They automated signal discovery in what kind of data? I have heard their use of unconventional data sources is the source of their success, not automated signal discovery.
In a way, the way contracts and legal documents are written now are already an antagonistic optimization against humans natural language ability. A typical contract is pretty difficult for a human to parse without experience. Imagine how much more tricky and complex legalese can get if lawyers start battling machines!
How correlated is your models output with a simple measure of the volume of interactions a user has with an app?
Main question...is this a domain where ML really adds a lot of value beyond the basic concept of High Interaction (lots of time spent and clicks/pages visited) -> High conversion Probability?
I could imagine piping system stats into this and then setting up alerts when something "unusual" is happening.
The super amazing thing would be multivariate analysis of multiple series so you could see when the relationships between variables becomes anomalous as well.
One thing I can say is that it's possible to feel the way you feel even for someone in a situation where they seem like they "have things figured out".
Read Seneca's On the Shortness of Life. It and other works of Stoic philosophy help me at times when I feel this way.
Changing jobs/cities/partners can sometimes be the answer, but finding a source of value that is robust to changing circumstances is the only hope we have in life.
I like the simplicity of the UI, but I can't seem to make straight lines horizontal lines between components, and I wish there were more options for annotating things (text, short description of a relation, etc).
Also most importantly, where is the option to export to an image or some other format?