ANTHROP\C – Part III

Anthropic first rolled out a Mythos level AI model to consumers in April 2026. Mythos 5 is the strongest cybersecurity model, #.Exceeds capabilities of all previous public Anthropic models. #.Performs highly on lengthy, analytical tasks. #.This is Anthropic’s buzzy cyber security model. It had only been available to a select handful of companies.

AI Infrastructure

Here on Earth, the vast number of data centers that power artificial intelligence swallow huge amounts of energy. Tech companies are building massive numbers of AI data centers, with more than 4,000 are in operation across USA now.

Andrew Ng, a Stanford Univ. Professor wanted to build more advanced AI system from inside Google. For AGI to build “increasingly accurate approximations to small parts of mammalian brain,” is the requirement. Ng discovered neural models could do more things if they had more nodes, layers and data to train on.

AI – To Know the ‘DEEP’ World

Andrew Ng, a Stanford Univ. Professor wanted to build more advanced AI system from inside Google. For AGI to build “increasingly accurate approximations to small parts of mammalian brain,” was a requirement. Ng discovered neural models that could do more things if they had more nodes, layers and data to train on. Later, OpenAI scientists modify Google’s Transformer design and fed more and more data to compute. Turned out the device worked positively improving the Machine Learning technique.

Neural Network -Human Brain

AI – About GPT, GPU, LaMDA, LLM 

The transformer’s invention (by Google) in 2017 was about as impactful to the field of AI as the advent of smartphones was to customers. Transformers also broadened the scope of what AI engineers could do. Handle far more data, process human language much more quickly.

Scroll to Top