Jeremiah Lowin is probably the smartest guy I know, and that is saying something. He is an expert in the fields of statistics, artificial intelligence, and risk management—among many other things. He is currently the Director of Risk Management for a private investment firm in the New York area, but has spent years working with machine learning and AI. This conversation is broken up into two parts. In the first part, we explore artificial intelligence, machine learning, and models. Then we shift to what risk means in a portfolio and how it can be managed or at least redistributed (which starts around 40 minutes into the conversation).
2:19 – (first question) – Looking at Jeremiah’s background and how he got started in a place where he could work in artificial intelligence
4:43 – The first exposure to necessary approximations and how that led to building models
6:10 – His time at Amaranth and how that led him to getting an advanced degree
8:35 – How does machine learning relates back to the simplest of models and where it goes from there.
10:50 – How does machine learning help break down data sets more quickly
13:35 – Explaining this concept with the example of google’s exploration of machine learning in translation.
16:30 – Google has had a huge leap in accuracy as they had time into their models and a recurring neural network
18:01 – The creation of the Lowin Data Company, what they do, and what has worked for them
21:01 – The Chinese Room Argument (John Searle)
22:19 – The Turing Test
22:55 – What will all this machine learning mean in terms of replacing jobs in society
26:27 – What are the jobs that will tough to replace with a machine
28:02 – Cultivating relationships and the movie Her
29:46 – Why AI machines aren’t intelligent
32:39 – Defining intelligence in machines as possible decision making or leaps of intuition
33:30 – Could discovery by a machine be the sign of intelligence in machines
35:55 – What can we learn from the findings of Lowin Data Company that can be applied to our own ability to learn and improve ourselves
38:35 – What led Jeremiah to start Lowin Data
Geoffrey Hinton – A Practical Guide to Training Restricted Boltzmann Machines
41:17 – Jeremiah’s working definition of risk
44:34 – The tenements of risk management
46:22 – A big part of risk management is shifting the distribution around
49:59 – How much of this is quantitative vs qualitative
53:14 – Why one of the most important skills in risk management is uncovering unknown unknowns and how Jeremiah applies this in his day-to-day job
56:58 – Getting information out of a closed system
59:59 – How often did Jeremiah’s involvement in a portfolio have a major impact on the decisions that were made
1:02:13 – What are the dimensions that Jeremiah plots out in his 3D graph that he uses to analyze a portfolio
1:05:15 – Where does Jeremiah fall when it comes to a simple index portfolio verse a more active portfolio
1:08:23 – What has driven the change that has led people away from active managed funds
1:11:01 – Has anything changed internally among hedge fund managers that could lead to a shift towards more passively managed funds
1:14:11 – Will artificial intelligence based tools become a bigger part of the investing process going forward?
1:17:52 – Looking at Jeremiah’s most memorable day
1:20:36 – Kindest thing anyone has ever done for Jeremiah
For more episodes go to InvestorFieldGuide.com/podcast.
Sign up for the book club, where you’ll get a full investor curriculum and then 3-4 suggestions every month at InvestorFieldGuide.com/bookclub
Follow Patrick on twitter at @patrick_oshag