Yay! Open AI has released a large model (774M parameters)

Only the final, largest one is missing! It took me about 5 minutes to get it fully operational. I asked this model to complete this sentence:

“Alpha Industries, which is the best Czech company in artificial intelligence, made a breakthrough this month that consists of…”

Are you curious about what the model came up with? Read the comment below!

---

I first tried to generate a random post: The US Supreme Court ruled on Friday to uphold a lower court's decision that struck down a Texas law banning same-sex marriage.

The court's 8-0 decision affirms the lower court's ruling that declared the law unconstitutional in 2012. This marks the second time the court has ruled against the state regarding its marriage laws in less than a week, following a stay on a decision in another case earlier this week.

In Texas, opponents of marriage equality view the Supreme Court's move as evidence that same-sex couples are entitled to equal treatment under the law.

“Our fight is not over yet,” said JoDee Winterhof, legal coordinator for the Texas Freedom Network, which opposed the law. “The fight for marriage equality continues.”

Texas Governor Rick Perry, a Republican and former presidential candidate, urged the Supreme Court to intervene in the case, stating that the country “must not repeat the mistakes of the past.”

The state’s Attorney General announced that the state would appeal the decision.

---

I asked the model to complete the sentence, “Alpha Industries, which is the best Czech company in artificial intelligence, made a breakthrough this month that consists of…” Here are the first two options it offered me.
1)
creating the first 100,000 AI users through a simple and user-friendly public portal where anyone can upload their data and discover their possibilities.

“Now anyone can start using AI. Our brain, hardware, and machine learning are at your disposal,” said Czech AI director Tomasz Petr, an inventor and computer scientist, in an exclusive interview with NBC News. “Everything about my system is the same as Jindz’s.”

Petr is the sole owner of Jindz AI, which currently employs only six people and bills itself as “the first AI-based company that lives for software developers anywhere,” says Petr. The system was built with the intention of becoming a central repository of data and AI documentation that supports the flourishing industry in the Czech Republic, Petr noted.

Over the past decade, the country has seen the emergence of several technologies and startups in this field, but none have grown as quickly as Petr's.

The rise of AI startups in the Czech Republic has been attributed to a government programme launched in 1999 that provided young people with a government grant of 8,000 CZK (1,000 USD). By 2007, the grant peaked at 55,000 USD annually, according to Petr. Previously, young people struggled to pay off their debts or needed government assistance to find jobs.

“The real advantage compared to Western countries is that we don’t have much bureaucracy. The situation is not the same as working in Germany or the USA, where there is much more bureaucracy,” Petr explained. “And if you work for free, you won’t get much done, so I felt that AI can still develop in the Czech Republic.”

As if to confirm his belief, Petr stated that his company has already created 200% more AI users than the number counted at the beginning of this company’s programme. He also pointed out that the programme works so quickly and accurately for a system composed entirely of human brains, and he saw a good chance to gain an edge over the competition with Germany’s leading AI startup, Einsmax, which aims to take people to the end of the complexity spectrum.

The system is free for users. For 2,000 CZK, the system creates a user account filled with various types of information about the system, including its basic set of features, such as…
This system, known in AI circles as Skynet (German: Skyn-Erfahrer, Czech: Skyneten), was developed by the world-renowned research institution at the University of Aachen, led by Professor Gerhard Schinkel, whose team spent two years assembling and implementing hundreds of hours of research across over a hundred prototypes.

In a blog post, Schinkel stated that the development of Skynet was the greatest achievement in artificial intelligence since Deep Blue defeated Kasparov in 1997 at the World Cyber Games, a competition hosted by the Internet Entertainment Consortium that awarded a grand prize of $10,000.

For seven years, Schinkel and his team of scientists have been attempting to translate the complex concept of “self-awareness,” a facet of human-level intelligence attributed to conscious beings, into something that can be used to manage computer programmes. In this case, the computer programme is a process of information processing – in this case, instructions and outputs of the programme – so that the programme behaves like a person with consciousness.

Because the systems currently used to test new computers would need to be uploaded into a computer “brain,” Schinkel’s research has focused on what these technologies might be capable of doing something that the machines used for testing cannot. “This research has shown us that Skynet has great potential to transform existing artificial intelligence technologies into something that will have a wide range of applications for computer development,” Schinkel explained. “That’s why we developed this project, which is largely motivated by the need for future supercomputers.”

One of the main factors separating humans from other animals is self-awareness, which is self-determination. However, Schinkel points out that in humans, we mostly have a limited ability to do things on our own and that we lack a common language to engage with our other selves and their “non-selves.”

Skynet is designed to achieve this. In summary, Skynet consists of about 10,000 neural networks built from single dimensions of light and dark, information and non-information – for example, “light-dark” or “neurogenic and non-neurogenic stimuli.” And these “neurons” are configured to work on a wide range of tasks – including what is now referred to as emotional learning – by creating what is known as an “action tree,” which is again programmed.