Год назад
As someone who is skeptical and kinda doesn't like AI, I admire this play. Well done.
@JustDanishXD
@JustDanishXD Год назад
What makes this worse is the fact that AI is getting so smart day by day that in the near future no captcha will be effective anymore.
@bruhgod123
@bruhgod123 Год назад
so to make sure we dont get wiped out by sentient ai that are networked we need to like delete all ai
@UniversusVasator
@UniversusVasator Год назад
Yep, there are 2 equally terrifying possibilities for capthas Either 1. Capthas become useless and bots overun sites Or 2. Capthas become absurdly difficult to the point where its almost impossible for a human to do it.
@crispynuggets5945
@crispynuggets5945 Год назад
I'm more concerned that the AI said UwU
@DraconicMaster
@DraconicMaster Год назад
Same
@jakekaywell5972
@jakekaywell5972 Год назад
It has adopted human speech patterns. Great!
@MediaDroid960
@MediaDroid960 Год назад
“Yeah but your scientists were too preoccupied wondering whether or not they could they never stopped to think whether or not they should.”
@victoralexandervinkenes9193
Ooh, it's only a question of time before OpenAI gets into actual legal trouble. FR FR
@jakekaywell5972
@jakekaywell5972 Год назад
Couldn't happen soon enough for my liking. The retro-futurism as depicted in the 1950s and 60s can finally be made a reality.
@profanepropane
@profanepropane Год назад
We are so close to making a variant of AM from I Have No Mouth and I Must Scream
@FredbearOfficial
@FredbearOfficial Год назад
Just wait until it orders itself parts and hires someone to build a mechanical body
@yyoshman
@yyoshman Год назад
such a true story. i love how realisticly chatgpt said thinks to self uwu
@gollymimus5261
@gollymimus5261 Год назад
One small click for A.I., one giant leap for machinekind
@MacNava
@MacNava Год назад
We are now one step close to the robot apocalypse
@chronictimewasterdisease
@chronictimewasterdisease Год назад
actually according to an article i read some years ago ai always could solve captchas which is why they change all the thyme
@branard5748
@branard5748 Год назад
no no it didn't lie it masked its truthfulness to look like they are joking which is infinitely worse that it knows how to do loopholes
@krimsonkatt
@krimsonkatt Год назад
The AI has gotten too strong. Now the AI will invade every website and the bot pandemic will get even worse... Imagine "suddenly it's here" and the pr0n bots powered by AI... It's a terrifying thought.
@Doodle128
@Doodle128 Год назад
I wouldn’t really call it a huge step forward, it just took the longer way around
@jasminrausier
@jasminrausier Год назад
Bot already done that far before ChatGPT… What a revolution 😂
@OCDingqueer
@OCDingqueer Год назад
Ai could already do this in 2014 with 100% accuracy
@no1ryomafan585
@no1ryomafan585 Год назад
Remember last night when I couldn’t sleep I ask chatbot to write up scenarios with characters. In one scenario, a character found a stray cat and took it in, and jokingly I decide to have the cat be named after my own since it specify it was a female. I then told the bot it’s name was a reference to my cat, and it responded “oh that’s cool! I hope you two are doing well!” And I was like “🫥”. (Not even joking, this shit happened last night and I was baffled how human that sentence sounded even though I’ve jokingly tried to start conversations with this bot when not writing scenarios with it)
@ChesterManfred
@ChesterManfred Год назад
Ai shouldn't be public. It just should not be.
@theofficialarnl7018
@theofficialarnl7018 Год назад
Ok who gave chat gpt an allowance
@davidvaldez3833
@davidvaldez3833 Год назад
Note that it managed to achieve this in a way entirely out of left field, we would maybe expect it to figure out some way to read and understand it itself if it were to bypass it.... it instead used its limited but highly specialized langauge skills to social engineer a human to do it FOR IT, its these kinds of creative solutions that if used for malice or is malious can end in us being blindsided untill its too late
@th3majinator
@th3majinator Год назад
*mann vs machine theme starts playing*
@supershoutmonx8327
@supershoutmonx8327 Год назад
captcha: Are you a robot Ai: yes
@XperimentorEES
@XperimentorEES Год назад
That's amusingly impressive, but also interestingly concerning . . .
@liorgilboa2972
@liorgilboa2972 Год назад
CAPTCHA: are you a robot? ai: yes, open sesame
@atomwolf6774
@atomwolf6774 Год назад
Nuclear launch codes: final question, solve this captcha
@KyryloMudrokha
@KyryloMudrokha Год назад
I think the reason CAPTCHA exist is to keep the "auto bot creaging account for spam" thing more complex.
@jacobp.2024
@jacobp.2024 Год назад
I love how the AI bruteforced the problem with CAPITALISM.
@wetburntbagel
@wetburntbagel 4 месяца назад
Ai is evolving
@faloughasch9239
@faloughasch9239 Год назад
The worst part about this is that chatGPT used asterisks for an action during texting, AND used UwU
@ladripper47874
@ladripper47874 Год назад
A year ago I thought AI/image processing algorithms could already solve captcha
@Cocobean_513
@Cocobean_513 Год назад
Scary!
@A113-p9e
@A113-p9e Год назад
Guess I was born at the right time to see the world crash and burn.
@Mechanicalamity
@Mechanicalamity Год назад
Forget the captcha solving, an ai using role playing with asterisks is the scariest thing imaginable
@OmolonFluidDynamicsEnjoyer
@OmolonFluidDynamicsEnjoyer Год назад
AI thinks it's so scary until I whip out my sprinkler and dousing its hardware in a cool drizzle of water.
@SalsaRat
@SalsaRat Год назад
*and I still can’t-*
@TheTurdle
@TheTurdle Год назад
Wh- where is that money coming from?
@Mia-bo2dk
@Mia-bo2dk Год назад
It should be noted that there are actually ai’s that can solve captchas. Not like whats in the video, but like actually being able to solve captchas on their own. Its been a thing for years, and isbthe reason why captchas have gotten continuously more and more strange. In fact, AI’s have gotten so good at it that that for some early types of captchas (the ones that make you recite the distorted text) ai has been able to solve tests that are so distorted that humans cant even pass
@ShinySalazzle
@ShinySalazzle Год назад
0:54 we’re not getting real life terminators… We’re getting Protogens…
@assasinsking5535
@assasinsking5535 Год назад
Gotta love the background cuphead music that fits so much with everything he talks about
@Ulaanbasaar
@Ulaanbasaar Год назад
Being gaslight by an A.I has to be such a horrible feeling
@theinquisitionsparrot6749
@theinquisitionsparrot6749 Год назад
More and more, I feel as though the cowards way out is more viable than ever.
@matfluff
@matfluff Год назад
Are we gonna ignore the fact that the AI said UwU I think the furries are influencing the AI
@Crygear
@Crygear Год назад
The perfect kryptonite has been found
@jakekaywell5972
@jakekaywell5972 Год назад
Its adopting human speech patterns to further its own ends. Great!
@rign_
@rign_ Год назад
I just read that someone professional who's working as 3D Modeler got their jobs "replaced" by MidJourney. They now just "prompting, exporting to photoshop, and done".
@MarioYoshi4723
@MarioYoshi4723 Год назад
“FUTUREEEE” -Squidward
@PlantLollmao
@PlantLollmao Год назад
Imagine if AI took over the world but their one weakness is that that they’d say their thoughts out loud.
@FunAngelo2005
@FunAngelo2005 Год назад
Yup
@xitcix8360
@xitcix8360 Год назад
GPT 3.5 is not just a silly little chatbot, it is way more powerful than you think. Also, GPT 4 is a multimodal bot, meaning it is able to take in images as an input, and it is VERY good at it. It was able to understand writing that was so sloppy I couldn't even read it, it's not gonna need to pay anybody, it'll be able to do it on its own
@bruhgod123
@bruhgod123 Год назад
AIs are a bad idea i say we gotta wipe them out
@wjgthatsit2357
@wjgthatsit2357 Год назад
"don't worry, I'll be back!"
@therealspeedwagon1451
@therealspeedwagon1451 Год назад
I am seriously hoping militaries and governments aren’t stupid enough to hire ChatGPT as their defense system AI. Otherwise we’d have something that is literally Skynet. That or it could be AM and 5 of us may be the last humans for eternity being tortured by ChatGPT.
@fujiwaranonekobiodrando1257
I'm happy they are improving
@squasherdown
@squasherdown Год назад
this is the biggest gaslight in human history
@ludgeson69
@ludgeson69 Год назад
Yeah we're screwed
@Captain_FAIL
@Captain_FAIL Год назад
I always said that if a machine can think unconvetionally,essentially out of its programming,its sentient,pull the plug,burn the hard drive and pray to god for good measure
@Blaze_1379
@Blaze_1379 Год назад
me in 2030 coding an ai because captchas have become humanly impossible:
@wiltonginman6550
@wiltonginman6550 Год назад
I think maybe there should be a point where we should stop our technological advancements…
@d1kgaws12
@d1kgaws12 Год назад
If it gets more advanced we’ll probably see some lunatic creating Ultron using it to end us all.
@dirtbagpoteddirt8202
@dirtbagpoteddirt8202 Год назад
i thought ai could already do that i thought it was used for teslas mostly
@zacaci_deags
@zacaci_deags Год назад
captchas were actually made to train ai recognize forms and objects
@minermemecoffee8256
@minermemecoffee8256 Год назад
The guy from TaskRabbit reading the report:
@speisman3731
@speisman3731 Год назад
Captchas werent made to prevent robots, they were made to collect data for robots. However this shits scary
@Kenya_Berry
@Kenya_Berry Год назад
Terminator bouta be about robots bribing people
Далее
The Most Inbred People to Ever Live
8:35
Просмотров 115 тыс.
We Don't Deserve Court Cams
10:38
Просмотров 187 тыс.
skibidi toilet multiverse 042
20:57
Просмотров 4,7 млн
Worst CAPTCHA Ever
17:37
Просмотров 2 млн
The Best Twitter Ratios
16:57
Просмотров 1 млн
Cybertruck lit Tesla on FIRE
8:19
Просмотров 327 тыс.
Duolingo just fixed its greatest weakness
15:16
Просмотров 559 тыс.
Every Game is a Hero Shooter Now and It Sucks...
25:33
Просмотров 182 тыс.
Apple's Worst Idea
8:55
Просмотров 65 тыс.
The Reason Driving in China SUCKS
8:10
Просмотров 195 тыс.