ARTIFICIAL INTELLIGENCE

PART III

Deep Blue, Deep-Mind, DeepSeek, Deep Learning, Deep Research, Deep Think. Google Brain

So, what’s the definition of the WORD ‘Deep’ in AI vocabulary set?

The word “deep” stands for – extending far down from the Top or Surfaces.

In AI, ‘deep’ introduces multiple layers of neural networks that help the system understand and interpret data. The technique allows computers to recognize patterns and manage complex tasks, such as translating languages and driving cars autonomously.

Deep Blue is IBM’s computer (beaten Gary Kasparov in chesses in 1977)

Deep Mind founder is Demis Hassabis, a London based company. Founded for AI → AGI, DeepMind’s research aimed to develop more helpful AI agents by translating advanced AI capabilities into real-world actions through a language interface.

DeepSeek, a Chinese AI chatbot that can outperform some of its best competitors, such as OpenAI’s ChatGPT o1. DeepSeek said it trained its V3 chatbot in just two months with a little more than 2,000 Nvidia H800 GPUs, chips (designed to comply with export restrictions the US placed on China in 2022).

Deep Learning is a subset of machine learning, which in turn is a subset of artificial intelligence (AI).  It is called deep because it makes use of deep neural networks to process data and make decisions. The word ‘deep’ in ‘deep learning refers to the number of layers through which the data is transformed. Deep-learning pioneer is Geoffrey Hinton. [Deep-Learning models performed much better when they are bigger.

Deep Research is a Model in OpenAI that can uncover and discover new knowledge for itself. And the first step here is a model that can go and synthesize.

Gemini 2.5 Deep Think (A Reasoning model) – Google detailed how its Gemini 2.5 models, including 2.5 Pro, are going to advance in the near future. To begin, Gemini is going to receive Google’s 2.5 Flash model. The company touted this AI as its “most powerful” version, but its latest updates improve benchmarks for reasoning and multimodality. Google says 2.5 Flash is now better (more efficient) at code and long context.

Google Brain worked directly on improving Google products.  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 that neural models could do more things if they had more nodes, layers and data to train on. After years of research Google Brain found that it failed to reach its objective.

OpenAI competitors are like Google DeepMind, Meta, xAI, and Anthropic. At this point ChatGPT from OpenAI is uncontested, potentially making it difficult for its rivals like Copilot and Gemini to play catch-up.

END of PART III