YWR Intelligence

‘I don’t get all the hype about AI. I don’t use it for anything’

Portfolio Manager

The near future:

YWR… I’m on the bridge on the way in. Should be at my desk in 15. How’re we doing in Asia?

You’re losing $237,000. Your worst performer is Baidu. Down 1.3%

Baidu!!! FFS! Any reason it’s down again?

No news specifically, but I did find a Twitter post about Chinese web search trends which helps explain the lack of growth in Baidu’s search business. The link and summary are in your research folder.

And Europe?

Down $47,000. Vinci is down 2.2%.

Thanks. This is so frustrating. I keep losing money everyday. I’m having a terrible month. I can’t sleep. And Vinci. What do I do? Do I sell it, buy more or do nothing?

I entered the order to buy 10,000 shares as you requested yesterday. I recommend that after this trade you do nothing. Our research view and price target are unchanged. Definitely, don’t sell. 65% of the time you sell on the lows.

You’re probably right. Thanks for the advice. Maybe we need some new ideas. YWR, I want you to go through all the companies reporting results today. Find any stock +6% or more after earnings. Then listen to the management call and send me a summary of what’s new and what caused the move. For each company also run an EPS forecast for the next 3 years and compare it with consensus earnings. Use a bullish forecast assuming margins and revenues continue at the latest growth rate. Just run the forecasts and see what the numbers look like.

Certainly, but may I ask why? You don’t usually trade in this manner.

I’ll explain the behavioural theory behind it later. Another thing. Go through

Kuppy

‘s KEDM Monitor and find me two good ideas. It’s full of tons of information, I just never find time to sit down and take it all in. But I should.

I already did that. Here are the two KEDM ideas I found interesting.

1. Chewy Inc. (CHWY) - Buyback

Summary: Chewy Inc. (CHWY) has announced a buyback program and has high short interest. This combination can lead to a short squeeze, driving up the stock price.

2. Cheniere Energy, Inc. (LNG) - Buyback

Summary: Cheniere Energy has topped off its buyback program by another $4 billion through 2027, representing approximately 11% of its market cap. Since 2022, they have repurchased about 10% of their shares outstanding.

Those sound good. Should we buy both?

Erik…. It’s always good to look for new ideas, but I notice when you are stressed you tend to overtrade. Just be patient. I know you are worried about performance, but our research and portfolio construction is solid. The ideas will work.

Thank you YWR. I appreciate your help. We make a good team.

---

This trading conversation is only a few years away. Maybe 2027.

Machines are acquiring new skills at an accelerating rate.

Citigroup: What Machines Can’t Master

We are also nearing the intelligence explosion. The inflection point where machines teach themselves. Full Artificial General Intelligence (AGI) probably occurs in 2027. Or, we might already be there. Some Microsoft researchers think ChatGPT 4.0 is showing early signs of AGI (Sparks of AGI: Early experiments with ChatGPT 4).

AI in Asset Management

Portfolio Management is one of the most dynamic, complex and challenging jobs possible. It requires high levels of information processing, pattern recognition in chaotic systems, decision making under uncertainty, and behavioural psychology. Which is why AI will be so transformational.

In the conversation above I was trying to demonstrate 4 areas where AI will be useful.

1. Idea generation. Increasingly, AI will connect the dots between financial results, corporate actions and investor positioning. For example, ChatGPT notices that if Chewy does a buyback it creates a short squeeze. It’s learning different ways to win.

2. Earnings Forecasts at Scale: I love this one. Earnings modelling will be a discussion with your AI assistant. Load the previous financial results, then guide the AI as it makes forecasts. It cuts out so much of the tedious data entry from the earnings reports. Instead, you will be sitting back with a coffee saying ‘Good first cut. Now grow costs less, and have operating margins expand 0.5% every year. Also have the company buy back 2% of shares every year.’ It can already do this. But imagine running this across an entire index and creating your own bottom up 2027 estimate for the S&P 500.

3. Portfolio Management: The AI assistant will combine and size your ideas for optimal return and volatility. ‘Erik. You have enough ideas correlated to inflation. I suggest increasing the Tenet Healthcare position.’

4. Trading psychologist: You are going to talk to your AI trading assistant, and ask what it thinks, especially when you are stressed. -

AI Weak Points

There are three weak spots I see in the application of AI in asset management. They will probably improve over time.

1. Sense of time. AI’s don’t have a good sense of the passage of time and yet time is an important dimension to investing. It takes time for ideas to work and knowing when to be patient and when to change is a key decision skill. It turns out Temporal Awareness is an important area of research to achieve full AGI.

2. Cause and Effect: Machine Learning is based on statistics and pattern recognition. It’s surprising what AI has been able to achieve with these techniques, without really knowing what is going on. For example, the machine knows with 97% certainty A & B occur together. But it never knows ‘why’ something is happening. It doesn’t know whether A is causing B, or vice versa.

   The machine never asks why the apple falls from the tree. It doesn’t care. Apples fall from trees with 100% certainty. That’s all you need to know. It’s similar to how a child speaks English from pattern recognition without knowing any grammar.

3. Lack of certainty in long-term investing: In machine learning there is usually a right and wrong answer. The picture is of an apple or it isn’t. The credit card transaction is fraudulent or it isn’t. The machine makes a prediction, sees how well it did, and then adjusts. But is JP Morgan a good investment? Maybe. Who knows. We won’t really know until 5 years down the road. There is no right or wrong answer.

   We want the AGI and its massive brain to make great long-term stock picks, which make us rich, but this skill is going to take longer to learn, and requires improvements in Temporal Awareness. CIO’s and PM’s still have a job. Relief.

YWR Intelligence

Yes, AI it isn’t perfect yet. It make mistakes, and there are things it doesn’t understand, but it’s improving rapidly. The last 2 years have been shocking.

What’s obvious is the most important thing is to get in motion. Quickly.

So I made us our own investment assistant using ChatGPT4.

YWR Intelligence.

YWR Intelligence has been trained in how we like to invest and what we are looking for. I’ve also loaded its knowledge base with YWR posts. It will use the power of its enormous brain to make money.

Give it a try.

It’s GPT4 so you can upload files for it to analyse, or work on earnings models together (in csv format).

Oh… almost forgot. Nothing it says is financial advice.