While we are still debating and guessing mostly what A.I. and smart machines will bring in the future, the future is already slowly sneaking its way into the present.
Earlier this year in July, Deepmind, Google’s A.I. division was showcasing that their machine taught itself to walk. It is really fun and interesting to watch how a machine learns to walk:
And then this month, the same division made public Alpha Go zero, a machine that taught itself to learn GO and then proceeded to beat the previous Alpha Go; oh, and it kicked its ass!
Now both these advancements could be considered and compared to games. From Wikipedia:
A game is a structured form of play, usually undertaken for enjoyment and sometimes used as an educational tool.
Also worthy of note is that humans learn a lot from play too. A game generally has components/elements/pieces and rules/constraints/conditions. Last but not least, it also has a clearly defined objective.
But we can mostly say the same thing about problems humans are trying to solve, scenario simulations with the goal to find a solution and other experiments.
What humanity learned about go over 3000 thousands of years, Alpha Go Zero learned in 40 days.
Let’s put this into perspective: that is 2.7 million percent faster!
In the future we will simulate and learn and advance with help of machines a lot faster. Our skills and knowledge might actually focus more on how to best use their potential and capabilities and less on actually mastering all of the skills ourselves.
So here are some other “games” that I see computers playing together with us with the goal to win:
- Keeping humans alive on a Mars/Titan or any other future human colony
- Learning a new alien language from scratch.
- Study how to use a new discovered resource by running all tried experiments but also new ones
- Intergalactic battle? Let me just log in to my A.I. support simulator.
Any other interesting “games” I should add to the list?