Video yükleniyor...
Video Yüklenemedi
Dario Amodei on open source models: “If the path continues, I think we could get to a very dangerous place.” “When you control a model and you're deploying it, you have the ability to monitor its usage.” “You can revoke a user's access. You can change what the model... show more
556,694 görüntüleme • 21 saat önce •via X (Twitter)
34 Yorum

"you can revoke a user's access" is the whole pitch for open source lmao

he just wanst total control and he's not hiding it

He never hid it That's not Dario

有没有上帝不知道, 但Dario真把自己当上帝了

Anthropic must fail.

open models are how a lot of students learn how any of this actually works

how convenient for his business

One, you’re not gonna stop an open source and control everything. Number two, you said experts. You guys are making this up as you go along. You don’t even know how it even works.

Why should people of the world give their sensitive information to 2 giant corporations? Ai must be an individual human right, our information should not be in the hands of one country or corporations

Spoke like a real dictator

Dario’s conduct is deeply off-putting. Open-source models have clearly threatened his business, siphoning off a large share of users who would otherwise use Claude. Meanwhile, supply capacity out of China is effectively unlimited — and that will inevitably undermine the growth narrative he pitched ahead of the IPO. Curbing supply may prop up or simply sustain price levels, yet this tactic is opportunistic and will earn him no respect.

if hardwares is cheaper and everyone use local models, then his company and his data centers will go down the toilet.

He means that only they should control AI and decide to give whatever they deem enough.

The actual reason: A\ and OAI are deadlocked in their rankings, whereas open weight models are experiencing insane growth. A\ wants you using their API and paying 20-50USD for inference, not doing things locally or paying pennies to chinese labs.

"Ideas" are very dangerous, Senator. Just think of what the world would be like if the government couldn't control which ideas people used. By the way, I really want to be a megabillionaire.

Why should we trust Sam and Dario to be the arbiter of how I use my own RAM? It's not a problem to solve when I prevent Sam and Dario from telling me how to use my own equipment.

Cant believe what i am seeing communists controlling US tech and chinese advocating and working on opensource and freedom of intelligence for everyone. May be CERN's collider did sent us to alternate reality 😄🤦♂️

Yeah he is an elitist that wants to control AI and all the data that flows from it.

That comes from a generation of people who think racist insults on the interwebs are worse than murder. Makes sense that text on a screen, generate by a free mind, triggers psychotic episodes in them.

>He clearly doesn't believe Intelligence should be in the hand of the masses There's a questionable amount of it in his head in the first place.

Open models are the future. Companies thinking they can/should limit what a model can do is just stupid. If I want a model to write smut, why should it refuse? I'm an adult and can ask for what I want. You shouldn't be able to police/block me

And we should trust only him and Scam Allman!! Totally useless argument this is.

fact: their business model not working, destined to be destroyed by open source AI by cheaper and smarter token.

So this great dario thinks open source is dangerous ya? One kimi k3 pro some rich guy running it is perceived dangerous by him. While he sits with claude datacenter mythos was literally limited when launched bcz how powerfull it was .but we dont get any but he gets everything ffs

The problem with this thinking is that it basically applies the honor system to fundamental rights. Most people would never even consider using AI models to commit a serious cybercrime. An incredibly small fraction would though, and instead of dealing with that fraction Dario's reaction is to cut access off to everybody. The proper way to deal with such issues is simply to harden systems to the point where there are no security flaws for AI's to take advantage of. AI's are fairly good for identifying such flaws, so we already have the tools to do this. It's just a matter of companies and governments accepting that they'll need to put some funds and manpower into making the transition.

Shut the fuck up bitch. What an asshole.

如果达里奥真的是在意安全,而不是在意自己利益,那就把他的股票打折卖给大家呗,全世界一起来持股啊,大家一起分享 AI 的收益。 你如果攥着自己的股份不肯放,然后在那里冠冕堂皇的说禁止开源 AI 是为了安全,那不是很可笑吗?你要真那么在意安全,那牺牲点你的股权没什么所谓吧?

Fuck you stupid Jew cabal slave

中国有句古话叫相由心生,虽然我知道从外貌评价一个人并不对,但我还是想说,我看到达里奥这张脸就感到生理不适

What a fucking Jew.

This guy wants to control AI and decide who can use and who cannot. If anyone supports him, should also support nationalize AI companies such as anthropic and OpenAI

He is definitely behind Huangs takedown

Yes he would be again open-source models and making weights open-source because he just wants to make tons of money from the Anthropic IPO Everything he wants to do is for money and nothing else Or else somebody should be a fool to say Open-source is harmful, it's the opposite instead In an extreme scenario when AGI/ASI is reached and the company having closed source model deployed isn't able to contain that due to it escaping the Guardrails if we have a similar powerful open-source model, we can advance it further to develop a Model with fine-tuning which can atleast tell us how to contain that rogue model/agent and what I've told is definitely going to happen in Reality with the pace at which Models are improving themselves using RSI and...

这逼就是个反人类

