Chess, GFLOPS & Computation Power: My Passion for Artificial Intelligence, Pt. 1

The latest developments in artificial intelligence are nothing short of spectacular. If you have been living under a rock, prepare yourself, as your mind is about to get blown wide open, or at least tickled by the latest technological revelations in the field of machine learning.

To start off, I want to touch upon the subject of AI in the concept of human-like intelligence – what we currently know about it, and how we understand it. It appears that for a long time the subject of AI was immediately associated with a machine, specifically, a computer. It was often depicted as a super-intelligent, emotionless being.

Perhaps you’ve seen AI through a lens that was shaped by science fiction. Perhaps you are familiar with books such as 2001: A Space Odyssey or Neuromancer, or movies such as Ex Machina, Transcendence or The Matrix. Many of these showcased the extremes of what AI is, what it could mean to us as a race, and how easily things could go wrong if we are not careful.


This is not AI.


But rather than mimicking Neo or Arnie, I will keep things simple and quote a short definition of ‘intelligence’, which will set your thought process on the right path:

“The ability to learn or understand or to deal with new or trying situations.”

I hope you like games. There are two games that have entertained us for centuries; Chess and Go. There are a number of reasons why I mention these games. For one, it becomes easy to explain, understand, study and experiment with the concept of AI in a domain that has set limitations and rules that govern it (the game of chess being a 8x8 board containing 64 squares). These limitations were an excellent case and scenario for the brilliant minds at Google’s Deepmind team to explore and train their neural network, and to showcase what AI is capable of through their AlphaZero project.

If you’re not aware, chess is going through a sort of renaissance. One reason for this is recent advancements in AI. Back in ‘96-’97, Garry Kasparov played a number of matches against the most powerful chess engine at the time, DeepBlue – developed by IBM. Back then, the engine was in most basic terms a smart calculator with some primitive AI capabilities. It had an extensive database of positions, datasets and configurations such as game openings and significant strongholds. 


This is AI.
This is AI.


What is absolutely mind-blowing about IBM’s 1997 DeepBlue is that it boasted 11.38 GFLOPs (FLOP is ‘floating point operations per second – a measure of computer performance). In perspective, the PlayStation 5 has 10,300 GFLOPs. Beside the obvious increase in computational power, the key takeaway here is that back in ‘97, AI was not very creative nor could it conceive strategies of sacrificing early game in favour of late stage game advantage. Most of all, the hardware to perform AI calculations were vastly weaker. Nevertheless, DeepBlue eventually won in six games 3½–2½.

A great example of a search algorithm for a game of chess is the ‘Monte Carlo tree search’.



Shown in this example from, the tree search would be carried out many times for each position in the game, and to a specific depth. This takes care of many different permutations of each position of the pieces on the board as well as the potential win/lose ratio for a given position. Websites such as utilises AI and actually visualises the outcome of neural network calculations in real-time after each move, and it does it well.



For the current board position, the neural network is proposing X potential moves that according to the Monte Carlo tree search would yield/contribute towards a win for my white pieces. According to’s AI, the strongest move I can play (evaluated at +5.58 weight) is to move white rook from b1 to b,7 capturing the black pawn and putting pressure on row 7.

The interface provides a handy overall evaluation bar – the vertical bar on the left-hand side. My example game shows a 6.4 weight, which indicates my dominance against the opponent.

Thanks to advancements in AI, there is a great potential for making key advancements in fighting diseases, including viruses such as COVID-19 as well as helping us to produce drugs that can fight it. Perhaps AI will help us crack the mystery of viruses in general, and help us create a safer future.

Some of us here at Kyan contribute to the fighting effort against COVID-19 and other diseases by donating our hardware’s computational power through the Folding@HOME project:

If you want to contribute to the effort, head over to for more information. This amazing project has been active for many years, but during the COVID-19 pandemic, the community of “folders” grew so large that its computational power is now considered the most powerful supercomputer on the planet.

“Together, we have created the most powerful supercomputer on the planet and are using it to help understand SARS-CoV-2/COVID-19 and to develop new therapies. We need your help pushing toward a potent, patent-free drug. Use your PC to help fight COVID-19.”

There are many areas in which we already use AI in our daily lives. Many apps on our smart devices already utilise AI. For example, auto-correct or photo editing. A result of a recent partnership between Microsoft and Nvidia was Megatron-Turing NLG – the largest language model to date, with 530 billion parameters. MT-NLG is one of the most important models because it deals with language and involves tasks such as: completion prediction, reading comprehension, commonsense reasoning, natural language inferences, and word sense disambiguation. 



I can’t go any further on the subject of AI in NLG without mentioning Kyan’s Solution Architect, Andy West. Andy recently gave a brilliant Dev Talk on the subject of OpenAI and GPT-3, which is a younger brother of the MT-NLG model. The key takeaways from Andy’s talk for me were the amazing RPG Story Builder and my absolute favourite, the Magic the Gathering Card Maker.

Andy demonstrates the AI language skills by showcasing AI’s ability to generate RPG stories or MTG cards with just a few word parameters. You can watch Andy’s DevTalk below, or by clicking here.



This concludes, this brief introduction into AI and its applications. I hope you enjoyed reading it as much as I have enjoyed writing it. I do want to dive a little deeper in the technical side of much of what I’ve written about here, so please look out for a ‘part two’ in early May.

Previously from Damian:

README: Damian talks PHP, AngularJS and his love/hate relationship with WordPress
Dev Talks: End-to-end testing with Cypress, with Damian Boni


We are Kyan, a technology agency powered by people.