Open AI CTO Interview (Sora videos)

[Juan]

She dodged some questions and it gave me some sus vibes at times. I’m optimistic about AI stuff but at times I feel like they are not fully transparent of how these things works or that they are aware of stuff but don’t want us to know (like the possible dangers of AI, etc.) 

[Leo Gura]

Of course they themselves don't know the dangers of AI. And even if they knew they would still rush to monetize it. They are there to make money and be famous, not to save mankind.

Edited March 17 by Leo Gura

[Juan]

19 minutes ago, Leo Gura said:

Of course they themselves don't know the dangers of AI. And even if they knew they would still rush to monetize it. They are there to make money and be famous, not to save mankind.

I see, thanks for your feedback. 

[Juan]

The creators themselves.

[ExplorerMystic]

Another aspect is to keep AI as much as possible open source which these companies preached earlier.

https://www.bloomberg.com/news/newsletters/2024-03-15/openai-tumult-raises-question-of-how-open-an-ai-company-should-be

[Vrubel]

5 hours ago, Leo Gura said:

They are there to make money and be famous, not to save mankind.

They are there to save me a colossal amount of time.

Edited March 17 by Vrubel

[Seth]

7 hours ago, Leo Gura said:

Of course they themselves don't know the dangers of AI. And even if they knew they would still rush to monetize it. They are there to make money and be famous, not to save mankind.

Although not everyone agrees about the risks, safety is a huge issue in the AI community. It is talked about constantly. People also constantly complain about how much companies censor their models for safety (and mainstream appeal). It feels like you are assuming every billionaire or adjacent is evil now. What makes you so confident? 

 

[Leo Gura]

1 hour ago, Seth said:

Although not everyone agrees about the risks, safety is a huge issue in the AI community. It is talked about constantly. People also constantly complain about how much companies censor their models for safety (and mainstream appeal). It feels like you are assuming every billionaire or adjacent is evil now. What makes you so confident? 

In a sense, the fearmongering in the field of AI is a misdirect. The danger is not that AI will go Skynet and kill humans, the real danger is that these AI companies will become capitalist monopolistic monstrosities driven to maximize profits at great cost to society at large, driving ever larger income inequality that corrupts and preverts society.

I don't take their talk of safety seriously. Because the danger is not in the AI, the danger is in the corpo ghouls who run the company.

Edited March 17 by Leo Gura

[thenondualtankie]

I agree with Leo 100%, not that my opinion particularly matters.

AI has no survival agenda, it simply acts as it is trained. If the desire to destroy humanity isn't trained into the AI, it won't magically start doing it.

Unbridled capitalism is the biggest risk associated with AI.

[Girzo]

Current AIs are glorified databases, they do good only on tasks and questions to which there are some answers on the internet or pre-made examples it trained on. So they can correct your writing mistakes, as there are many examples of how correcting text works, but they can't solve a novel physics problem for you as they don't have anything akin to a model of how the external world works, only what operations can be made on text.

They would be much better databases if they had accessed to copyrighted scientific materials, like digital copies of Elsevier books or something. Then they would be able to give answers on for example complex medical topics, as they know and can recall a correct answer from the book, serving it in a manner that relates to your question.

I bet 0% on what OpenAI is currently doing becoming a self-acting agent. 

Edited March 17 by Girzo

[mmKay]

@Leo Gura what are some specific ways you envision these money hungry corporations exploiting AI ?

[Leo Gura]

23 hours ago, mmKay said:

@Leo Gura what are some specific ways you envision these money hungry corporations exploiting AI ?

I see a handful of massive companies like Google, Microsoft, Amazon, Apple, Tesla, etc using AI to become such massive monoplies that they buy up everyone and everything in the field of tech, sucking up tens of trillions of value while little people are left with crumbs.

That's the real danger.

Edited March 19 by Leo Gura

[Basman]

1 hour ago, Leo Gura said:

I see a handful of massive companies like Google, Microsoft, Amazon, Apple, Tesla, etc using AI to become such massive monoplies that they buy up everyone and everything in the field of tech, sucking up tens of trillions of value while little people are left with crumbs.

That's the real danger.

What kind of regulations or laws do you think would be ideal for combating this?

[Hello from Russia]

What I find fascinating is that these companies are already similar to small governments in a sense

They can build their own cities from scratch (and they actively do), they can have complex care plans for their employees, they educate you, they feed you, they take care of your children (if needed), they give you meaning, they finance you, they lift you up socially if you deserve it from their pov and etc

Edited March 21 by Hello from Russia

[thenondualtankie]

6 hours ago, Basman said:

What kind of regulations or laws do you think would be ideal for combating this?

I think getting them to open source the models would be a big step.

You'd still need the compute power to be able to run the models though. Which might end up being cheap enough in the end.

[Bobby_2021]

9 hours ago, Leo Gura said:

I see a handful of massive companies like Google, Microsoft, Amazon, Apple, Tesla, etc using AI to become such massive monoplies that they buy up everyone and everything in the field of tech, sucking up tens of trillions of value while little people are left with crumbs.

That's the real danger.

On the one hand you say the AI hype isn't real and isn't even worthy of attention.

And then you say there is real danger in how AI companies will eventually rule the world while simultaneously asserting AI is merely a hype.

Truth is that AI isn't nearly as scary as you think it is. No corporation will survive long enough to outlive a single human lifetime, unless they do some serious innovation and deliver value. 

And no, Apple or any big corporations isn't interested in buying your house and drain you by extracting rent. They can't just buy up everything and anything they want. That's simply not their business model. I have said this before that these corporations stay relevant by contributing disproportionate value into the economy. The doomsday scenario you mentioned is hard to come by even if you tried. This should be regarded with the same level of significance as some fringe leftists conspiracy theory.

Just uphold strong individual rights and most of these things fall into irrelevancy. You are taking up needless fears. There are real problems to solve, but you are not going to solve them while spreading all this fear. 

[Bobby_2021]

Most of these corporations will be dead even while you are still alive.

Only Apple and Microsoft will make it long enough. The rest will fade into irrelevancy faster than you think. They might be around, but not nearly as powerful. 

7 hours ago, Basman said:

What kind of regulations or laws do you think would be ideal for combating this?

- Strong individual protection laws.

- No communism in any form or shape whatsoever.

- Make the basics (Food & Shelter) of life accessible to all citizens. If not all, atleast to the most hardworking citizens.

Reward hardwork. Have meritocracy as much as possible in any given scenario. 

- Do not let government rig the open market. And don't have too many regulations. Especially for small business owners.

If any needs serious regulations, it should made exclusively for giant corporations. 

Edited March 19 by Bobby_2021

[lostingenosmaze]

 

[Leo Gura]

Obviously all these AI companies are at the very minimum scraping the entire public internet and every major social media platform for their data. And then on top of that they are probably accessing paywalled stuff.

[Danioover9000]

@Hello from Russia

On 19/03/2024 at 11:35 AM, Hello from Russia said:

What I find fascinating is that these companies are already similar to small governments in a sense

They can build their own cities from scratch (and they actively do), they can have complex care plans for their employees, they educate you, they feed you, they take care of your children (if needed), they give you meaning, they finance you, they lift you up socially if you deserve it from their pov

   Case in point Mitsubishi, Yamaha, Sony, and other Japanese large companies. It's more capitalism there and it's so integrated than the USA that toxic work culture, Neoliberalism, hyper conformity, along with multiculturalism and egalitarianism is leading to rapid decline of birthrates in westernized secular countries.