Visualizing "The Future According to You"
We want to provide a canvas on which you can paint your predictions of the future. Not just a passive canvas, but an active canvas: one that can hold not just your predictions of particular times, but also your predictions about trends and changes: a system where you can input your beliefs and watch the future move.
Our questions are designed to take your beliefs as input and see what those beliefs imply. We tried to make as few assumptions as possible, and we tried to make our assumptions as non-controversial as possible; we want the interesting parts of the future to come from you. (We did have to introduce some assumptions to simplify our model — we wanted the questionnaire to predict your next sixty years, not to take your next sixty years to fill out. If you'd like to learn more, check the frequently asked questions or dive into the gory mathematical details).
Besides making our canvas active and giving it room for your beliefs, we aim to provide a canvas that can hold your full uncertainty. Most futurism is about telling a single, specific story about what might happen. Specific stories are great in cases where you know exactly what will happen. However, if you're dealing with a wide swath of future possibilities and you're a human being, chances are you don't know what will happen. You have some guesses, you know some things are more likely than others, but you can't say "In September of 2038, John Smith will invent the Such-And-Such Widget, which will allow him to build human-level AI".
In the Uncertain Future, we let you input the full range of your uncertainties — and we show you the distribution over futures that your current, uncertain beliefs imply. We help you create your own picture of the uncertain future.
How to use our system to draw your picture of the future:
We'll ask you a sequence of questions aimed at getting your answer to the question: "When will human-level AI be invented?" We'll model your beliefs in four parts:
- Is human-level AI possible in principle?
- Assuming business as usual (no nuclear or other catastrophes, and no brain-boosting methods that seriously speed up research), how long until human-level AI?
- What if a nuclear or other catastrophe interrupts AI research?
- What if future brain-boosting methods speed AI research?
The Uncertain Future requires Java 6 or later. Click here to get a Java runtime environment.
We recommend using the browser Firefox or Opera on Linux/Windows and Safari on OS X (Mac). Not all of the features are available in Internet Explorer for Windows or on Firefox for OS X.