My background isn’t in technology or computer science, and I see that INFER is focusing on AI and Microelectronics. I’m intrigued by the questions, but should I actually be contributing my forecast if I’m not an expert — and if so, how should I tackle my first forecast?
The short answer: Yes.
The long answer: Yes, please.
The longer answer: You are the perfect kind of person to forecast on INFER. Even if you don’t have subject matter expertise or background in AI or Microelectronics, you still can and should make predictions. To explain why, I want to borrow from the philosopher Isaiah Berlin, who once made a distinction between “foxes'' and “hedgehogs” in his aptly titled essay, “The Hedgehog and the Fox.”
For Berlin, there are two ways that people view the world. Berlin defined "hedgehogs" as those who view the world through the lens of a single defining idea. His foil for hedgehogs was "foxes," who he described as people who draw on a wide variety of experiences and apply an interdisciplinary approach to solving problems. As the Greek poet Archilochus said, “the fox knows many things, but the hedgehog knows one big thing.”
Right now, you’re a fox – you’re not bound to any one guiding principle in the sphere of AI because you aren’t an expert. This is a very good thing. While there is certainly value in hearing from experts, often they have built their careers on one organizing principle. This can lead to bias and error, but that is perfectly okay – the very power of crowdsourced forecasting is that collectively our errors cancel each other out. Your forecast will have error too, but because they arise from different processes than everyone else’s, the group’s forecast is strengthened. By forecasting on a topic that is new to you, you can learn from the crowd (some of which might include experts) and they can learn from you.
Let’s walk through what happens when you want to forecast a question you find interesting but don’t have a lot of background knowledge about it...
Start by reading the background info for the question that you want to answer – INFER will always provide useful context, source data, or article links for further reading. Next, you can find other articles on your own, ask more questions (“What are microelectronics? Are they just really tiny electronics?”), and explore different viewpoints before you submit your forecast.
Keep in mind that you can always update your forecast. Maybe your first forecast is based on a few minutes perusing through one or two brief articles, but then a week later, you find another article that slightly changes your mind.
At some point, you'll also want to read other people's forecasts and comments. You may find that you agree with some people's rationales and disagree with others, but all of them will help you form an opinion. Perhaps you’ll realize that you were wrong about one assumption or a base rate. You may even realize you were wrong entirely, and change your opinion again. This could happen five times over, but it's best to change your forecast in small increments over time vs. one big swoop.
My goal is that you see that this entire process is critical to developing a well-informed forecast. You’re refining your predictions based on new information, and you’re seeing your forecasts as hypotheses that need to be tested, as opposed to authoritative predictions.
I engourage you, dear reader, to give this a try with a few questions that I've found to have pretty fascinating discussions:
- What percentage of "very high impact" AI projects on Github will have Chinese contributors in 2022 according to the OECD AI Policy Observatory?
- How much revenue will Intel report for the second quarter of 2022?
- Will the U.S. Congress pass a tax credit for semiconductor manufacturing or design before 1 January 2023?
Best of luck forecasting and keep those questions coming!
Your forecasting ally,