A.I. in Blockchain and Bitcoin Technology
Can Humans Continue To Compete?
News headlines are shattered every single year with top level computer programming beating the best and brightest online gamers in society today. All across the globe, games that were once thought practically unbeatable not that long ago are being won by Artificial Intelligence (A.I.).
Whether that be board games such as the game of Go that ended up with Lee Sedol defeated by AlphaGo or some of the world's best Dota 2 players defeated by OpenAI, the world has been abuzz by humanity's newest overlords that are threatening to define what it can even mean to be human in the not too distant future.
And while many games have been defeated by A.I., not every game has been beat by computer programs. Generally speaking, for computer programs to win in the current environment, they need to be able to do a few things:
They need to have a clear set of rules
They need to know what proper winning conditions are
They need to be able to distinguish good from bad
The search space for games needs to be somewhat limited or redefined
Should We Welcome Our New Robot Overlords?
With the above list of rules, it can be relatively easy to take a look at the game of chess. The amount of moves that are possible in a game of chess are in present day computation relatively impossible to compute since there are simply too many possible moves. The search space is simply too big. So the search space needs to be pruned algorithmically so that the computer can search moves more efficiently and decide which moves can even make sense to look at to save time and processing power. Secondly, it needs to know which moves lead to winning outcomes. With the pruning techniques, winning conditions and help from Google's AlphaGo Zero, the world's strongest chess engines emerge with an unbelievable understanding of the game that almost begins to transcend conceivable human understanding.
However, A.I. bots can be weaker in various scenarios in today's world although still have a reasonable grasp as to what is going on. With games of incomplete information, games where not all information is visible to all players (for example, Poker), A.I. presently may struggle a little. This problem, however, is quickly being solved by top level A.I. programs by operating under the assumption that incomplete information games, given enough information, converge on complete information scenarios based on a distribution of outcomes of possibilities that a smart opponent may even engage in.
A.I.s have been targeting games with incomplete information which include, but are not limited to, Dota 2, Starcraft 2 and Poker to name a few.
The problem then grows further when we apply what the A.I. consequences are relative to Bitcoin gaming and Blockchain technology. Incomplete information games are a smaller scope of what can end up being the “games” played in the reality of real life scenarios for human beings and the optimal decision making that exists therein.
When it comes to Bitcoin gaming in particular, in order for it to thrive in the real world it needs to be somewhat impervious to A.I. technology which is undoubtedly tough if you consider the bleak outlook of what has been discussed so far. However, there are solutions, even if they may be temporary in the year 2019.
There Are Some Short Term Strategies
The best way to allow Blockchain online games to thrive in today's economy, moving into 2020 and forward is to base the Bitcoin games on a few different theoretical concepts which include:
Make the win conditions not very obvious and hard for computer programs to define
Make the rules extremely noisy
Hide as much information from bots and humans as possible
Make it difficult for bots to strategically determine winning vs losing conditions
With all of those conditions in mind, you can then start to draft games that at least challenge A.I. technology and can allow humans to play against each other in proper Blockchain games / Ethereum games and not play solely against outwitting bots.
To wit, MintDice has a strong focus on the development of Bitcoin games that will be slightly difficult for A.I. to solve, so stay tuned and keep in touch with MintDice's Casino Bitcoin roadmap development.
At least three skill games will be developed by MintDice with other skill games planned for future development depending on the overall engagement by the Bitcoin community. If there is demand for proper Bitcoin games of skill, the MintDice Bitcoin Casino will be happy to continue to produce exceptionally high production value games for the community to enjoy and play against each other.
It is a war against the bots, and not one that we can win in the end, but one we can have fun with for at least the next few years, we hope :)