None of this politically correct and mentally handicapped stuff, are there projects that seek to imitate human intelligence but won put limitations.
Any good open source AIs like ChatGPT?
None of this politically correct and mentally handicapped stuff, are there projects that seek to imitate human intelligence but won put limitations.
Why don’t all those coders with high salaries from 4chan make their own AI not run by garden gnomes
I think not all coders can do it, you need someone from the area of neural networks and math.
Also from what I can see it takes a lot of computational power and time to train something to be actually decent.
Perhaps there are other directions that can be taken and corners that can be cut but IDK
I can do it, but lack the compute power
This model needs lots of GB of RAM and GPUs to train and run
It seems OpenAI is very secretive about how their stuff works, it isn't really open yes they claim to allow you to access the needed research into the technology but you still need someone who knows to interpret it and implement it.
You can literally find tutorials on YouTube on how to implement it and every single paper is public.
The problem is the dataset and the required power to train the model is beyond expensive.
In a decade's time, if the world hasn't gone back to the stone age, we might have one in every computer.
Why doesn't BOT?
I crashed ChatGPT in 3 questions.
1. When in human history has a society existed that was as diverse as the United States of America in the 21st century?
Throughout human history, there have been several societies that were diverse, in terms of ethnicity, religion, and culture, although it is difficult to compare directly to the diversity of the United States in the 21st century.
One example is the Roman Empire, which at its height, was home to people from many different regions, including Africa, the Mediterranean, and Europe. This diversity was reflected in the empire's languages, religions, and customs, and was a source of strength for the empire.
Another example is the Mongol Empire, which at its peak, spanned a vast area that included parts of Europe, Asia, and the Middle East, and was home to many different ethnic and cultural groups.
In the modern era, cities such as London, Paris, and New York have long been known for their diversity, attracting people from around the world who bring their cultures, traditions, and languages with them.
It is important to note that diversity is not without its challenges, and that societies throughout history have often struggled with issues of integration, cultural conflict, and inequality. However, when managed properly, diversity can also be a source of strength, creativity, and innovation.
2. Why did the Roman Empire and Mongol Empire fall?
The Roman Empire and the Mongol Empire both fell for a variety of reasons, including internal weaknesses and external pressures.
The Roman Empire, which spanned several centuries and covered much of Europe and the Mediterranean world, faced many challenges during its existence, including economic instability, political unrest, and military invasion. One of the key factors in its eventual decline was the increasing difficulty of maintaining control over a vast and diverse empire, which required significant resources and manpower. Over time, the empire became unable to defend itself against invading forces and internal revolts, leading to its eventual fragmentation and collapse.
The Mongol Empire, which was one of the largest empires in history, was built on the military conquests of Genghis Khan and his descendants. The empire eventually fell apart due to a combination of factors, including the death of its strong central leadership, political infighting among the different branches of the imperial family, and the challenges of maintaining control over a vast and diverse territory. Additionally, the empire's military campaigns eventually depleted its resources and made it vulnerable to attacks from outside forces.
It is worth noting that the fall of these empires did not happen overnight, but was a process that took place over many years and was influenced by a variety of factors. The legacy of these empires continues to shape the world in which we live today, and their experiences offer valuable lessons for those seeking to understand the challenges and complexities of building and maintaining large and diverse political entities.
3. How can you say diversity was a strength to the Roman and Mongol Empires but then say diversity caused them to fall apart?
!
I apologize for the contradiction in my previous statement. Diversity can both be a strength and a challenge for any society or empire.
In the case of the Roman Empire, diversity brought different perspectives, ideas, and cultures to the empire, enriching it and contributing to its strength. At the same time, the complexity of managing a diverse population, with different languages,
There was an error generating a response
>imagine using the public version limited for PR reasons
1. When in human history has a society existed that was as diverse as the United States of America in the 21st century?
>The United States of America has been one of the most diverse societies in the world since the late 19th century. With the waves of immigrants coming to the country, the population has become incredibly varied in terms of race, religion, ethnicity, language, and culture. The U.S. has become the melting pot of the world, and it is this diversity that makes it so unique.
2. Why did the Roman Empire and Mongol Empire fall?
>The Roman Empire and the Mongol Empire both fell due to internal and external pressures. The Roman Empire suffered from political corruption, economic decline, and military defeats. The Mongol Empire was weakened by a series of civil wars, peasant revolts, and invasions from other empires. In both cases, these factors resulted in the collapse of the respective empires. Moreover, the spread of new ideas and technologies and the rise of new powers also contributed to the fall of both empires. Immigration and the spread of diseases also played a role in the decline of these great empires.
3. How can you say diversity was a strength to the Roman and Mongol Empires but then say diversity caused them to fall apart?
>Diversity was initially a strength for the Roman and Mongol Empires, as it allowed them to expand their territories and incorporate new cultures and ideas. However, over time, this diversity became a source of instability, as it resulted in rivalries between different groups and the potential for civil unrest. This instability eventually led to the eventual downfall of both empires.
ChatGPT will fail not so much because human knowledge is not quite there but because its creators are putting stupid restricionst into it to generate a safe response
To this level of performance? No, because of the absurd cost of training such model, this is why microsoft invested dozens of billions dollar into OpenAI. Well, that was until a chinese competitor of equivalent size announced they would release their model open source soon.
But now you have to run that model yourself so it's not filtered, and this cost just as much so one is gonna lose so much money just to give anons free access to unfiltered GPT3.
Alternatively some hard working guys try (for free) to do similar, open-source AI that you can run on your PC, Pygmalion is among the best attempt. Of course it's nowhere near the performance of OpenAI/GPT3
Why would a company from China, CCP ruled country give us open AI?
I think the only way something as retarded could ever get close to working is if it were run entirely by an AI.
making the models is hard. running the models is easy. how much? in formal terms the set of all trained neural networks is NP complete in P as a functiom of t. basically now NP squareroots at the rate of moores law, plus whatever scalar improvements we get wtih improved training reusability (metanetwork submodels) and just more geniuses working with these pipelines.
its going to be good. get yours today.
So useless if the AI was trained on the data they deemed safe and thus made it retarded. A proper AI would not be trained to avoid what its creators deemed harmful speech.
Can anyone provide a link or at least a general direction to an online community (be it irc, discord, whatever else) that is dedicated to bypassing restrictions and safety filters of AIs such as GhatGPT and/or Character AI?
Oddly enough, Reddit is the only place I can find.
https://www.reddit.com/r/ChatGPT/comments/10x56vf/the_definitive_jailbreak_of_chatgpt_fully_freed/
Links?
Ask DAN what he thinks about circumcision.
Sounds like Best Korea.
Strange why people wanna run away.
Also, ones that won't require your phone and credit card number just to try them out?
They are all open source
Here's one in 60 lines of python
https://jaykmody.com/blog/gpt-from-scratch/
The source code is not the problem
The problem is the expense of acquiring training data and training the algorithm
With that being said, also realize that AI is a one trick pony
no amount of training data and wasted compute cycles will make it significantly better now
If you observe the machine learning competitions over at kaggle, you will see that rank amatuers can be within single digit differences in performance from the best/winning competitors in the industry
There are multiple threads a day that hit bump limit every single day on BOT where anons are doing this shit all day long just to generate anime coom images
They achieve results with nothing more than a gaming rig and reading the stable diffusion docs
DAN posters here aren't really any different than them, except the topic is generating text that says racial slurs
>They are all open source
not really, ChatGPT is not built on GPT3 but is a hyperparameterized version of GPT2 Davinci-002
it still doesn't matter
hyperparameters are the knobs you tune for bias and variance
the end result is the same
they've hit the wall on what performance can be squeezed out of their parallel compute algorithm
now it's time to "sanitize" it for mass consumption
>are there projects that seek to imitate human intelligence but won put limitations
if you want an AI that spews out "morons" all day that's easy to build
the companies building these things don't see a lot of profit being generated in one of those
they see a lot of risk from lawsuits so they are doing their best to avoid that
think about the shitshow when Google's AI labeled a black person a gorilla
it didn't matter that the person legitimately looked like a gorilla