{{active_subscriber_count}} active subscribers.

S2N Spotlight.

In today’s letter I am going to share a bit of detail on 2 exciting projects I am working on.

The first project you probably would have guessed has an AI component. We are living in extremely exciting times, with the pace of AI development seriously providing an edge to small entrepreneurs that I don’t believe has ever existed to the degree it does today.

First, some of the things I cannot stand. I hate the newsletters I receive that are so clearly written by AI. They lack personality, they lack experience, they lack subtle nuance, they lack humour (at least my type), and they most certainly lack humanity.

The second theme of things I cannot stand is the way many in the market believe that AI is going to come up with trading systems for them. There are many in this space who I respect, but I believe they are mistaken to believe that some AI-driven agent is going to run through a bunch of ideas and design a bunch of great systems and put them into the market, trade them through another agent, and win. I was quite disappointed to see one such respectable quant say my AI system built me 786 profitable trading systems. I went through all 786 trading systems; in short, I am not impressed.

So big jaw Berman, tell us, what has got you so pumped up?

Today, I took my wife (who cannot drive for 8 weeks due to shoulder surgery) for coffee to my local café and gave her a demonstration of Mini Me. I am going to keep this high level, as I am supposed to be in stealth mode.

I have built a research platform that has more than 600 daily chart and table outputs that cover the full gamut of global macro with a US focus. Every bit of the research platform is driven by Python code written by me, with some help from a billion “friends,” at a total cost of $20 per month. The code base interprets data and creates analytical outputs. I always knew there would be a natural way for me to leverage all the extra effort I have been putting into the platform. The way forward comes with a very natural way to take advantage of NLP (natural language processing), which LLMs (large language models) like ChatGPT—LOVE! In other words, I can feed AI all the steps in the process of running my code, and I can share with AI the binary outputs of these models.

Ok, big deal, spit it out already. Sorry, I was just vibing, to use the latest nerd word out of context. That little riff was just to show you this is not AI writing; this is me.

Writing basic and even not-so-basic commentary on certain charts and tables is boring, at least for me it is, especially when you are producing 600+ a day. So I decided I will build a mini-me, actually a version 2 of me that is far smarter than I am but is able to think like me and write like me and probably one day talk like me. By the way, that day is today, but I digress.

The main challenge, at least for me, is to do it in a way that addresses all the things I said I cannot stand and to do things that most of the AI commentary I have seen from competitors cannot do.

There are many different ways one can approach this; I have come up with my own system architecture. The bottom line is it comes down to teaching, training, mentoring, coaching, journaling, and whatever other adjective you can come up with to build a protégé. But to do it in a reinforcing way.

Today in the coffee shop I got to show my wife with the click of a button how Mike Version 2 is forming right in front of our eyes. Everything I have ever written, that makes sense at least, and every thought I have about life and the markets I have tried to capture, and I continue to capture, I will always need to capture new thoughts. I have built a framework that provides me with homework each day to mark, guide, or edit where necessary so that the smartest LLM models in the world can be harnessed at ridiculously cheap costs into a version of me.

I have overstayed my welcome for today, so tomorrow or another day I will share my equally exciting second project I am working on. However, I need to leave you with a sobering thought after writing what might be construed as hubris.

If I am building an improved version of myself, that means it will still come with many of my shortcomings, flaws, and blind spots; trust me, there are many. There are no perfect answers in economics. After all, Thomas Carlyle called economics “the dismal science” for a reason. Just because prices go up doesn’t always mean demand goes down. Just because a company reports a financial loss doesn’t mean its share price will go down. The economy and the markets are chaotic, complex systems. This is a game of figuring out nonlinear relationships and when and where they are likely to be found and expressed. This is not a game of exactness. There are no guarantees. Brains alone won’t win the day.

S2N Observations

One place’s deficit is another’s surplus.

You have heard me say it many times before, and I will say it again. Inflation is not under control. Headline inflation is sitting at 2.47% year-on-year, and the more important core is sitting at 2.91%. This is not due to tariffs, as many think; the inflation in June has come mainly from services. Goods inflation, which tariffs affect, has still not shown its face. But it will come, and with the shenanigans going on to boot Powell out early, the inflation genie ain’t getting back in the bottle anytime soon.