Chess
|
|
A complicated, but not complex, system. Therefore not a great grounds for concluding that software is going to eat the world and turn we mortal meatsacks into battery-pods for Skynet, but that won’t stop Daniel Susskind from leaping to just that conclusion anyway.
Will chess-playing AI chatbots take your job?
Ask yourself “is my job like chess?” If it is, you should get your coat. The sooner you start looking for a job that isn’t like chess, the better.
| Feature | Chess | Not Chess |
|---|---|---|
| Complexity | Complicated | Complex |
| Rules | Simple. Static. Common. | Unclear, changing, often differing between players, incomplete, liable to change without warning. |
| Logic | Fully logical. | Basically irrational, except by accident. |
| Outcome | zero-sum | non-zero sum |
| Data | All relevant data available to both players at all times. | Incomplete, mainly absent. What data there is will be ambiguous, unevenly distributed, and may only emerge once it is too late. |
| Language | Mathematical. No scope for ambiguity | Ambiguous, metaphorical; requiring interpretation and psychology. |
| Boundaries | Entirely bounded: two players, 8x8 board, 16 pieces each. | None. Whatever you bring to the party. |
Now, it is true that algorithms, and combinations of algorithms, can help crunch data and provide you with information that you might not otherwise have that give you more tools for making those value judgments that wicked games impose on you. But this is no different to how technology has helped us since plough. Ploughs don’t plough fields by themselves, after all — and they can’t help us with when, or which field, to plow.