While in chess, after the first 2 moves, there are 400 possible positions, but in Go, after the first 2 moves, there are over 129,000 positions. The number of games of Go is so many that it is more than all the atoms in the known universe combined, and until 2016, everyone thought that a machine would never be able to conquer this feat.
Google DeepMind, Alpha Go went against Lee Sedol and came out the victor. This was the start of the AI revolution in the World. Elon Musk and Peter Thiel decided that if they did not do anything concrete urgently, Google would take control of the AI industry completely.
Elon Musk and Peter Thiel decided on a Dream team. One which was set to destroy Google’s dominance over the AI market. The Team consisted of Sam Altman, Elon Musk, Greg Brockman, Ilya Sutskever and Peter Thiel. They pledged 1 billion for a non-profit venture, which had no shareholders, no profit motive and no corporate creed. They wanted to build a God-like AI which could do anything from calculating a rocket’s trip to Mars, predicting the death of people and even perform surgeries. So basically GOD-like AI.
But as always, there was an unexpected fight between Elon Musk and Sam Altman, which was because Elon wanted OpenAI to be a for-profit venture but Sam disagreed that he wanted OpenAI’s developments to benefit all humanity. Elon disagreed and left OpenAI and later started competing with OpenAI with Tesla’s very own Self driving cars
In 2017, A group of 8 Google Researchers published a research paper named “Attention Is All You Need” for most people this was just another boring research paper published by google, what most people did not know was that this research paper contained a groundbreaking blueprint for something called Transformer Architecture.
Transformer Architecture was a technological advancement which in simple words improved the LLM’s memory significantly, basically before Transformer Architecture the AI would read sentences word by word, For Example, “The cat sat on the mat and it ran away” The Language learning module would read the sentence word by word, as in it would first read the word cat, then it would read sat separately, by the time it reached the word “mat”, it would forget everything it read before that. So, you get the basic understanding of it.
After Transformer Architecture, The Language learning module would read the whole sentence, so if you gave it the same sentence. It would read the whole thing at once, basically instead of reading one word at a time, it would read the sentence as a whole “The cat sat on the mat and it ran away” This type of architecture at the time was so revolutionary. Google could have built an extremely efficient LLM with this knowledge, but you know what they did, they did NOTHING!!
Google did not make an efficient LLM out of their knowledge of transformer architecture because they were scared that the LLM would say something racist, or it might help somebody make a bomb or even spread misinformation that causes a riot. For you and me, these things are very rare, but when you are a 500 Billion Dollar company, these things are very possible!!
While Google was paralysed by fear, someone was watching closely. He did not have 75 Billion Dollars or even 1000’s of researchers. This was Sam Altman, he and his team were different from Google because they had nothing to lose, they did not have a 500 Billion Dollar company and were relatively unknown.
So, Sam and his team went through every Research Paper published by Google and slowly took small steps to make the perfect LLM but everything wasn’t going well for Sam and his team because at the time, OpenAI was racing towards bankruptcy, and he did not have the funds. Do you remember the 1 billion-dollar pledge? Well, out of that, only 164 million dollars had come.
But even while taking steps to build the LLM, OpenAI was still building bad and troublesome LLMs, so after thinking for a long time, Sam made OpenAI a capped-for-profit company. Which basically means that an investor invests some money. The Investor can only get returns up to 100x of his investment and Returns earned after that would go to the Non-Profit mission, and the core shareholders of the non-profit mission will be the leaderboard in this team.
Even after making OpenAI a capped for-profit company, They were still going towards bankruptcy, in comparison, Google DeepMind itself had a budget of 442 Million Dollars. But suddenly an unexpected saviour came to their help. Microsoft!! Microsoft invested 1 Billion Dollars in OpenAI in2022 and in 2023 invested 10 billion dollars in exchange for all of Open AI’s technology and research. So, with enough funds and all the research done by Sam and his team. Open AI launched the killer product GPT -3 , and the rest is continued history. But now that OpenAI is bleeding in losses and Gemini, Claude and Deepseek are catching up and Gemini again has a Billion dollars to fight OpenAI, how will OpenAI fight these competitors, and who will be the leader of the AI Revolution? This question only time can answer.
In 2016, the world witnessed something nobody thought possible. See, for decades everyone thought that chess was the ultimate test of the intellect, but in 2016, everyone realised that they were wrong. In 2016, Google put its Google DeepMind to the test against one of the greatest intellects in the world. Lee Sedol. At the time a champion of a seemingly unknown game called Go.
While in chess, after the first 2 moves, there are 400 possible positions, but in Go, after the first 2 moves, there are over 129,000 positions. The number of games of Go is so many that it is more than all the atoms in the known universe combined, and until 2016, everyone thought that a machine would never be able to conquer this feat.
Google DeepMind, Alpha Go went against Lee Sedol and came out the victor. This was the start of the AI revolution in the World. Elon Musk and Peter Thiel decided that if they did not do anything concrete urgently, Google would take control of the AI industry completely.
Elon Musk and Peter Thiel decided on a Dream team. One which was set to destroy Google’s dominance over the AI market. The Team consisted of Sam Altman, Elon Musk, Greg Brockman, Ilya Sutskever and Peter Thiel. They pledged 1 billion for a non-profit venture, which had no shareholders, no profit motive and no corporate creed. They wanted to build a God-like AI which could do anything from calculating a rocket’s trip to Mars, predicting the death of people and even perform surgeries. So basically GOD-like AI.
But as always, there was an unexpected fight between Elon Musk and Sam Altman, which was because Elon wanted OpenAI to be a for-profit venture but Sam disagreed that he wanted OpenAI’s developments to benefit all humanity. Elon disagreed and left OpenAI and later started competing with OpenAI with Tesla’s very own Self driving cars
In 2017, A group of 8 Google Researchers published a research paper named “Attention Is All You Need” for most people this was just another boring research paper published by google, what most people did not know was that this research paper contained a groundbreaking blueprint for something called Transformer Architecture.
Transformer Architecture was a technological advancement which in simple words improved the LLM’s memory significantly, basically before Transformer Architecture the AI would read sentences word by word, For Example, “The cat sat on the mat and it ran away” The Language learning module would read the sentence word by word, as in it would first read the word cat, then it would read sat separately, by the time it reached the word “mat”, it would forget everything it read before that. So, you get the basic understanding of it.
After Transformer Architecture, The Language learning module would read the whole sentence, so if you gave it the same sentence. It would read the whole thing at once, basically instead of reading one word at a time, it would read the sentence as a whole “The cat sat on the mat and it ran away” This type of architecture at the time was so revolutionary. Google could have built an extremely efficient LLM with this knowledge, but you know what they did, they did NOTHING!!
Google did not make an efficient LLM out of their knowledge of transformer architecture because they were scared that the LLM would say something racist, or it might help somebody make a bomb or even spread misinformation that causes a riot. For you and me, these things are very rare, but when you are a 500 Billion Dollar company, these things are very possible!!
While Google was paralysed by fear, someone was watching closely. He did not have 75 Billion Dollars or even 1000’s of researchers. This was Sam Altman, he and his team were different from Google because they had nothing to lose, they did not have a 500 Billion Dollar company and were relatively unknown.
So, Sam and his team went through every Research Paper published by Google and slowly took small steps to make the perfect LLM but everything wasn’t going well for Sam and his team because at the time, OpenAI was racing towards bankruptcy, and he did not have the funds. Do you remember the 1 billion-dollar pledge? Well, out of that, only 164 million dollars had come.
But even while taking steps to build the LLM, OpenAI was still building bad and troublesome LLMs, so after thinking for a long time, Sam made OpenAI a capped-for-profit company. Which basically means that an investor invests some money. The Investor can only get returns up to 100x of his investment and Returns earned after that would go to the Non-Profit mission, and the core shareholders of the non-profit mission will be the leaderboard in this team.
Even after making OpenAI a capped for-profit company, They were still going towards bankruptcy, in comparison, Google DeepMind itself had a budget of 442 Million Dollars. But suddenly an unexpected saviour came to their help. Microsoft!! Microsoft invested 1 Billion Dollars in OpenAI in2022 and in 2023 invested 10 billion dollars in exchange for all of Open AI’s technology and research. So, with enough funds and all the research done by Sam and his team. Open AI launched the killer product GPT -3 , and the rest is continued history. But now that OpenAI is bleeding in losses and Gemini, Claude and Deepseek are catching up and Gemini again has a Billion dollars to fight OpenAI, how will OpenAI fight these competitors, and who will be the leader of the AI Revolution? This question only time can answer.
