Technology evolves on a very fast pace and it’s often difficult to predict the future of technology or some specific directions that technological development will head to. Still, last week there was an event that unarguably defines some strategic directions that technology development will surely emphasize. So last week, AlphaZero won a 100 games chess marathon against Stockfish. Not only did AlphaZero win, but it didn’t lose a game at all! It won 28 games and drew 72. The spectacular aspect from a tech perspective is that AlphaZero learned the game in only 4 hours.
For those who are not familiar with chess, Stockfish 8 is the strongest chess engine right now, winning the Computer Chess Chanpionship few weeks ago. All chess grand-masters (GMs) and amateurs are using it to prepare for their games and to strengthen their game. Stockfish 8 is what I would call a “brute force” engine that calculates thousands different variations more than 25 moves deep. How deep it calculates each possible variation depends a lot on the available computing power. After all these calculations it decides what the best move is.
AlphaZero, developed by Google as a re-purposed AlphaGo AI, takes a total different approach: machine learning. Developers provided the chess rules and AlphaZero started to learn by playing. After 2 hours, it already got a super-human chess understanding. In conclusion, an AI beat a “brute force” engine.
The implications for technology are huge. First, this proved once again that an AI can learn things that humanity needed thousand years or more to understand. In this circumstances I think it’s natural to ask what could an AI do in medicine, cancer research, environment research and so on. And in my opinion this underlines a clear switch from “brute force” computing to “clever” computing. And this means, in the end, that all of us involved in the IT sector will have to shift our thinking paradigms. We need to change the way we look at computers, the way we architect modern applications and, in the end, the way we do our job. Are we really ready to do that?
How useful was this post?
Click on a star to rate it!
Average rating / 5. Vote count:
Latest posts by Dan Patrascu-Baba (see all)
- Configuration and environments in ASP.NET Core - 25/11/2019
- GraphQL in the .NET ecosystem - 19/11/2019
- A common use case of delegating handlers in ASP.NET API - 12/11/2019