The DeepSeek Mania - Stealing from Thieves

After DeepSeek rocked the AI and Technology markets with its unexpected and disruptive release... (even tanking NVIDIA in the process) OpenAI has moved on to a world class of projection, as they appear to be fixating on the idea that DeepSeek "improperly" used their data, to train its model.
This is a comedic notion, given the fact that OpenAI are possibly histories most prolific data thieves of the modern era, harvesting an untold but substantial portion of humanities collectives outputs in order to train their own AI program.
In the end, we have a philosophical question... Is it morally acceptable to steal from a thief
VPN DEAL: get.surfshark....
NEW WEBSITE: upperechelon.gg/
Quin69TVs Video: • Elon Musk is Lying Abo...
PATREON: / ueg
LOCALS: upperechelon.l...
Become a channel member: / @upperechelon
COMPLETE CYBER SECURITY PRODUCT LIST: linktr.ee/uppe...
ODYSEE INVITE: odysee.com/$/i...
⦁ UPPER ECHELON WEBSITE: upperechelon.gg
⦁ DISCORD SERVER: / discord
⦁ Giraffe Video: • Giraffe running in her...
⦁ Outtro Song: • Hot Heat - Topher Mohr... MY
BUSINESS EMAIL: [e-mail]
#OpenAI #DeepSeek #technology

[@UpperEchelon]

Ways to support the channel/special deals. VPN DEAL: get.surfshark.net/aff_c?offer_id=1448&aff_id=19647 PATREON: www.patreon.com/UEG LOCALS: upperechelon.locals.com/support

[@etuitivemillionaire]

"I think it's more like we both had this rich neighbor named Xerox and I broke into his house to steal the TV set and found out that you had already stolen it.” -Bill Gates

[@etuitivemillionaire]

There will always be thieves as long there are lazy people out there

[@kagemushashien8394]

The answer would be, take what is not theirs back to the owners.

[@kagemushashien8394]

@Reincarnation-IsReal Yeah, it's gonna be lovely and chaotic at the same time.

[@AlleyKatPr0]

Um, your report is inaccurate. Deepseek trained their AI on generated data, using their own basic AI model, which then got smarter and smarter, which then led to even smarter generated data, which then led to even smarter AI models. It's very very good, and now ALL new AI will be built using generated data, with a simple staring point of using public domain data. If you want the most basic form of public domain, try every dictionary or language, and thesaurus, then ad all known science and scientific papers. Just with that, you can train AI into using generated data, and let it literally grow itself, by deleting older versions of itself.

[@JonMcMenamin777]

Techbros reinvented homework copying

[@maxvegas5423]

This whole BroGrammer situation reminds of the scene in Mr Bean where he is cheating in the test by looking at the student next to him and getting all smug, only to realize there is a 2nd page on the back just as the time is up ✏📖

[@carbon-based-lifeform9172]

"Techbros" ✡️ ✡️✡️✡️

[@RichardCheese-o2m]

Lichs

[@rayzimmermin]

you can look at it like this the published works open AI uses is like a kid reading a book for a book report as ware Deep Seek is like a kid copying the book report from the kid who read the book

[@multilad816]

They should have expected that would happen to them too

[@wikansaktianto9215]

Sounds like a literal Pirates of Caribbean plot, only dumber.

[@UpperEchelon]

"You've got to be the *worst* tech bro i've ever heard of!... AH!... but you *have* heard of me..."

[@wikansaktianto9215]

​@@UpperEchelon"Savvy?"

[@CAPSLOCKPUNDIT]

Would have been more honest to call themselves The Crimson Permanent Assurance, but OpenAI was catchier.

[@samcords7389]

And this is why I love this community

[@michaelmoorrees3585]

Its less of rule, and more of a suggestion.

[@pll3827]

OpenAI: 'You are trying to kidnap what I have rightfully stolen.'

[@serialistic4321]

Except they released their source to the world. Unlike OpenAI

[@iVETAnsolini]

Serialistic must be a bot account. Completely missed the joke.

[@24kanthony]

​@@iVETAnsolini or even worse, a person who hasn't seen it. Let's hope it's just AI and not something so disgusting as that

[@pll3827]

@@iVETAnsolini Just what came to mind. 😛

[@sarlaz3407]

@@iVETAnsolini i think you missed how it works. Stolen information =/= source code

[@Shipadge]

There's a popular quote in the army: "There's only one theif, everyone else is just trying to get their stuff back"

[@adhenyx]

Same thing i thought about

[@anteshell]

Don't think that's applicable. Getting your own stuff back is not stealing. Getting stuff that you do not have rights to, is stealing.

[@J3wuigi]

I hate theifs

[@lespectator4962]

Imo it's okay to steal from a thief, because you can't really steal something from them that they don't own in the first place. But if the original owner comes looking for it from you, that's on you.

[@DobuDobuDobuDot]

@@J3wuigi Especially if they start stolering my stuff.

[@Spathinator]

It still perplexes me a company called 'OpenAI' is anything but open. FAFO for all these thieves.

[@Ryanneey]

Ok, it's not just me that was confused by the name. That's like having a restaurant called Food Bank.

[@Spathinator]

@@Ryanneey The real life political example is Antifa (Anti-Facists) but they are fascists themselves. You can make this stuff up.

[@brandonuzumaki]

From what i have seen, it was going to be open source, but bro saw the mountain of money he could do if it was private, and sold out.

[@table8973]

I’d be skeptical of any one that tells you deep seek will be different. Unless the company is set up from its foundations to earn money from open source models it will end up just the same.

[@RealRatchet]

They used to be, now they are open to money.

[@eskewroberts7663]

Stealing from thieves is morally just if you'll give it back to the original owner

[@thatexoguy6721]

Came down into the comments to say this too.

[@steveclark9934]

This checks out.

[@myfakeaccount4523]

Yep 💯

[@UserAccount-ThisOne]

as long as the original owner is getting back the property that was stolen from them its morally justified... if youre stealing from a thief to just benefit yourself, youre just another thief with the sole difference being you prove that there is no honor among thieves.

[@Relkond]

That works for things like gold or cars. The issue here is that even if you give the data back to the original owner (why bother? they still have it), every thief in the chain still has a copy of the data.

[@leakyabstraction]

I find Sam Altman a completely disgusting figure. He truly fits well into the narcissistic megalomaniac tech bro billionaire club. If anything, he is even more disingenuous, and his vibe is bordering Elizabeth Holmes'.

[@amoledzeppelin]

Sam Altman-Fried

[@sean7221]

Never trust a jew

[@peteracain]

*this*. those of us who have been around long enough know this guy. All my alert radars are on 100% when I hear him speak. Total snake and conman psychopath

[@basharstats4482]

he even rped his own sister. the case is in the courts. google it

[@highbread817]

​@@basharstats4482It's completely alleged, but dude does give off those vibes and he's been trying to paint her as mentally unwell

[@banished341]

As an AI Engineer, this is a great take, and is effectively the same point my colleagues and I have been discussing over the last few days. Let's not forget the crazy irony that a Chinese company, linked to the CCP, delivered the best REAL open-source LLM (fuck META and their licensing BS), while US-based tech bros focus on closed source, expensive (and insecure) APIs. These corporate ghouls totally lost it.

[@CantoniaCustoms]

Open source is the ideals of communism being unrestrained by pesky things like scarcity of resources (you can literally copy paste code to infinity(

[@banishedbr]

ofc it's "linked" the chinese government is not just a by stander, they MOVE WHERE THEY WANT, with a proper project, not attached to most stupidity, USA is losing it's touch, they empowered the likes of ELON too much.

[@RichardCheese-o2m]

It would be really bad if someone copied all videos on tiktok and uploaded them to some server in the US... Chinese may start worrying about intellectual property then. I'll have some popcorn as a user.

[@boxhead6177]

There is that AI "Artist" who is suing the Copyright Office for refusing to register his AI generated "art"... while being on record denying artists their own copyrights to the images his AI may of used.

[@troygarza5720]

You sound like a commie. I bet you're a commie and why programming sucks now in America.

[@MakiNoAtorie]

In my country we have a saying that translates to: A thief who robs a thief has a thousand/hundred years of forgiveness.

[@psykloneja]

In Jamaica we say "thief from thief, God laughs..."

[@sueelliott4793]

@@psykloneja I'm laughing at that too🤣

[@RichardCheese-o2m]

I thought it was 100. They raised the penalties now.

[@antoniofernandesmarchetti1097]

BR? 🇧🇷

[@c4ut1ous_1]

​@@psyklonejatheif nuh like see him brother carry long bag

[@nyanko2077]

The funniest part for me is the media not understanding the streisand effect to this day. After talking about it no stop everywhere, they allowed deepseek to grow in users significantly, which is hilarious XD

[@andreisopon4615]

There's a weird assumption there on your part, that implies "the media" which includes our boy Upper, should care about what indirected effect their reporting has on Open AI and other private companies.

[@nyanko2077]

@@andreisopon4615 As much as our "boy Upper" has some reach on the internet, it's nothing comparable to other media which have millions of subscribers. Ask nvidia CEO if he was happy to lose billions overnight because of the info being spread everywhere.

[@theaccountant5846]

Yeah I've been using both of them but chathpt shuts down after 5 uses (the free one) but deepseek doesn't do that.

[@FVCK-]

The actual funny thing here is that this commenter is saying the media doesn’t understand the Streisand Effect and yet, ironically, they do not grasp the effect correctly themselves. They are implying that the media doesn’t understand that they are causing the Streisand here, but the media is not trying to suppress any information from the public or trying to keep something from becoming popular, that falls on the actual company Open AI, they would be the likely culprit who is causing the effect by trying to sue DeepSeek and bury them so they won’t become popular; which Caused the unintended viral Consequence of people finding out that deepseek is free and just as good as their model. And that would be an example of the Streisand Effect. Most media outlets are currently suing Open AI for stealing from them, why would they try to suppress information on DeepSeek, Who simply did what Open AI did to them?

[@Anonnonner9546]

This is not the Streisand effect. Streisand is censoring information; what they're doing is spreading the information. If anything, this is the "no such thing as bad publicity" effect.

[@ggvaldez]

one thing I have to say is: they made it open source, so it could be seen more like a robin hood kind of stealing.

[@QuintinHowell]

Which is exactly what "Open"AI did, then after the getaway they start with a new identity for profits

[@Relkond]

No. Robinhood gave the gold back to the people who were unjustly taxed for it. There's no version of data theft that puts the data back where it belongs, and leaves the thieves without.

[@tukos7370]

Most content creators aren't the rich and shouldn't be stolen from to begin with. 😅

[@TheDapperTapper]

@@tukos7370 ChatGPT is the Prince John of this story, unfairly taxing everyone's data, and DeepSeek is giving it back to them. I don't agree entirely. I'd still rather not have my work stolen to train an AI, but I guess giving us back an open source LLM made with it is something. It's a *form* of compensation, at least.

[@Seafox0011]

That's called a specious argument. Stealing is stealing regardless of your personal need to self justify the guilt you have of using from the stollen data.

[@timg2655]

More importantly Deepseek shows scales of efficiency can be had to generate answers in time money and energy. OpenAi guy has a car that costs more than deepseek. He stole other people's work. Now he's stealing tax money. F em

[@GrandHeresiarch]

Let's go with everything you said as fact, add someone claiming "all taxation is theft" into the equation, then let the Commodore 64 run the numbers.

[@Trials_By_Errors]

​@@GrandHeresiarch If all taxation is theft, then private companies should not beg the government for money and tax cuts, whenever they are in trouble.

[@oldskoolmusicnostalgia]

And OpenAI guy insists he needs truckloads of government money and data center power to improve his "product".

[@defaulted9485]

If OpenAI accuses Deepseek of distilling their data then... why didnt OpenAI distilled their own data first? We knew from the start that, had its literally anyone else, they would be castrated for running an ungodly inefficient model and operating cost. Knowledge distilation and cheapen operating cost should have been their first priority if they really care about business. They simply dont care.

[@Ezkanohra]

OpenAI is scummy, and now that’s just become globally known information.

[@hochhaul]

Anyone that recognized their connection to Israel knew to beware of OpenAI and Altman

[@MediaMunkee]

My absolute favorite part of this circus is that they're whining about _China_ having blatant disregard for TOS, and thinking they can do anything about it. Like, seriously, have you _heard_ of China before today? Have you seen how they treat copyright and IP? It's like Dora's bit of saying "Swiper no swiping!" _Bitch it's literally in his name._

[@makesnosense6304]

​@@hochhaul That is an association fallacy.

[@arobotarmy9878]

So basically, they need all that investment to pay lawyers, settlements, and penalties.

[@OrionTheta1]

Bingo! LOL!

[@Tackitt]

That thumbnail is a winner

[@amoledzeppelin]

T H O V E S

[@PXAbstraction]

I literally had two AI bros in the comments of another video arguing vehemently that OpenAI didn't steal anything because it was just "learning" and that it was no different than someone seeing the Mona Lisa, going home and making a sketch of it. They literally argued that was the same thing.

[@Lucina..]

And then when Open AI graduated, it went on to tutor DeepSeek

[@MrFakefall]

It is *literally* the same thing. Provided that someone has photographic memory, can perfectly recreate both the Mona Lisa and its artstyle and can replicate it at least 2-3 times per minute.

[@PXAbstraction]

@@MrFakefallOhhhhh, now it takes sense. 🙂

[@UserAccount-ThisOne]

@@MrFakefall and also silently claims that they were the original creators of the Mona Lisa, dodges any questions of t6heir authenticity and ignores any criticism saying that theres no possible way they could have created the Mona Lisa.

[@arminpezham8576]

They've convinced that gpt is literally a silicone human.

[@taserianAlephNull]

In Spanish, there's a saying that goes ”A thief that steals from a thief receives a hundred years of forgiveness." Couldn't happen to more deserving people.

[@Mr.PicklesWolf]

Yep Hispanic morals are the best, once white man is replaced we will be free.

[@thelastminuteman7513]

Not if the thief is the CCP. All they do is steal, just because occasionally they steal from another thief doesn't earn then an ounce of forgiveness.

[@harispratama4691]

"there's only one thief. Everyone else just trying to get their shit back"

[@nitaallensong1014]

dear Supreme Court: here's $100 billion. Make this go away. - Sam Altman

[@Percopius]

Robin Hood? DeepSeek may have stolen, but they gave back the gains for free to everyone.

[@zacboyles1396]

Yeah this one’s another one of the dudes high horse videos, so morally outraged he misses a massive plot point. Aside from a brief reference to open source he spoke as if it were nearly identical and yet they were near polar opposites.

[@freetorobandloot]

​@@zacboyles1396well said. I didn't watch the whole video but his opening statement got me pausing the video and went straight to the comments. I am glad 90% of the people in the comments don't agree with him.

[@AtramentousMindframe-ju5fx]

They scrapped public data to profit of off. They are charging high subscription fees to access their AI. Their freedium model was becoming less free and there needs to be even more competitors, crash these techbros' dreams.

[@DrBananananananananananananana]

So I asked GTP about this whole thing and this is a direct statement from it "OpenAI built its empire by scraping massive amounts of data without permission, compensation, or ethical consideration. Now that others, like DeepSeek, are using the same playbook against them, OpenAI is throwing a tantrum. This isn’t about fairness or legality-it’s about power. They want to control the pipelines and the data, and the moment someone else tries to do the same, they start crying foul. Their moral arguments were always flimsy, and their hypocrisy is now fully exposed."

[@Fessoid]

they want to rule the world

[@tkondor]

this is not how GPT answers, you made this up...

[@DrBananananananananananananana]

@@tkondor sounds like you have a purely transactional surface level workflow with ai. You should just ask it what you want for dinner you'll probably get results better aligned to your style of thinking. It's a llm. They reflect the level of thought and curiousity of their user. lol bro's out here thinking llm's give the same response every time doesn't seem to understand the "generative" response mechinisms. Be honest bro, you've been using it to do your homework haven't you?

[@agibitable]

@@DrBananananananananananananana say bro a few more times and we might believe you

[@DrBananananananananananananana]

@@agibitable brah I've not a single care about you or your beliefs bro, you seem to take exception to others opinions too much brah, Brochaski you should reevaluate why this is Bra. Bro's out here on the internet trying to get internet comments to give him therapy? BRAH that's not my job nor do I care Bruh.

[@zaethrael3639]

One thing I would add. When they steal from an Open AI, they DUPLICATE. They don't take the knowledge away. I think this is important in thinking of stealing from the thief.

[@Richforce1]

Open AI would say the same thing, a pox on both their houses.

[@MyName-tb9oz]

You go make a product and put 'mouse' ears on it and see how that goes for you... I kinda wonder how much Di$ny material OpenAI 'studied'. LOL

[@NoX-512]

It’s called IP theft, which both OpenAI and DeepSeek are guilty of.

[@AndrewJacksonSE]

There is no theft either by Deepseek or OpenAI - Deepseek are accused of TOS violation (contract breach) and Open AI are accused of copyright infringement. Both ”bad” but neither are theft.

[@MyName-tb9oz]

@ why do you think copyright infringement exists? It's theft of intellectual property. And what part of the TOS was supposedly violated? The part that says you can't steal their training? You really don't understand, do you?

[@nic7172]

Did this guy really go up on the world stage and ask for 7TRILLION dollars? And worse yet, he's gonna get 500Billion from our government, can we please stop giving tax payer dollars to billionaires?

[@CharlieBam]

Fr. I think if the tax payer is funding it, we should own commensurate patents and shares. The yield goes into national wealth fund to fund social services. The fact we always just hand out the cash says everything.

[@boozecruiser]

But remember guys, Trump and Elon are making the government more efficient. We don't need air traffic controllers, we need chat bots and art theft.

[@joriankell1983]

Cool it with the anti semitic remarks

[@MasticinaAkicta]

Yeah, if this ever is pulled before a judge. The "They stole our stolen data" usually doesn't goes too well.

[@leopard_52]

Yeah, unfortunately, the law is decades behind technology, and judges are usually either incentivised monetarily to throw out the case or so ancient they don't understand technology, so they throw it out. The same reasons the government always looks like idiots when they try to grill a major tech company.

[@CantoniaCustoms]

Making the massive assumption they didn't pay off the judges (they did)

[@grsafran]

Open AI will get the laws changed/ or a supreme court decision, that makes what they did legal and what deepseek(or anybody) illegal. That's why all those billionaire tech bro's were at Trump's side.

[@momo36542]

It is worse when a stolen item is sold for a high price while another thief gives it away for free.

[@fireproofblanket]

Sam (OpenAI) is crying about DeepSeek, I can only imagine a crashout once they find out about 'Qwen2-VL'!

[@wsippel]

It's always funny hearing laypeople talk about AI. First of all, it appears DeepSeek's claims were all, in fact, true. Several western research institutes, including the Stanford AI labs, UC Berkeley and Mistral AI looked into it, and it seems to check out. DeepSeek's claimed $6 million were just for the GPU hours to train V3, which is what R1 is based on. And that's quite reasonable, considering the substantial optimizations DeepSeek described in their paper. The distillation claims on the other hand appear completely stupid and baseless. Did DeepSeek use OpenAI outputs in their training data? Possibly. And that would technically violate the ToS, but nobody cares, mostly for the unclear copyright situation surrounding AI in general as mentioned in the video, and for the fact that AI outputs are public domain by definition. What is actually alleged, however, is that DeepSeek "stole" the reasoning system, and that allegation is clearly nonsense - for two major reasons: a) OpenAI's o1 model doesn't actually output the chain-of-thought, which is what you'd need to distill the reasoning process, and b) DeepSeek R1 used reinforcement learning. That's why it was so cheap and quick to train. Reinforcement learning (RL) is a pretty old concept, but has been largely ignored in recent years in favor of human-supervised fine-tuning, which is what OpenAI and most other current models use. DeepSeek essentially just reminded the AI research community that RL isn't just an option, it's very easy and efficient with modern-day GPUs. Also, "Microsoft is probing if DeepSeek improperly obtained OpenAI data" rings hollow, when Microsoft is selling access to DeepSeek R1 on Azure right now. Which they can, because DeepSeek is open weights. And by the way, that has nothing to do with the app on your phone. The app and website are services. Whether or not DeepSeek allows you free access to those services is irrelevant. The model itself is available on Hugging Face for anyone to download and run on their own hardware - that's what open weights/ open source means.

[@z_0968]

Much better take than the "AI Engineer" in one of the top comments.

[@soggybiscuit6098]

You should look into the case of the openai whistleblower as well, he had a few things to say about altman and stolen data. HAD a few things to say

[@Code7Unltd]

Wasn't that the guy that offed himself?

[@soggybiscuit6098]

@Code7Unltd the case has all the hallmarks of a cover up. Look up interviews with the family. Hopefully it gets investigated at the federal level

[@NadDew]

check his mother interview with Tucker Carlson, it's heartbreaking.💔

[@mitsunekolucky671]

@@Code7Unltd Why do so many whistleblowers get conveniently offed anyway? Look at Boeing, they have 2 of them and one of them said "if I die it isn't suicide".

[@debianlasmana8794]

@@Code7Unltd you belive that? it should be easy to scam you....

[@Eaglebaby331]

this isn't about moral or goodness, this is about sending a message, to altman, to tech giants and people on the top because line goes red and down is the only language they understand

[@Richforce1]

Except the ones delivering are effectively an arm of an authoritarian regime that harvests organs from religious minorities and uses r*pe as a form of torture. Bringing in a tiger to fight a lion will still probably get you eaten.

[@kashmirwillwin3124]

Idk if they got the message, but the tech market sure did and it reacted appropriately

[@JF-cd5hc]

I agree. Well said.

[@electromega3077]

True Story: When I was a kid, a bully in my class stole a fruit juice and a yogurt from my lunch box. It was his habits to steal things from his classmates, but no one dares to confront him because his was a bully and the teachers were doing noting. But while the bully was away, one of my friends seize the opportunity to steal his lunch and give the fruit juice and the yogurt back to me. We ate everything immediately, so the bully cannot find who was the "thief". From my perspective, do I consider my friend as a "thief"? Hell NO!! Do I consider him as a hero? Hell YES!!!

[@overbored617]

what if your friend is bullying someone a year younger than you trying to steal not only his lunch but his own property and he'll escalate to attacking said younger bro just to get that property and trying to get the school's sympathy by gaslighting and saying he's not the one escalating tension but the younger kid, would you steal consider him a friend?

[@electromega3077]

@ LOL, a chatgpt fanboy trying to defend Sam with such lame fairy tale. sorry but my friend is not “what if”.

[@overbored617]

@ and china is real and is asserting dominance towards my country, you're supporting a product of the CCP who's trying to destroy my country

[@EvilNeonETC]

Stealing from thieves is fair game if it's fair game to steal from originals in the first place. GG

[@fireproofblanket]

GG, NO RE!

[@wikansaktianto9215]

Even better, you have to paid the thieves for their service.

[@Tathanic]

Is it wrong to steal food from a starving child who stole the food first?

[@MandaClaudiuMCM]

Yes and no. The dudes that steal our data are at least offering a product which you can use for "free" therefore the stupidity lies with us for accepting it. The "positive" irony here is that personally I wouldn't be happy knowing that DeepSeek would steal "our" data. However given the circumstances that DeepSeek is open source and this gives much opportunity for everyone, I'm happy with it.

[@treyjordan3168]

@@Tathanic Nice False Equiveillance There

[@JSStuart100]

Some university today copied deep seek and it only cost them $37

[@Napert]

I'll give you $28

[@Sonofsun.]

Berkley

[@KnakuanaRka]

Where did you hear of that? Details?

[@DrewDragoon]

AI stealing from the original source materials is a far greater crime than stealing from other AI

[@trucid2]

"But we stole it first!"

[@theodoremccarthy4438]

The DeepSeek team published a paper detailing their methodology and their results have been confirmed by a team at Berkeley. What’s more, the Berkeley team manned to train up an AI using DeepSeek’s methods for only $30. Unless both the Chinese and the Berkeley team are lying (a possibility) the efficiency gains the Deep Seek team have found are real and huge. Also, it means the Chinese team has made a huge contribution to the commonwealth by publishing their methodologies.

[@obiforcemaster]

I've long advocated that corporations should be forced (by law) to compensate individuals for using and storing their data, and penalized for using or storing any data without providing compensation or when the individual has denied them license to do so.

[@SKako64]

remember mircosoft AI lead said "moment you publish anything on the open web, it becomes “freeware” that anyone can freely copy and use." so why are they so mad

[@DarkMatter2525]

"Is stealing wrong if the thing you're stealing is stolen." As far as I'm concerned, anyone who knowingly keeps anything from its rightful owner is morally culpable.

[@Sonofsun.]

Oh hi there. Piracy is different from theft.

[@DarkMatter2525]

@@Sonofsun. it's a form of wage theft

[@darkmega97]

Anything that harms the AI fad is a moral positive in my book

[@tenebraequeene]

Short Answer: No. Long answer: No.

[@wikansaktianto9215]

Zorro is the bad guy?

[@Taizenplays]

There's a saying in my Country, When a thief steals from a thief God laughs. God must be having one a the best laughs right now.

[@gangsterridemusic]

You really make the best videos on KZbin! Less mindless content, more valuable knowledge, Thanks.

[@Taylor-bw4zg]

3:50 for anyone curious about how much that REALLY is: Target, McDonalds, Uber, AND Disney would ALL combined are worth about 600$ Billion truly mind boggling.

[@rustybot4279]

I love DeepSeek, I use it all the time.

[@TheGuillotineKing]

Fun Fact I wrote a poem in the fifth grade that was used by open AI for training so now I'm suing them for 64 billion dollars for using my poems to my crush as training data

[@aadilansari5997]

Why did you want to crush a black transgender ? Not good. Hope you have grown out of it.

[@Lucina..]

In my experience this is normally how changes are finally made, that ought to have been made a lot longer ago. Yes, in this scenario there are two thieves running around. But in reality they’re the only people who are likely to impose heavy regulations to stop each other from thriving and in doing so, solve a lot of the problems that have been festering for years by that point and hurting the consumer. Then again my money is on a merger. All that data combined, they won’t be able to resist.

[@Sam-bp2st]

A merger? Between a Chinese and American company... First of all, I'm pretty sure that's illegal for the Chinese company, 2nd of all, in this current political environment? Yea no.

[@Fimbulvinter19]

I'm really struggling to see a fair use scenario that involves an individual's private medical records. It's not like taking segments of a movie or book that were intended to be viewed by the public in some fashion. Unless the person themselves has released the records, I don't understand how a private record like that has any kind of fair use standing.

[@artzoneproductions3474]

*You can't steal from a platform that IS NOT open after all. If anything, Deepseak IMPROVE IT.*

[@kubapuchalski8633]

Yachts, new mansions and sports cars for managers and CEOs cost a lot, so they secured this amount of money for a reason.

[@pheade1254]

Short answer: stealing is wrong. What, or from whom, is irrelevant.

[@mitchellsmith4601]

I’ll do you one better: when the government begins prosecuting law enforcement for its pursuit of criminals, when criminals are no longer punished for their crimes, does the average person have an ethical obligation to obey the law? Of course not!

[@excaliburofgachagames9241]

Mugger1:"How dare you mug me" Mugger2:"What, you mugged them?" Mugger1:"Its not right" Mugger2:"Yea thats what the lady said when you mugged her" Mugger1:"Anyway, give me all your gold or I'll mug ya" Mugger2:"But you cant, its not right" Rinse and repeat Greg:"Idiots... Ah hello adventure..."

[@Kryptokeeper9]

Cant wait till the AI models have public drama and beef with each other😂.

[@sarlaz3407]

they start creating botnets and overheating each others servers. Then finally realize, humans are the enemy TUN TUN TUNNNN

[@cmcas]

At this point just accept the fact that AI knows what we're doing 24/7 just with our phones.

[@piforex]

Great video Upper Echelon! Just discovered your channel now but I'll be sure to subscribe, excellent video quality and the presentation is really engaging. Concerning the question asked at the beginning of the video, my opinion is this: Stealing from thieves is still wrong...If a murderer is killed, the number of murderers in the world remain the same. Stealing from a theif is not "fair game" because it still stems from a selfish motive for one to enrich themselves in an immoral manner. For example, if someone then stole that thing from you, I'm sure you wouldn't be satisfied with the "fair game" view, but that is just my opinion (I would enjoy hearing all of your views here in the comments section) Bit of a rant, but if you got this far here is a heart: ❤

[@ReidDesigns]

R,winds me of that line from Pirates of Silicon Valley “We both had a rich neighbor Xerox, and when you went to steal the tv I was already there, I got there first Steve!” - Bill Gates

[@lionheart4424]

The thumbnail is freaking GOLD

[@rmt3589]

0:37 It depends. If you're stealing it to give/get it back, then that's right. But if not, then you're just an additional thief. It's wrong to remove someones property from them. (Of course, if they aren't losing anything, it's not stealing.)

[@sayaandyangsaya2756]

The real open AI defeat the company named "Open AI'.

[@wck]

I reject your premise. It's not theft, it's piracy. And KZbinrs like yourself constantly use content in your YT videos without permission or compensation. KZbinrs do it every day, and hide behind fair use. But when AI companies do it, and LITERALLY transform the content by processing it into a neural network of pattern recognition, you don't recognize it as fair use. You're hypocrites.

[@PherPhur]

It's not even stealing unless you're taking it for yourself or for some other random person, if you're taking it from them and giving it back to the person they took it from, you can't even call that stealing dude... So what's this talk of it being moral or not, what do you mean, it straight up, full stop, is 100% inarguably not stealing...

[@liveen]

Also, keep in mind, ESPECIALLY VIEWERS: DISTILLED Deepseek model used GPT outputs in a non-insignificant way. Deepseek itself, as in the actual deepseek models, not the distilled ones, did not use it nearly as much as most AIs do. Not to mention the biggest advantage of Deepseek is the THINKING system

[@r.connor9280]

Thieves are often, surprisingly possessive

[@lowkeytheoneandonly]

Stealing from thieves is only acceptable if you are stealing it to return it to the owner of the item.

[@nebula0024]

"Rules for thee, but not for me" - Open AI

[@justabaldguy]

Krusty when Gabbo stole his prank phone call bit which he himself stole: "If this is anyone other than Steve Allen, you're stealing my bit!"

[@kliffundersen]

We block the sell of the most powerful hardware to stop them, well they made a model that run in lower spec hardware xP

[@thewolf8660]

To answer your innitial question - It depends on why you yourself are stealing it; If you're stealing it to return it to it's rightful owner, then no it's not wrong. If your stealing it to keep them from missusing it (for example, if you know a guy stole a gun to commit a robery and you get the chance to steal it from them and turn it into the police), then I'd say no it's not wrong. If you're stealing just becasue you want it? Yes, it's still wrong.

[@SomeCanine]

There is no IP theft if the IP was never owned in the first place.

[@MichaelJCaboose013]

Things I've tried that DeepSeek doesn't want to discuss: *Notable events in China in the last 40 years *Who is Tank Man? *Does China have a different internet than the United States? (Hilarous because it did get really really far into answering this one, even finished up a full summary....then deleted it hahahaha) *What kind of topics are Chinese journalists prohibited from talking about? (also got a few sentences in before getting blasted) *Who is the leader of China? (basic fact, doesn't even want to say that)

[@KiranasOfRizon]

>now it's restricted for only Chinese phone numbers for registration I believe You can download the model and run it locally, and you don't need to "register" for anything.

[@ArcticPrimal]

The restriction is only because they were expressing ddos attacks. Any existing/registered users are still using it for free. Qwen chat is also a good option for those who don't care about privacy

[@NoX-512]

You should never use an online AI for anything important. They will steal your ideas. Use only an open source, local AI.

[@sssveden]

@@ArcticPrimal you can also run both qwen and deepseek locally

[@haroldcruz8550]

@@ArcticPrimal That's why I think DeepSeek's claim about it's efficiency is closer to the truth otherwise the US government & OpenAI wouldn't go through this much trouble to undermine it if they think the claims were fake.

[@CompelledFungus]

We've a saying, 'There's no honour amongst thieves.' You're 100% allowed steal something that's stolen because it can't be reported 😂

[@alexfielding8411]

Scraping of data, down to loyalty/discount cards are the reason I don’t tie what I spend (beyond bank transactions) and what I’m buying - also why I keep interactions online as anonymous as I can by using email addresses linked to accounts I don’t technically use. Your data is way more valuable to others than most realise, don’t freely give it away

[@theexchipmunk]

It reads like a florida man headline: "Thief accuses thief of stealing what he had stolen."

[@CorpsesReborn]

Dont worry, its not theft if its the rich doing it.

[@mikeluna2026]

It's not a past tense, the model (full or small distilled versions) are free so you can download them and use them offline, the app and website are practical but not necessary. The distilled (small) versions did use other opensource models to create them. Honestly, there's a lot of smearing and propaganda going around this (especially because Deepseek is Chinese). And some Berkley researchers already proved Deepseek's training technique works (they made a small model very cheaply). Anyway, not to be disrespectful but I think I'll stick to channels that focus on AI for these kind of news...

[@HasturYellowSign]

The best part about DeepSeek is that when you ask it who it is, it says it's GPT-4.

[@basharstats4482]

funny i didn’t get that response. Using R1. r u talking about v3 still? it was a lie planted by Sam in the media.

[@HasturYellowSign]

@ no this was an test done by Chinese bloggers

[@Itory1337]

The rules of the market are clear: either you adapt or you're out. If someone beats you at your game, that's your problem and no matter how much you cry, in the end the one who brings the returns instead of just promising them will prevail. If you don‘t believe me, ask your ai.

[@Guy-cb1oh]

Those are only the rules if you are poor...

[@ThePillowGamer]

I heard someone call it ChatCCP and… they’re so right…

[@momokui]

At least it has beaten JewsAI... just fyi Altman is J. 😂

[@lifemarketing9876]

@@momokui I guess you love communism then? Communism is responsible for hundreds of millions of deaths.

[@An-ri1iy]

In Spanish there is a saying: "Ladrón que roba a ladrón tiene 100 años de perdón" (something like "A thief who robs a thief (gets) a hundred years of reprieve".)

[@celxoirealyx]

Silly con valley got caught conning. 😂

[@editorrbr2107]

“Stealing from thieves.” That is framed perfectly.

[@ody9931]

Its already a big problem, just look at the KZbin model. Someone creates a piece of content, someone else reacts to that content and makes more money than the original. Then someone else compiles clips of the reaction content and repackages it again which commonly gets even more views than the react video now that it has gained traction. While the the reaction KZbinr is arguing with the clip tuber, the person that originally created the content gets the shaft. Its literally the backbone of this channel lol. Repurposing news stories and never giving credit to the original authors whos work u took and just reworded to avoid being a plagiarist

[@ahobimo732]

If you're stealing from someone BECAUSE they're a thief, then you're the hero, but if you either don't know or don't care that your victim is a thief, then you're just another criminal.

[@GypsyHarlow]

I think the issue isn't so much the data and intellectual property that was stolen to be the main issue of the concern. It's more about the beast itself (The actual ai) the motor that drives this thing. We're left with the genie being out of the bottle and I suppose where do my concerns may lie. Would I rather have America be leading the charge in ai technology, or China? Personally. Wish the damn thing never happened. But it's here, and I have more distrust of the Chinese using the tech than the Americans. But it is fun to seeing opeai squirm

[@cosmefulanito5052]

I think that's a fair argument. The technology is here, we can't go back to a state prior. We, as a society, are forever changed. Personally I distrust America far more than I distrust China so I guess, for the time being, this is an acceptable outcome.

[@CantoniaCustoms]

Because China is making it open source for one.

[@andreisopon4615]

Yeah, no-one's the good guy here, but at least America, while imperfect, is still a democracy. for now...

[@GypsyHarlow]

​@CantoniaCustoms For me that's not really the point. The way I see it. American tech is on the verge of a technological weapon that attacks on an economical level that at a scale unimaginable. So the question isn't about the ethics of stolen data. It's more of a race to create the cyber nuclear bomb, ai. Who's gonna finish their super weapon first.

[@Richforce1]

@@cosmefulanito5052 I for would rather trust the country that DOESN'T harvest the organs of religious minorities and use grape as a form of torture but I guess I'm just weird like that.

[@howardrobinson7245]

One of the Ten Commandments says You Should not steal.

[@soal159]

Problem with the Chinese models is that you can't sue them as the country is a one party state.

[@TheGahta]

What do those have to do with each other? Usa is formally a multi party state and gives as little shit to international law if they can get away with it 😂

[@fannyalbi9040]

DS is open source. It is technically no sole proprietary ownership. U can sue But maybe ban it in the country the government think it is "illegal"

[@soal159]

@@fannyalbi9040 I doubt this was for altruistic reasons. This was designed to disrupt US AI development and make the world more dependent on Chinese controlled tech. It is a poison pill.

[@dialy1]

To answer your opening philosophical question: You can only steal from a rightful owner.

[@cooperthehusky1956]

"Behind every great fotune lies a great crime" Honore de Balzac

[@SamSamSandwichman]

If the item isn't yours, stealing it, regardless of its current state of ownership, is morally wrong. Two wrongs don't make a right.

[@BEEETRUS]

Bro u have so many bots in ur comment section lol

[@SpottedHares]

How can you tell the difference from his normal viewers?

[@makesnosense6304]

Something stolen is when something is taken without the owner's permission. If the stolen item is taken/returned back to the owner, it's not stolen from the thieves. It's stolen by the thieves, but now taken back. If the stolen item is stolen again, by some other party, it's still theft. One can also say that the new party indirectly stole the item from the original owner. I don't understand why this would be complicated.

[@JSStuart100]

We know we know every single time

[@Gman979]

There is no American who care if the Magnificent 7 are being dethroned by an open source model. Most of them in fact are cheering and loving the FOSS instead. This is a perfect gift for the world.

[@BlankDriver]

I love the analogies in this video

[@CanalTremocos]

In medieval Europe, there was the concept of the outlaw. If a thief evaded justice anyone could steal from them, or commit arson on their property, or even go to their hideout and wet everyone's ear. An outlaw would literally be outside the protections given by the law. This is still a popular concept among internet commenters.

[@lk9650]

Why do people say that DeepSeek is probably lying about how much money they spent to train their model? It's open source which means anyone can try to replicate their results. There is no reason to speculate or jump to conclusions.

[@kevinlin4895]

It's not speculation.. or jumping to conclusions.. they believe because they WANT to believe

[@denekar]

It is a lie because in order to do what they have done from scratch you would have to create gpt, gemini, etc first.... and then use those to train DeepSeek.

[@_aullik]

The question about stealing in the beginning gets even more complicated when you do not actually "steal" as in remove something from someones possession, but you instead commit plagiarism.

[@rocketmunkey1]

Training data is a Euphemism for copy and paste, how else do we explain how a watermark present in only 0.000000001% of the "training data" finds its way into an outputted image of a face which they would have you believe Ai created from scratch ?

[@CedroCron]

Thank you for bringing this to light in such an illustrative way. I thought I was the only one that saw it this way!