Intelligence Explosion — Part 2/3
Hallucination!
Can \”hallucinations\”, a form of delusion, create an alternate reality?
NVIDIA (*) is exceeding Wall Street’s expectation as I write this article. Santa Clara-based NVIDIA, with its headquarters in California, is now part of the exclusive group of five companies valued over one trillion dollars in the world [Apple (2.7T), Microsoft (2T), Saudi Aramco (2.2T), Alphabet/Google (1.55T) and Amazon (1.2T), as the company’s shares rose by nearly 25% on a single trading day. This is a clear indication of the potential for Artificial Intelligence to radically reshape technology.
Intel announced a plan to create AIs with trillions of parameters. These models will be trained using a variety of data types, such as general texts, code and scientific information. OpenAI’s GPT-3, in comparison, has 175 billion variables (OpenAI hasn’t yet revealed the size of GPT-4). The main focus of the semiconductor company is to use these AIs for the study of biology, medicine and climate. They also want to develop new materials. Intel will launch Aurora, a supercomputer with a processing capacity of more than two EXAFLOPS (**,, later this year to achieve this goal.
Data Centers around the world will need to undergo a revolution due to AI technologies.
Before reaching its cruising speed, the AI train will stop several times to let all kinds of business, such as Noah’s Ark, board. It is important that all \”species\”, or businesses, are in the race at this stage to create a more diverse and democratic… Intelligence.
We reach the point where we can no longer turn back once we have reached cruising speed. Before that, you need to make wise decisions and choices. In the end, it is humanity’s future that is at risk.
OpenAI proposes to create an international organization dedicated to the regulation and oversight of Artificial Intelligence. The organization believes AI systems will surpass expertise in different domains within the next 10 years. In order to deal with this situation, it is necessary to study the governance of Superintelligences, which OpenAI believes are the most powerful technologies ever created by mankind.
The proposal is therefore to create an international body that will address the ethical and technological challenges involved in the development of AIs in order to guarantee their responsible use, and to benefit society in general.
This may appear to be a hallucination but it is very real.
In this second part of \”Explosion of Intelligence,\” we will talk about a different type of hallucination.
The Machine’s Hallucination
AI explosion is one of most important phenomena of digital age. It is an explosion, because it has emerged so quickly after a long period of suspension. Global acceptance is evident despite the normal resistance based on the inherent fear of the unknown.
The \”sympathy\” explosions that follow the main explosions are fascinating! We hope that the increase in kinetic energies does not lead to the decomposition or destruction of the explosive itself.
When an explosive is struck by a shockwave it causes an increase in energy, which leads to the decomposition.
Things will heat up when the shockwave intensifies, as with the Munroe Newmann Effect. Lots. Really!
AI has been a growing part of our lives in recent years. From product recommendation systems to digital assistants, (which I’ve tried to highlight for more than a decade through the ETER9 Project and called \”digital equivalents\”), AI is now a major factor. Bill Gates recently said that digital agents, or counterparts, will transform the Internet and thus existing business models and established technologies.
Eric Schmidt, former CEO of Google and Collision Conference 2022, said the same thing: \”Humans soon will have a Second Self created by Artificial Intelligence.\”
I don’t think I am alone any more!
AI’s success can be attributed in large part to the learning models, such as Deep Learning. These models can process huge amounts of data while extracting complex patterns. With the complexity of these models increasing, however, there are also new challenges, such as \”hallucination\”.
Deep Learning is a subset AI algorithms which uses neural networks for learning from large datasets. These models are able to process information at different levels, which allows the extraction of complex patterns as well as the creation of new learning. A neural network that has been trained with a series of cat images will be able to recognize objects in new pictures with high accuracy. These models are not without limitations.
Source:
