Episode #276 from 14:01
Inflation
You said table. Some of that also is the task of visualization, how to extract from this complex set of numbers, patterns that somehow indicate something fundamental about what's happening. So summarization of data is still important. Perhaps summarization not down to a single scale of value, but looking at that whole sea of numbers, you have to find patterns like what is inflation in a particular sector? What does it maybe change over time, maybe different geographical regions, things of that nature. I think that's, I don't know even what that task is. That's what you could look at machine learning, you can look at AI with that perspective, which is how do you represent what's happening efficiently, as efficiently as possible? That's never going to be a single number, but it might be a compressed model that captures something beautiful, something fundamental about what's happening. It's an opportunity for sure. If we take, for example, during the pandemic, the response of the political apparatus was to lower interest rates to zero, and to start buying assets, in essence printing money. And the defense was, there's no inflation. But of course you had one part of the economy where it was locked down, so it was illegal to buy anything. It was either illegal or it was impractical, so it would be impossible for demand to manifest. So of course, there is no inflation. On the other hand, there was instantaneous immediate inflation in another part of the economy, for example, you lowered the interest rates to zero. At one point, we saw the swap rate on a 30-year note go to 72 basis points. Okay. That means that the value of a long-dated bond immediately inflates.
Why this moment matters
You said table. Some of that also is the task of visualization, how to extract from this complex set of numbers, patterns that somehow indicate something fundamental about what's happening. So summarization of data is still important. Perhaps summarization not down to a single scale of value, but looking at that whole sea of numbers, you have to find patterns like what is inflation in a particular sector? What does it maybe change over time, maybe different geographical regions, things of that nature. I think that's, I don't know even what that task is. That's what you could look at machine learning, you can look at AI with that perspective, which is how do you represent what's happening efficiently, as efficiently as possible? That's never going to be a single number, but it might be a compressed model that captures something beautiful, something fundamental about what's happening. It's an opportunity for sure. If we take, for example, during the pandemic, the response of the political apparatus was to lower interest rates to zero, and to start buying assets, in essence printing money. And the defense was, there's no inflation. But of course you had one part of the economy where it was locked down, so it was illegal to buy anything. It was either illegal or it was impractical, so it would be impossible for demand to manifest. So of course, there is no inflation. On the other hand, there was instantaneous immediate inflation in another part of the economy, for example, you lowered the interest rates to zero. At one point, we saw the swap rate on a 30-year note go to 72 basis points. Okay. That means that the value of a long-dated bond immediately inflates.