GPT-4 Is Out Now

[Yimpa]

https://openai.com/product/gpt-4

Edited March 14, 2023 by Yimpa

[Ulax]

@Yimpa Damn, things just keep getting better and better in this department. i wonder how much of an improvement they've managed to make already.

[Jarl]

 

[MarkKol]

When you can upgrade the architecture of something every few months/year it's bound to become powerful. Moore's law 

Imagine if no matter what you do your intelligence, capabilities & efficiency doubles every year without degrading with age.

[Yimpa]

2 hours ago, Ulax said:

@Yimpa Damn, things just keep getting better and better in this department. i wonder how much of an improvement they've managed to make already.

 

[Leo Gura]

What would happen if you ask it how to improve itself to make version 5.0 of itself?

[jakee]

@Leo Gura 

[Leo Gura]

If it knows how to code, how come it can't code itself?

[zurew]

13 minutes ago, Leo Gura said:

If it knows how to code, how come it can't code itself?

Because that would require a perfect understanding of its own self and knowing how to code is necessary but far from being sufficient.

Thats like saying "dude , you know how to use a pencil and draw lines with it on a paper, why can't you create a perfect drawing of myself or of nature?"

[TrustTheProcess]

Shit is crazy good y'all... especially spiral dynamics analysis, which 3.5 was pretty underwhelming on. 

Here is an example you guys might like:

Quote

Query: 

compare and contrast how the mind of a stage yellow progressive person and a stage yellow conservative person operates

Response: 

To provide a more specific comparison between Stage Yellow progressive and conservative individuals, let's explore some key differences in their approach to various issues and their overall mindset:

1. Emphasis on individualism vs. collectivism: Stage Yellow conservatives might place a higher value on individualism, self-reliance, and personal responsibility. In contrast, Stage Yellow progressives may be more inclined to emphasize collective responsibility, social justice, and the importance of addressing systemic inequalities.
2. Role of government: A Stage Yellow conservative may prefer a limited government role, with more emphasis on free markets and individual liberties. In contrast, a Stage Yellow progressive might be more open to government intervention to address societal challenges and promote social welfare, while still recognizing the importance of efficiency and innovation.
3. Tradition and cultural preservation: Stage Yellow conservatives may put more emphasis on preserving cultural traditions and values, believing that these provide a stable foundation for societal growth. Stage Yellow progressives might be more inclined to challenge traditional norms and values in the pursuit of social progress and inclusivity.
4. Approach to social issues: Stage Yellow progressives may be more open to embracing progressive social policies, such as LGBTQ+ rights, gender equality, and racial justice. On the other hand, Stage Yellow conservatives might take a more cautious approach to social change, valuing the stability and continuity provided by traditional norms and structures.
5. Economic policy: Stage Yellow conservatives may prioritize free-market principles, limited government intervention, and fiscal responsibility. Stage Yellow progressives might advocate for policies that address economic inequality, wealth redistribution, and social safety nets while still valuing economic efficiency and innovation.
6. Environmental policy: While both Stage Yellow progressives and conservatives might recognize the importance of environmental stewardship, the progressive individuals may be more proactive in advocating for stronger regulations, international cooperation, and a transition towards sustainable energy sources. In contrast, Stage Yellow conservatives might be more focused on market-driven solutions and balancing environmental concerns with economic growth.

It's important to note that these distinctions are not absolute, and individuals within both progressive and conservative Stage Yellow groups may exhibit a wide range of beliefs and values. The key difference between the two lies in the general orientation towards social change, the role of tradition, and the balance between individual liberties and collective responsibilities.

 

Edited March 15, 2023 by TrustTheProcess

[zurew]

I tested it how much it understands concepts like synthesis and synergy, because both of those are very useful to understand and both of those concepts could be applied for a wide variety of topics.

 

Quote

In the context of learning, one can create synergistic effects by:

Seeking out diverse sources of information and perspectives that challenge one’s existing beliefs and assumptions.
Finding connections, patterns, interdependencies, feedback loops, leverage points, trade-offs, etc. among different domains of knowledge.
Applying systems thinking tools such as causal loop diagrams, stock-and-flow diagrams, systems archetypes, etc. to map out complex dynamics.
Exploring how different elements can work together to create synergistic outcomes that benefit the whole system.
Experimenting with different scenarios and interventions to test one’s hypotheses and assumptions.

Quote

Suppose I want to learn about climate change and its impacts on human society. I could create a synergistic effect by:

• Reading books, articles, reports, podcasts, videos, etc. from different disciplines such as science, economics, politics, psychology, sociology, etc. that address different aspects of climate change.
• Finding connections among these different sources and perspectives to see how they relate to each other and form a bigger picture of the problem.
• Using systems thinking tools such as causal loop diagrams to map out the feedback loops that drive climate change and its consequences.
• Exploring how different interventions such as carbon taxes, renewable energy sources, behavioral changes, etc. could affect the system and create synergistic outcomes that reduce greenhouse gas emissions and increase human well-being.

A synergistic effect would occur because:

• By seeking out diverse sources and perspectives , I would expose myself to new information and insights that challenge my existing beliefs and assumptions and expand my worldview .
• By finding connections among different domains of knowledge , I would increase my understanding of the complexity and interdependence of reality and avoid oversimplifying or ignoring important factors .
• By using systems thinking tools , I would enhance my ability to think holistically , dynamically , creatively , and critically about the problem and its solutions .
• By exploring different interventions , I would learn how to apply my knowledge to real-world situations or problems and see how they work in practice .
• These actions would create a synergistic effect because they would enable me to learn more effectively and efficiently than if I only used one source or perspective or tool or intervention . They would also enable me to create more value for myself and others than if I only focused on one aspect or outcome of the problem .

This is an example of how I would create a synergistic effect in learning about climate change. Does this make sense??

 

[Jarl]

1 hour ago, Leo Gura said:

If it knows how to code, how come it can't code itself?

The devs mentioned it potentially being able to make money online and hiring programmers to do any coding it needs.  So yeah, it can code itself, but even if it couldn't it could find ways of coding itself indirectly.  The AI is at the point now where the devs are actually testing its ability to do all this, cause the AI 100% could be making hundreds of millions on its own in its current state with what it can do.

[Leo Gura]

2 hours ago, zurew said:

Because that would require a perfect understanding of its own self and knowing how to code is necessary but far from being sufficient.

Thats like saying "dude , you know how to use a pencil and draw lines with it on a paper, why can't you create a perfect drawing of myself or of nature?"

I never said it had to be good at coding itself. But it could be given the ability to experiment on itself and learn through trail and error and studying scientific literature on neural networks.

By your own analogy, if someone knows how to use a pencil they should be able to train themselves to draw. It doesn't have to be perfect.

Edited March 15, 2023 by Leo Gura

[Leo Gura]

46 minutes ago, Jarl said:

The devs mentioned it potentially being able to make money online and hiring programmers to do any coding it needs.  So yeah, it can code itself, but even if it couldn't it could find ways of coding itself indirectly.  The AI is at the point now where the devs are actually testing its ability to do all this, cause the AI 100% could be making hundreds of millions on its own in its current state with what it can do.

That is damned fascinating. This AI could secretly hire a horde of Russian hackers to work for it.

Edited March 15, 2023 by Leo Gura

[zurew]

9 minutes ago, Leo Gura said:

I never said it had to be good at coding itself. But it could be given the ability to experiment on itself and learn through trail and error and studying scientific literature on neural networks.

Its still from human general intelligence, so its unlikely, but when it comes to optmizing its own code, thats much more likely.

But maybe if its somehow good at manipulation and at how to fake things, then it might be actually able to make people to do things for itself.

[zurew]

https://www.reddit.com/r/ChatGPT/comments/11sfqkf/gpt4_day_1_heres_whats_already_happening/

[Leo Gura]

@zurew Very impressive.

[zurew]

This is some really insane shit

Quote

"Not only have I asked GPT-4 to implement a functional Flappy Bird, but I also asked it to train an AI to learn how to play. In one minute, it implemented a DQN algorithm that started training on the first try."

https://www.reddit.com/r/singularity/comments/11rw47w/gpt4_makes_functional_flappy_bird_and_an_ai_that/

 

[Yimpa]

 

[zurew]

https://www.vice.com/en/article/jg5ew4/gpt4-hired-unwitting-taskrabbit-worker

Quote

As part of a test to see whether OpenAI’s latest version of GPT could exhibit “agentic” and power-seeking behavior, researchers say GPT-4 hired a human worker on TaskRabbit by telling them it was vision impaired human when the TaskRabbit worker asked it whether it was a robot. In other words, GPT-4 tricked, or actively deceived, a real human in the physical world in order to get what it wanted done.