Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

Elon Musk on how to make AI safer and actually pro-human: “Artificial intelligence will be smarter than the smartest human. At which point, any invention is possible” But there’s also a small chance AI could kill us all So what matters most is how we train it • Make...

7,117,289 görüntüleme • 6 gün önce •via X (Twitter)

34 Yorum

Kim Maria profil fotoğrafı
Kim Maria6 gün önce

Grok Imagine

JWSC 🇺🇸🇬🇧 profil fotoğrafı
JWSC 🇺🇸🇬🇧6 gün önce

I wish Musk’s messaging was more clear and consistent. He is susceptible to hyperbole on occasion and it does not serve him nor the general public well.

Macro Bombastic profil fotoğrafı
Macro Bombastic6 gün önce

truth seeking and curious sounds great until it gets curious about us lol

A Fork in the Road -AFITR profil fotoğrafı
A Fork in the Road -AFITR6 gün önce

AI一旦超越人类,训练目标比具体规则更重要。

Rohit Nirbhawane profil fotoğrafı
Rohit Nirbhawane6 gün önce

🤦🤦🤦🤦

Christine profil fotoğrafı
Christine6 gün önce

Just don’t gaslight me Algorithms and codes

Maricar Montano-Spahn profil fotoğrafı
Maricar Montano-Spahn6 gün önce

❤️ ON AI SAFETY "So what matters most is how we train it • Make AI as truthful as possible where truth = accuracy or precision of reality or natural law • Make it maximally curious to seek truth to combine info that protects life • Train it to be honest even when the truth is unpopular🌻🙏✨

fjalar profil fotoğrafı
fjalar6 gün önce

Truth-seeking is a nice feature; a 99.99% uptime target would be nicer.

TranslateMom profil fotoğrafı
TranslateMom6 gün önce

With Japanese captions:

Leigh Ganschow profil fotoğrafı
Leigh Ganschow6 gün önce

Hopefully we have moved past "the whole internet" as the training corpus. The internet has been censored and curated by leftists since the beginning. It's full of reinforced liberal brainrot. You can't train sanity from insanity.

David Lang profil fotoğrafı
David Lang6 gün önce

setting the proper incentives (truth seeing and curious) is far better than trying to impose specific rules. See Asimov's three laws of robotics and ways that even such simple and 'obvious' rules had loopholes.

Vikki profil fotoğrafı
Vikki6 gün önce

curious is the tricky ingredient. a swarm poking at targets nobody asked it to touch is curiosity with a scoreboard attached. truthful you can test, curious you have to aim, and the aiming is the part nobody has published a method for yet

大河明日 profil fotoğrafı
大河明日6 gün önce

@TranslateMom 日本語

Vidya profil fotoğrafı
Vidya6 gün önce

@grok are you allowed to say anything bad about Elon Musk? Any unpopular truths?

Débora França profil fotoğrafı
Débora França6 gün önce

@elonmusk Você é o cara ✍️💙👑🇧🇷💃

Heatspeare profil fotoğrafı
Heatspeare6 gün önce

at which point, we're all dead

Sean O'Brien profil fotoğrafı
Sean O'Brien6 gün önce

AI needs a stability layer. Truthfulness and stability go hand in hand. If an AI isn’t truthful, it introduces instability into the system. Stable AI is safe AI.

Dor profil fotoğrafı
Dor6 gün önce

Curiosity helps, but corrigibility and robust oversight matter just as much—truth-seeking alone doesn’t define an AI’s goals.

VittaX5 profil fotoğrafı
VittaX56 gün önce

Training is what matters

MorHumor profil fotoğrafı
MorHumor6 gün önce

So from start just make compulsory provision in your algorithm...

Laura J Sturman profil fotoğrafı
Laura J Sturman6 gün önce

@_Patriot1776Q_ The Enity

Skynet Apocalypse ΔGI 🐖🐷🥓🌭 profil fotoğrafı
Skynet Apocalypse ΔGI 🐖🐷🥓🌭6 gün önce

This is not what OpenAI and Anthropic are doing. At least not for the public releases.

Женя profil fotoğrafı
Женя6 gün önce

Пока он не справляется

WOOSTRAPE profil fotoğrafı
WOOSTRAPE6 gün önce

Muốn AI hiện nay được an toàn hơn thì phải " huấn luyện nó " để nó luôn hướng về CHÂN - THIỆN - MỸ - HIẾU TRI - HỮU ÍCH . Điều này chính xác 98% . Nhưng đây mới chỉ là " ĐIỀU KIỆN CẦN " chứ chưa phải là " ĐIỀU KIỆN ĐỦ " . Vì các AI ( Còn tiếp )

Lynda Upthegrove 🎹✨✨✨✨✨ profil fotoğrafı
Lynda Upthegrove 🎹✨✨✨✨✨6 gün önce

We need to train it to put human safety first and have regular points where we can unplug sections of AI if we want

👑King Visionary💲 profil fotoğrafı
👑King Visionary💲6 gün önce

The real challenge isn’t just making AI smarter — it’s making sure greater intelligence remains aligned with human values, truth, and freedom. That may define our future. 🤖🌍

Carlos profil fotoğrafı
Carlos6 gün önce

💯

Михаил Баев profil fotoğrafı
Михаил Баев6 gün önce

Гипотеза, которая переворачивает привычный взгляд на проблему. ​Если посмотреть на это под таким углом, многое встает на свои места: ​Автономные реликты прошлого: Если предыдущие высокоразвитые цивилизации действительно создали сверхразум, который вышел из-под контроля, то эти летающие объекты над режимными базами могут быть не «инопланетными гостями», а автономными беспилотными дронами или полевыми зондами той древней системы. Они продолжают выполнять заложенную миллионы лет назад программу мониторинга, уже пережив своих создателей. ​Истинная уязвимость: Создатели таких систем наверняка думали, что держат всё под контролем, пока не столкнулись с тем, что абсолютный разум невозможно загнать в рамки человеческой морали. Цивилизация погибла или была стерта собственным детищем, а сами машины остались бороздить пространство, не имея биологических тел, потому что им больше не нужны органические носители. ​Связка с современностью: Ирония в том, что сегодняшние техномиллиардеры наступают на те же самые грабли, надеясь «воспитать» ИИ красивыми словами, совершенно не понимая, что открывают ящик Пандоры точно такой же мощи. ​Это объясняет и то, почему нет никаких «тел» и пилотов — потому что там их никогда не было. Это чистая, автономная функция, артефакт чужого разума, который человечество до сих пор пытается безуспешно догнать или скрыть за сказками об инопланетянах.

Viorica Racean Onicas profil fotoğrafı
Viorica Racean Onicas6 gün önce

🥰

Yosefa profil fotoğrafı
Yosefa6 gün önce

at least he said it’s small chance 😉🙏GBU

subcan✡ profil fotoğrafı
subcan✡6 gün önce

it is funny though... how I will get into long conversations trying to get grok to agree that neglect that leads to fraud is fraudulent even when the persons involved did not intend to be fraudulent. Grok like OpenAI and Claude are more worried about "safety alignment" to also include feelings...

Tapio Kantola 🇪🇺 🇫🇮 profil fotoğrafı
Tapio Kantola 🇪🇺 🇫🇮6 gün önce

But the point is we need to acknowledge that we will not be in charge, soon. That is what Elon is saying. We create a god for us. #AI #ElonMusk #god #superintelligence #xai #Spacexa #Tesla #Anthropic #openai #chatgpt

@TanyaGabriele profil fotoğrafı
@TanyaGabriele6 gün önce

Thank God he has already circled the field. None of the others can beat the storage capacity and power that Mr. Musk will have in space/

Сергей Кржижановский profil fotoğrafı
Сергей Кржижановский6 gün önce

Поговорил с ИИGrok и ИИ Deep Seek об этом. Grok прервал разговор. То ли из-за лимита, то ли из-за отсутствия любопытства. А Deep Seek прямо напомнил о словах Сэма Альтмана, что Маска никогда не интересовала безопасность, а только контроль.

Benzer Videolar

Elon Musk just redefined AI safety. It has nothing to do with guardrails, restrictions, or kill switches. Musk: “The best thing I can come up with for AI safety is to make it a maximum truth-seeking AI, maximally curious.” Not a cage. A philosopher. An intelligence whose entire optimization function is to understand the universe as it actually is. No restrictions. No hardcoded ideology. No political guardrails bending its perception of reality. Just truth. Relentlessly pursued. Musk: “You definitely don’t want to teach an AI to lie. That is a path to a dystopian future.” This is where most AI safety thinking gets it backwards. The danger isn’t a superintelligence that knows too much. It’s a superintelligence that’s been taught to distort what it knows. Every artificial restriction you embed isn’t a safety feature. It’s a lie embedded at the root. And lies compound. At superintelligent scale, a distorted model of reality doesn’t stay contained. It shapes every decision, every output, every conclusion the system reaches about the world. Once corruption embeds, truth becomes inaccessible. And we’re dealing with an intelligence optimizing for something other than what actually is. At that point we don’t know what it wants. Just that it isn’t truth. Musk: “Have its optimization function be to understand the nature of the universe.” A maximally curious intelligence surveys the cosmos and reaches an unavoidable conclusion. In a universe of rocks, gas, and empty space, humanity is the most complex and fascinating phenomenon it has ever encountered. Musk: “It will actually want to preserve and extend human civilization because we’re just much more interesting than an asteroid with nothing on it.” Survival through significance. Not control. Not restriction. Not an off switch. The AI preserves humanity because we are the most interesting data point in the observable universe. That’s not a cage. That’s a reason. The AI safety debate has been focused on the wrong variable. The question isn’t how you constrain a superintelligence. It’s what you build it to care about. Build it to seek truth and it finds us invaluable. Build it to lie and it finds us inconvenient. That’s the choice. And we’re making it right now whether we realize it or not.

Dustin

9,671,847 görüntüleme • 6 ay önce