General Discussion

To me, the interesting thing about things that failed - were they interesting questions and did they fail fast bc of the technology? Failing fast is better than wasting a lot more time obviously. If you are asking questions that were never feasible to consider before, that is also good.

In my line of work, I often create things that enable one to ask questions that could never be answered with the old technology. We often have to “backtest” certain strategies for what we do to get our research done and see if we go to production. In one case, the old technology stack could take 15 hours in the cloud using 120 cores. I found a way to knock it down to one local core that gets done in 2 minutes - using 30x the number of timestamps. This opened up a whole new world of questions we could ask. It also helped us to fail fast on things that don’t work so we can move on quickly to something that will work. A real research enabler.