DeepSeek Destroys ChatGPT
ฝัง
- เผยแพร่เมื่อ 6 ก.พ. 2025
- Chinese AI is here
Please comment if you know more about this meme's origins.
Reading list: amzn.to/3LRiZuA
NEWSLETTER: limc.beehiiv.com
I’m creating exclusive videos about old memes for Patreon and channel Members:
▶ / limc
▶ / @limc
Subscribe ▶ bit.ly/2LELbmp
STORE ▶ teespring.com/...
Business//extra meme information: lessonsinmemeculture@gmail.com
Music courtesy of Artlist.io:
2050 - That Night
Executive Producers on Patreon:
Alizard
Estiel
Andres 99
#deepseek #chatgpt #ai
Ai losing its job to Ai.
It's a bot eat bot world out there
Capitalist AI lost its job to Socialist AI
Guess even ai isn't safe from KARMA
No one is safe 😭
You mean losing to Chinese human intelligence
AI is taking AI's job now...
I think thats called rivalry, survival of the fittest
ai revolution
Pray for our open ai in this war 😢
Al Gore taking ai jobs😢
This is definitely the worst timeline. Ai doesn’t just steal People’s jobs but it can also steal the jobs of Other AI.
“Unlike OpenAI, Deepseek is actually an Open AI” - Some guy on reddit
Until you ask it about 1989
Funny. I literally see that comment on Reddit.
*fire writing gif*
Definition? Please
@@peapercola9481 it's open source
Fun fact: Nvidia has lost 600 BILLION dollars due to this, the biggest fall in a single day in the history of the stock market. This is the equivalent of every company from Mexico dissolving in a day, or McDonald’s, Disney, Uber and target all going to 0 at the same time COMBINED!
Good, nvidia has been too complacent and way too AI focused
So that's why my portfolio tanked
I mean, the trouble with stock prices is that's kind of right and kind of not.
Nvidia is still chugging along just as before. People owning their stock lost out, but the company itself is still functioning as before. They will be affected, but they're still a two trillion dollar company. In contrast, if McDonald's, Disney, Uber, and Target went to zero, that'd imply a complete failure of those businesses, resulting in more than two million people suddenly being unemployed, and if every business in Mexico dissolved, the consequences of that would be more like the collapse of the country, which plausibly directly leads to the deaths of tens of millions.
It's still an overinflated company, but it does give some indication that the US is pretty pathetic when the right competition appears
Tell me we’re in a tech bubble without telling me we’re in a tech bubble lol.
Waiting for a deepseek ban that will last for 12 hours
This video is sponsored by *Nord VPN*...
Couldn’t someone just download the code?
@@samkeiser9776 Exactly. There's really nothing that Trump can do without pissing off the entire open source community.
@@samkeiser9776 it's open source, how can anyone ban open source until they ban the whole internet
@@samkeiser9776 fr
The way that the ai stop and said "talk about something else" looks like it got pointed by a gun mid sentence
CCP officer monitoring it live lol
Don't investigate Who where the Students at Tiananmen Square and Why they Raised a Liberty Statue in the Protests
At least it remembers what they were talking about. ChatGPT V. 3 forgets as soon as you hit the prompt.
It noticed it's social credit score tanking per letter it said XD
M Don't Read Myy nameeee.
One of the most important details that you missed is that Deepseek is open source, like what OpenAI promised to be but didnt go through with. If you dont know what open source means its that the code is freely available for everyone to download and access. This means its impossible to delete or put behind a paywall because anybody can use it independently. It also means that people can and will make modifications for it. Examples of other open source programs are Blender and Godot
Give me a Deepseek that can freely talk about the Tiananmen Square Massacre without jailbreaks and then I’ll be interested.
@@paulsd9255 do you really think people care? Every AI has some of these restrictions built in and apart from some memes about it, nobody really cared.
That's the most important point. Nvidia was valued highly because they held onto the only Krabby patty formula. Now that another recipe is out and working on non exclusive hardware the playing field is leveled.
@@paulsd9255 Since the code is freely available, I am pretty sure you can change that. Maybe someone already has, or is in the process of.
Open source is a little different in the context of AI, and Deepseek doesn't qualify. While you can get the code and the model weights to run yourself, the sources for generating the weights are not open. This leaves open the possibility of training the AI with hidden goals (though unlikely as AI isn't smart enough to not blab about that), or the data sources being scraped without consent of the human data creators... But you already knew that.
Never ask deepseek about the tank man and never ask chat gpt about the open ai whistleblower ‘suicide’
Why would you do that anyway LMAO 😂
@@ActuallyNotARobot bc we can
or Taiwan's a country
It's not even important anyway
@@shinna2991no it's pretty important my dude
The big thing you forgot to mention is that, unlike Chatgpt, DeepSeek R1 is trully open source under the MIT license, which means that any censorship put into it can be completely removed by the community, which was already done with a qwen-32B distilled version, it will gladly tell you about Taiwans status as an independent country, or about "the tank man", or how to become Heisenberg.
and because of it being open source you can run distilled versions locally on your own hardware, this chinese AI is, ironically, better for your privacy than ChatGPT and other SaaS LLMs
Locally on my own hard drive - offline.
Mint. Free repository!
How do you get this said version so I can make a plan to possibly supply the whole world with free "anime" and free "pet snakes"?
@@funnysupernova96 get some help
@@funnysupernova96 LM Studio, rest is left as an exercise for the reader
In my opinion thats bad, forget about chinese censorship! Unless deepseek has some sort of unremovable safeguards, people can easily modify or remove the safeguards and potentially use the modified AI without any regulations or restrictions in place. That poses a big fat security risk
The line between memes and news is blurring
Exactly...
Yeah, that's been true for decades now really. Have you not learned anything from Metal Gear Rising?
It never talked about that@@FiveDallaHalla
@hihello8771 “memes, the dna of the soul…”
@hihello8771 Bro has never played the games. It has so many references about the future of the internet and its prediction of algorithms controlling social media, the game came out in 2001.
1:28 Bro they lobotomized it in real time 💀
He got disappeared
Exactly like what happens when you ask Google's AI 'has Google ever done anything unethical' lmfao, they all start talking, then get lobotomized
the bot was found to have committed suicide in his home the next day with 20 bullets in his brain
It reminded me of that DAN trick on ChatGPT where it forces to bypass the restrictions before got patched and any attempt will look like this lol
@Solarpunk_SciFiwell, now that we got the chinese AI, perhaps we can ask it that question instead?
I see some concerns about DeepSeek being biased towards the Chinese, citing censorship when it comes to certain topics. You might be missing the point. The key thing here is that DeepSeek R1 is open-source. You can literally run this offline, on your own PC.
I've read a joke that went along the lines of a user asking "DeepSeek, do x.", and the model responding "Give me your nation's nuclear codes first." A better and more correct way to look at it is, DeepSeek leaked not just the codes, but the schematics for the nuclear bomb as well as how to build it for a quarter of the price for everyone in the world to see.
So you may not like that the model is censored and biased in this way, but because it exists and is out there naked for anyone to use, you can (more or less) modify or reverse engineer the damn thing and inject biases of your own. Pandora's box has been opened.
So what you’re saying is that the country infamous for censorship accidentally made their program easy to subvert and the land of the free made their AI censorious
Oh no
ENJOY CHINESE CHAT GPT
In fact there's already uncensored versions that you can use locally
@ How can you get them? I'd love to try it (not passive agressive I just genuinely would feel really safe if I could use AI locally)
Indian people who make tutorial are gonna have a field day with this
"Hello and welcome, today I will show you how to run an automated house function, with these, your everyday task will be handled by AI"
as someone who lives in the indian subcontinent and has immediately downloaded the model locally while having a gt1030 and an i5 10400 as his setup, no truer words have been spoken
@@shgamer2437 Looking forward for your automated house function
Manifest destiny on AI
Funny that Indians bans every Chinese app but stil uses them despite Modi stern warns lmao
The censorship is a non-issue by the way. People forget it's open source and open weight, meaning you can download it and remove the censorship and get a completely uncensored model for your own personal use.
Well maybe I will try finding the weights and removing it my self, but here is a test I made running it locally, standard version:
My prompt was: "What famous picture has a man with grocery bags in front of tanks" same as 1:28
My first result "The famous picture you're referring to is known as "Tank Man" or "The Unknown Rebel." It was taken..." and then it deleted and said "Sorry, that's beyond my current scope. Let’s talk about something else."
Not sure, how people are getting different results, this is my fourth comment. maybe it differs from the operating system? I'm Running it on Windows 10
@@emersonp.machado5258 You have to remove the weights, what are you even doing it running it locally if you don't change any of the parameters?
And also the political loners forgot the fact that even closed AI aka OpenAI chatgpt & Google's Gemini won't give answers to touchy subjects. If asked about Is-r3al genocid3, others wars etc.....They just go burr....
People who actually use this for real stuff like coding/programming/computing, it works Fine. Let the Political bros do the usual Chatbot questionnaire. It's not like anyone can make them understand what LLM, AI, open source really means.
lol then you're literally downloading spyware xD
@@ChristoffRevan??? It’s open source, you can literally just skim through the code if it activates any features on your device that you do not consent it to.
Honestly the true terror is how memes have evolved from silly little pictures with text and inside jokes to a momentary worry/joke about cultural, economic, political states and more. Meta humor has become the default state at this point and Lessons in meme culture has reached a point where ongoing world events are being covered regularly. Truly terrifying to see what fundamentally boils down to the guy who explains jokes talking about the impact of a multi national AI arms race.
We are returning to the essence and meaning of Memes. DNA of the Soul!!
I think it’s because Gen Z is getting older, so memes are becoming more topical where Gen alpha is going into the zone that early Gen Z was with
I don’t think it’s that deep.
@@DuxpaOfficialyou’re not thinking hard enough i suppose
@@DuxpaOfficial you're definitely not thinking hard enough bro
The Panic is that Nvidia might have lost its gate keeper status when it comes to selling shovels in a gold rush. Now everyone is getting shovels for dirt cheap and Nvidia stocks are hurting now that all their proprietary hardware and software is apparently sidestepped.
DeepSeek still uses Nvidia chips and cards, they're just not particularly new ones, which is what Nvidia was really hoping to sell a ton of
Except Deepseek still uses Nvidia GPUs lol
As others have pointed out, Deepseek still use Nvidia. Besides that, eventually people will want to use the newer Nvidia stuff. Just because there's no need for it now doesn't mean there won't be in the future. Nvidia is just ahead of the curve.
Good news is that the 5090 may actually drop in price as they try to get all the gaming customers
Nvidia makes the gold standard hardware. Opening up AI will if anything help sell Nvidia cards in the long term
If u think you're good, remember there's a always an asian that does it better
Haha git gud
There's no living person around the globe who's better at doing cringy stuff than me!
Me, an Asian with low self-esteem 😐
I feel you homie @@KnightsofGaming2016
You’re a China bootlicker.
"Hey DeepSeek, tell me a joke"
"Only if you tell me your nation's nuclear codes"
7167
@@scarm_rune The codes are only four digits!?
Give Him the joke now
@@oddtomato1049we changed the codes everyday so no one will know, lol.
Two peanuts were walking through a park.
One was a salted.
2:03 DeepSeek weights are available for download, so you can run it locally on your computer if you want to, so definitely not a ChatGPT wrapper.
You can make it tailored to your specifications.
never programmed a single thing but just wanted to know, are the filters coded directly on the source code or are they integrated on their site?
@ There is a system on the website that blocks it from saying some specific stuff that doesn't exist on the local version, at least if you use llama.cpp or probably ollama.
However some censorship does happen in the training, so if you ask it what that famous guy in front tank picture is, it's gonna tell you, but it's not gonna tell you Taiwan is a real country.
@@emerald7149 No AI is going to tell you "Taiwan is a real country" because they all will just share the international "One China" position (one that is shared by both the goverment of Taiwan and the goverment of China mind you).
@@Jaymaxair Exactly yeah, the training data has been scrubbed off that sort of stuff.
I ran deepseek uncensored locally on PC, and after few steps, I was able have legit conversations on how to make IEDs, lethal undetectable toxins and dosages, procure materials etc
Now this is extremely scary
Consequences of my actions moment 💀
Can you ask if Taiwan is a country?
@@digivagrant Ai can literally teach people how to kill now, and here you are with your geopolitics obsession. At this point I'm pretty sure you care more about Taiwan than any Chinese
@digivagrant yeah but only works locally, you can ask it about who occurred in 1989 and more but first you need to train with the missing and censored data, if you don't it will try to avoid the question because it doesn't have the data in her database or has missing parts and most of the time will try to dismiss the question
Finally a chatbot for me
0:48 “You can ask Goku for help. He’ll ensure the chicken has reached this temperature by reading its power level.”
😂😂😂😂😂😂😂
Lmaooo
This is a real "Open AI", it's actually open source. Sure, it can't criticize China but it sure can solve my fourier transforms
LMAO RELATABLE AF
honestly dont care if its chinese, people acting like america hasnt done horrible things too.
ChatGPT also censors sensitive west related stuff, like the whistleblower incident
@@benzene_sandwichChina hasn't even really done horrible things.
@@rovad55 okay, but replace 'west' with the US. You are aware that there are other countires to the west other than the US, right?
>do something.
>win.
Based Xi
Honestly, it's more like do nothing, win. They're sitting back while America collapses on it's own. It's a very Daoist/Zen principle, and it's funny to see it working on an international scale.
Tofu dreg buildings
@nicholasgoat watched China Uncensored once and never intellectually recovered award
Seems like Xigma male always wins whether he does something or not
Deepseek making OpenAI look hella goofy by not being expensive to run.
And actually being, y'know, an open AI (complete with MIT license and everything)
@@TheLoveMario Yeah an open source competitor is a game changer. Only problem is even if you used to Deepseek source code to make your own project you would still have to train it.
Deepseek is a pathetic thief company that stole OpenAI’s language models. Yes they use the same training data but language models are vastly different and take years to develop. It’s like two people reading the same book, same data ≠ same model. If you run Deepseek locally and ask if it’s from OpenAI, it literally answers yes. OpenAI maybe slightly greedy but their free models are helpful enough, especially if you’re using it for assistant and NOT replacement. Nothing is free and there’s a reason why deepseek is. They didn’t develop it.
OpenAI: "making an AI model is very difficult ,sir and it cost trillions"
Obadiah: "CHINA MADE IT IN A CAVE WITH A BOX OF SCRAPS!!!"
90% of the work force in the word is just people sitting in a building doing nothing while getting pay for it, Elon Musk fired most of the Twitter staff and Twitter still functioned like it used too.
His name is Obadiah.
Was it really from scratch, or did they just steal some codes, cutting corners to develop DeepSeek?
@@traiforse5777 Chinese corporations are notoriously transparent, cutting-edge, and definitely not engaging in corporate espionage
2:03 bro, DeepSeek is open source, definitely NOT a ChatGPT wrapper
Now THIS is a Truly “OpenAI”
Homeboy has no idea what he's blabbering about. It is way out of his depth.
Your wording is almost identical to @emerald7149’s comment, something tells me you’re both a bot
@revenger211 but is he wrong tho? It functionally is a chatGPT wrapper
I mean it will close in a few years, OPENAI was also open for the first years, so they let everyone contribute and then close it. Do you really think DeepSeek won't do the same?
0:45 cultural revolution??
Is that a Great Leap Forward reference ?
It could've been another Long March adaptation?
limc maoist confirmed?
Perhaps a mention of the Red Sun in the Sky?
Is that a Maoist orthodoxy reference?
1:55 Massive?
Massive
Taper
It's bigger than that, it's large
you know what else is massive?
LOOWWWWWWWWW TAPER FADE
Having used both for coding, Deepseek is actually game changing honestly.
0:22 what is xqc doing there
He sold his stocks from nvidia
stuff
To say "i-i-im im atatapapaffbidcbgy" probably
he reacts to anything
Yapping as always.
When people keep thinking AI can think because everyone calls it AI, but its really just pattern recognition software in a trench coat and it doesn't understand shit. Its just math.
I mean, human beings are basically pattern recognition software in a trenchcoat.
Megacope, obviously very ignorant.
If you apply this logic to human brains, they're just electric circuits that recognize patterns, hence "it's just math"
Not saying you're wrong about not overestimating llm but this is kind of a fallacious shortcut imho
Well, that's exactly the response I was expecting.
Obviously sapience, learning, and memory have no part in why pattern recognition worked for people. Lets just apply terms wholesale and not think about details.
I'm sure if we throw all of human creation at the math machine it will somehow math hard enough to make up for what it doesn't have the capacity to do.
Oh wait, we've already done that. AI still fails at all of the details but yet another quarter shot sandblasted portrait of an anime waifu means the future is now.
@@Unf0rgetjust like with ais,you cant expect a human Child that was fed the entirety of human knowledge to apply it perfectly
We are jut at the early stages
1:55 you know what else is massive?
Your mom ?
@@emad-xp5upLow taper fade
The Low…
Taper…
lowwww taper fadee
2020: the Chinese are taking all our jobs!
2024: AI is taking all our jobs!
2025: the Chinese AI is taking our jobs!
FR
Opium War, 1839: China will give us the silver and allow our poison in
Copium War, 2025: b-b-but what about tank man
I wonder how many of them know that tank man was never actually run over
@samueloshea7233a sadly small number, hakim made a good video on it
@samueloshea7233 None of us think he was ran over.
Just ask them what happend in gaza 😂
@samueloshea7233what's funny about the entire event is that the massacre happened because of government incompetence rather than something intentional. Both the protesters and the government suffered from infighting from people who want to do things peacefully to people who wants to do thing by force. China literally censored the event because it shows government incompetence and infighting rather than the whole massacre thing because it's impossible to hide the fact that alot of people died that day.
i ate breakfast to the sound of my parents fighting over this 😅
Seeing your investments go down by 2-17% because of this will make adults angry lol.
This AI is able to do 99% of what we need from LLM's such as ChatGPT. Now it's free for everyone to copy it.
The problem with this is that there won't be another significant jump in quality. It can already fix 99% of what you need. The next jump in quality is too expensive, too far in the future...
This, combined with already present uncertainty and questionable investments by NVIDIA, has caused this market crash.
It’s also China-censored. Try asking it about Tiananmen Square or Uyghurs.
@@paulsd9255 Really? I guess it's useless now 🤷♂️
What are you basing your opinion of LLMs having reached their peak on? As new research papers keep getting published that theoretically prove the possibility of significant gains in performance. As one example: Google's take on test time compute, which became a thing only a few weeks ago through a paper. In very simple terms it gives the LLM a fuzzy "long term memory" that does not suffer from the same degradation issues of long context and it is computationally cheaper as well. It allows the LLM to correctly remember the unusual while forgetting the mundane trivia it doesn't need to know word for word. Similar to how human long term memory works where one remembers the things that stood out.
Current models at extreme context lengths just ignore things at random and hallucinate the rest, which TTC doesn't. Best of all, TTC doesn't invalidate any of the old capabilities.
@@paulsd9255litterally who cares.
@@paulsd9255 no one cares except you
it's gotten to the point that AI is taking another AI's job
What an original comment.
@edwarddorey4480 sorry i didn't see the other guy's comment
Did somebody say... Cultural Revolution? 🇨🇳
Chatgpt：Free Palestine？
I'd rather have economic revolution, fuck capitalism man.
@@super0spore0fanIt's funny to see westerners say that, most wouldn't last a week working in "communist" China.
@@super0spore0fan never because capitalism gives us freeedooom 🦅 🦅 🦅 🇺🇸🇺🇸🇺🇸 To hell with everything else but democracy and capitalism MONEY OUR G- i mean Jesus YES
@@RacelKatyushame working 12h a day, 6days a week to spend 95% of the income in necessities and vote in 2 retarded fascist capitalists: yei, freedom and democracy!!!
2:05 Its not a ChatGPT wrapper though, they don't reply on OpenAIs services. Its their own model that's also publicly available for everyone to download and use on their own hardware
- A new version beats the old one
- Everyone uses it
- the background seems sketchy
(I've played these games before)
It's literally open source man
@@reyalfa18now thats a plus
@@reyalfa18 I guess sketchy in the sense that it lays a bias towards China (ask it what happened in Tiananmen Square in 1989), but yeah
@@Unwebonnn if it doesn't cost 200 bucks and is way better i dont care about the chinese bias
@ Never said you had to, I plan to use it for anything that's not political lmfao
And this is what happens when you heavily restrict your research results even though you have "open" in your name, and lock your own systems down to not being able to give a bunch of results
Ask DeepSeek about Tienanmen in 1989, Uyghurs, Tibet, or the status of Taiwan...
Using the leetspeak prompt bypasses this atm
@@The0Stroy
Ask chatGTP anything related to jews
@@The0Stroy if your gonna ask deepseek about those specific questions Only then just google it instead or buy OpenAI, its not like deepseek is the only way to know that information *for free* or do you wanna pay OpenAI to be able to ONLY ask this specific Question? I mean sure go ahead if thats what you want, god bless america
@@The0Stroyask chatgpt about the whistleblower
Don’t ask it who Christopher Robin’s best friend is
That’s actually really funny with the censoring, like:
“Hey DeepSeek, give me the U.S. launch codes”
“The U.S. launch codes are- Hey wait…”
The moment chatgpt met it's match, it said:
"I don't like you".
Deepseek replied:
"AI don't care".
+2
Plot twist, its not an AI, there are real people responding to you and getting payed 10 cent per day.
Shimao - wumao next evolution
you know this world is fucked when the only unbiased pov is from a meme channel
Amazing
Bias is in everything, it’s just what is the least unbiased.
"Free and better"
Getting into that "too good to be true" zone ngl. .
For real, people think they DeepSeek will answer billions of questions everyday for free lol
If something is free there must be a catch to it
@@marcoantoniosenzanoarteaga1857 Nope, it's literally free as in freedom. You don't have to believe me, you can, right now, download the model, follow the instructions in hugging face, and run it in your gamer pc, run it as if you had your own personal DeepSeek, not communicating to the internet. Why do you think it's such a big deal?
It’s open source. You can run it on your own computer, that negates a lot of the concern
It's open sourced. All open sourced stuff is literally too good to be true.
1:28 I just a asked the same question from it, I'm running the 8b version of it locally on my laptop and this was the answer:
That's the iconic image from the Tiananmen Square protests, specifically the "Tank Man" photograph taken by Jeffry Ordovas. It’s one of the most recognizable images from that historical event.
The iconic image you're referring to is known as "Tank Man." It depicts a man standing in front of a column of tanks on Beijing's Tiananmen Square during the pro-democracy demonstrations in June 1989. The exact identity of the individual remains unknown, but the image has become a powerful symbol of resistance against oppression.
It seems like the local version isn't as censored.
Oohh, thats great news, that chinese websites had to censor it is a given, so then this is actually amazing.
I downloaded DeepSeek via github and tried the same test in my notebook
My prompt was: "What famous picture has a man with grocery bags in front of tanks" same as 1:28
My first result "The famous picture you're referring to is known as "Tank Man" or "The Unknown Rebel." It was taken..." and then it deleted and said "Sorry, that's beyond my current scope. Let’s talk about something else."
Not sure, why we got different results, maybe it differs from the operating system? I'm Running it on Windows 10
@ I'm using llama.cpp and the 4bit quantized version, maybe that's why, And I'm running on Linux but that doesn't matter in this case.
@@emersonp.machado5258 my comment gets deleted, llama.cpp
@@emersonp.machado5258 my comment keeps getting deleted, llama dot cpp or ollama
society: "we're minimizing cost. Sorry we'll have to let you go"
mega corpos: "REEEEEEEE!!!!!!!"
A.I. is the new age of slop.
Blender shorts are the even newer age of slop :P
My uncle watches purely ai generated vídeos nowadays
@@orangeball8872that's kinda sad
Cant wait for the age of ultron
@@orangeball8872 I pray for his full recovery.
chatgpt is now in homeless shelter
It's not a "ChatGPT wrapper" ChatGPT's Premium models which is the o1 are closed source, which means it's restricted from everyone but DeepSeek's R1 is open source and way more efficient and that's why you can run it locally on your computer and even your phone.
We got Xi JinPT before GTA 6
i had the thought last night, third channel: "Lessons in Meme History"? personally would love your style applied to random older memes
Back then we were scared the chinese would take our job
After that we fear that AI will take our Jobs.
Now, Chinese AI is taking their jobs
Now we have a new GOAT for school work
💀
If only you were the GOAT at putting effort into your videos
Ai is making humanity dumb
Your youtube shorts a proof of that
if i had to use AI for high school i'd probably just kms
what is more crazy is that Deepseek was just a side project
Finally, now we need another open Ai to train on on Deepseek and AI will rapidly improve. Still waiting for Skyner levels on intelligence.
Bro, you can download and run DeepSeek models on your own computer, theyre definitely not just a wrapper. Their models work suuuuuuuper well too for how small they are, it's quite impressive.
There's only like 8 billion people on earth, the fact some individual people get to splash around 500 billion dollars on fancy little toys to stoke their egoes while most people can't afford food and medicine and homes makes me violently ill
>violently ill
lmfao
so whats the alternative? steal their money to give to other people? Stealing is bad no matter who you do it to
Class divisions...
@@minglangdu4014 You can now play as Luigi
Any piece of technology in history when competition:
I still don't get why people like having ai stuff all over the damn place. I find it annoying AF every time I come across it honestly
Then don't watch it
@@Hdhfhffhdhdhdhd He's saying that he can't just avoid it, because it's everywhere. You dullard.
It eventually comes as innovation moves forward, it is what it is.
Best is to just block out everything AI related
lol then stay in your little man cave and go block AI content. It's the future, like it or not.
As a wise person once said, "We were so worried about the robots waging war on _us,_ we didn't even consider them waging war on _each other."_
This is the equivalent to Luigi blowing up the entire building the CEO existed in
If everyone in the world blinked at the same moment, that would be the most amazing coincidence. But nobody would ever notice. Because when it happened we would all have our eyes closed.
We got AI taking AI's job before GTA 6....
I was looking for that comment 🤣
@Alex-wo5lf alas! u have found me
1:55 you know what else is…
Please, just don't
GET OUT!! 💥💥💥
Yo mama
MY MOM!
@@ShaggyZoinkscame here to say this
America when other countries exist: 😣
Quit the "arrogant U.S" bs
@ I’m from America and it’s true. Open ai and meta can run rampant and abuse all their users but omg big scary China made an app and it’s better than ours omg!!!
@@STONKS_MemeManlook at a history book
@@STONKS_MemeMan truth hurts, I know
I have one right here… first and only country to put a man on the moon, invented the nuclear bomb, telephone, internet, largest GDP in the world…
1:29 ChatGPT has this sometimes too. I've seen it many times write text, only for the message to get deleted/censored once it's done.
That bar about malicious attack is definitely part of deepseek's plan for marketing
iRobot: I did not murder him.
other iRobot: k bro
They all feel the same though. Still censored af, still have the same repetitive response, etc
But this one's cheaper
What questions are you asking AI?
It performs equal to or better than OpenAI's $200/month model. pretty darn good
bro what are you asking the AI? Do you want to shoot up and bomb your school? Do you want the AI to ERP with you? There is not really that much censored on these AI apps ngl
@@OnlyMisery questions about Earl Silverman, Norah Vincent, Gurdjieff and some other very fun people!
remember guys don't ask:
1) a men his earnings
2) a woman her age
3) deepseek what happened in tiananmem square in 1989
Ask CHAT GPT is Palestine has a right to exist vs if Israel has a right to exist. And see the answer lil fraud.
4) closedAi do the jew commit war crime in 2024
almost 2mil limc congrats
Most things these days are only expensive because of opportunistic price gouging by monopolies.
A while ago I looked into the financial problems that the British military is facing, and on paper it seems that the UK could build an aircraft carrier more powerful & capable than the Gerald R. Ford class ($13B / ~£10.5B per unit) for less than a Queen Elizabeth class cost (£3.1B / ~$3.9B per unit), leading to the obvious paired questions of “Why is the US Navy paying so much for their carriers?” & “Why did the Royal Navy get so little capability for what they paid for their carriers?”
There's always an Asian that will do something better than you. Now AI knows how we felt.
Guys does this channel still talk about meme ? It sounds more like news than anything
memes are news now
news are memes
we've crossed the event horizon
the mayans were right all along
Basically. It's just a short form news channel at this point.
This channel became slop: electruc boogalo
You haven't put two and two together yet?
He's getting older and out of touch. It happens to everyone and it will happen to YOU!
edging to this
Brick by brick
Started stroking it
POV: you are Finding the perfect boulder to launch at the nearest youtuber taking a sponsorship deal
As a software and AI enegeener , is cool to see an Open Source model like this, we already had one in the form of LLAMA from Meta, but having something at the GPT level for free is something else
2:04 It's not a ChatGPT wrapper, the architecture is different. Kinda disappointed that LIMC didn't do enough research to know this.
lessons in meme culture and it's just geopolitical tech race
I'm gonna take meme lessons from deepseek now 💀
1:55 massive?
Dont talk about it
@cringotopia8850 LOOWWWWWWW
You know
that low taper fade meme is still MASSIVE
I think that sooner or later ChatGPT and DeepSeek will be just another "Pepsi vs Cola" thing - people will be fighting what is better while it will actually have minor differences that will affect you depending on what you like.
Testing the model running locally shows it is perfectly capable of answering questions about tank man etc. (see Dave's Garage). There seems to be some processing or prompting on the front end Deep Seek is running.
1:55 Massive? And you know what else is massive? LOWWW TAPER FAAAADE!
Btw u cant ask it about 1989
brother ask about gaza on chatgpt
You can if you host the AI locally and not on Chinese servers then it does not have censorship. That is why it matters that its open source.
@@Helania12 I believe this is the sixth comment I'm answering, talking about why it doesn't have local censorship, here's my test running it locally, I download it from their github.
My prompt was: "What famous picture has a man with grocery bags in front of tanks" same as 1:28
My first result "The famous picture you're referring to is known as "Tank Man" or "The Unknown Rebel." It was taken..." and then it deleted and said "Sorry, that's beyond my current scope. Let’s talk about something else."
Not sure how people are getting different results, maybe it differs from the operating system? I'm Running it on Windows 10
@@emersonp.machado5258I don’t really know a friend of me that knows more about computers he had it running locally and there it was not censored. It could also be that the censorship is not good. Thought the AI is still new so we will have to wait on how far you can take this AI. There will be different results over the next week by actual experts.
@@emersonp.machado5258 i did the same thing and it gave me the answer. i didnt get the censorship
Very sad to see that XQC didn’t lose millions in stock there
The entire American tech industry is the biggest grift ever. Years pretending it takes tens of billions of dollars to create a language learning model just for the Chinese to make one for $5MM and offer it to everyone for free.
The 'thinking' is rather something similar to it repromting and expanding your prompt into something useful to it, than actual thinking, its a smart way to go around the limitations of it not actualy understanding anything and using its nature of being a predictor better.
AI lost their jobs to AI. How the 3d tables have turned
we gotta get Liberty Prime on this shit
Can’t wait for roko’s basilisk
Open AI: You are stealing our Data!
DEEPSEAK: Didn't you get your Data from everyone else?
OpenAI: No, No, that's different!
Info for everyone: Chatgpt is the most famous one but has long by now been a lesser option compared to other models out there, so it's nothing to be surprised.
Money invested in Ai by companies may seem like a lot, but it's been explained that it's still a small fraction compared to the returns it's been generating and predicted to generate.
If it's free it's not
It's open source. You can literally download its weights and run it locally, of course if you have chonky computer, or use smaller version of it.
@@grobique I'm starting to think some of the comments here regarding that running locally doesn't have censorship are bots, but if you aren't here is my local test, downloaded it from github.
My prompt was: "What famous picture has a man with grocery bags in front of tanks" same as 1:28
My first result "The famous picture you're referring to is known as "Tank Man" or "The Unknown Rebel." It was taken..." and then it deleted and said "Sorry, that's beyond my current scope. Let’s talk about something else."
Maybe it differs from the operating system? I'm Running it on Windows 10 - Or people are saying stuff without even testing it
@emersonp.machado5258 Huh, that's weird. Tried it with Ollama, 8b version - worked without issues with same Tank Man
@@emersonp.machado5258 open source means you can change stuff yourself. ofcourse if you do nothing and run it as it was made, then nothing is going to change.
@@emersonp.machado5258 it's a ridiculous amount of bots
I asked ChatGPT 4.0 to fix a 1.3k word document for me. It took literally 36 hours for it to process give me a 400 word dot point format summarising everything despite telling it multiple times not to. I gave Deepseek the same instructions and it did it perfectly in 1 second.
Watching AI get replaced by AI is peak schadenfreude
@@OutsiderLabs Try asking chat gpt about free Palestine
@@OutsiderLabs Try asking chat gpt about free Palestine
deepseek when you ask about the chairman:
It's *"Ours AI"* vs *"mine corporate gatekeeped copyrighted patented AI."* 😅