AI Money is Heat-Seeking

Investingai

Jul 2

One of the best investment mantras I've heard is:

Money is Heat-Seeking

Momentum 101

Money goes where the action is, and like a fire, that action becomes self-perpetuating. Thus, one of the most successful, and highest-return-for-lowest-effort methods for investing is momentum investing.

While mo-mo or momentum gets a bad name, as even some professional investors have been known to hold stocks where they have no idea what the company does... it is a great way to make money consistently.

How to do it?

Buy what goes up. Sell what doesn't.

The wrinkle is that things can go down just a bit, so you can sell winners or have too much turnover if you sell on every down day. Technical analysis gives us an idea of not selling when a stock moves meaningfully lower, as defined by moving averages of recent prices.

Momentum investing is the execution of the concept that money is heat-seeking.

Market Turns & Leadership Changes

Sometimes the broad market has positive momentum (think January to now), and sometimes even when the market is going down, a subsector is going up (think Energy up 60% from it's bottom in the fall of 2021 while tech was melting down in that same time frame).

That's a change in leadership.

Changes in leadership tell you about where the heat (money) is leaving, and where it is going to next. They happen slowly, and then all of a sudden. Leadership changes may signal more risk appetite or less.

Leadership changes also happen between risky assets.

Such has been the case between Web3 and AI. Yes, bitcoin has rebounded back to $30,000 along with NASDAQ and the S&P rallying near all-time highs. Bitcoin is a risk-on speculative asset. Makes sense.

But the Web3 market overall hasn't rebounded the way Bitcoin has, and funding from VCs has gone AI red hot, while VC funding for Web3 is still rather ... cool (and not in a good way).

AI: ChatGPT to NVDA, follow the money

My grandfather used to say: "Want to understand what's going on? Follow the money!"

Hindsight is 20/20

We all saw ChatGPT become the fastest-ever adopted technology in history. I wrote about it in this newsletter. BUT, this is where historical perspective can hurt as much as help.

AI, like Web3, has experienced many winters, and for longer.

Remember the AI winter of the 1970s? No?

Here's a quick recap of the history of AI:

WW2: MACHINATION & CODE BREAKING (advances in compute + algos)

1942: ISAAC ASIMOV WRITES THE THREE LAWS OF ROBOTICS: #1 Don’t injure Human. #2 Obey Human. Self Protect if not #1 or #2.

1950: ALAN TURING PROPOSES THE IMITATION GAME (WHY NOT AI?) Is a machine imitating the sentient behavior of a human, itself sentient?

1952: ARTHUR SAMUEL’S PROGRAM LEARNS AND PLAYS (CHECKERS)

1956: DARTMOUTH HOLDS AN AI CONFERENCE

1957: FRANK ROSENBLATT BUILDS THE PERCEPTRON (NEURAL NET)

1970s: DARPA (US GOVT) CUTS AI FUNDING… (AI WINTER)

1997: IBM’S DEEP BLUE BEATS KASPAROV (CHESS)

2011: A NEURAL NET SEES CATS (CV: COMPUTER VISION)

2014: TESLA ANNOUNCED AUTOPILOT (COMPUTER VISION DRIVING)

2012: GEOFFREY HINTON UNLEASHES DEEP NEURAL NETWORKS (VISION++)

2016: ALPHAGO DEFEATS HUMAN GO CHAMPION (GO)

2019: GPT-3 “READS” INTERNET (NLP: NATURAL LANGUAGE PROCESSING)

LLM: 175 billion parameters/variables, 10X larger than previous models

2020: GPT-3 ACCESS GRANTED. Copy.ai, multiple startups

2022: ChatGPT RELEASED. FASTEST ADOPTED TECH (GENERATIVE AI)

So, I ask you, Where's the Money?

Decades, 80+ years of progress, and before this past year, where were the AI unicorns?

Billions were made via social media companies like Facebook, who used AI, but pure AI companies were NOT good investments. I heard all kinds of reasons. The AI was not allowed to grow, to mature.

Case in point, Google bought DeepMind for appx $500M and that was one of the biggest AI company purchases.

Especially in private markets, you don't know how your investment does until you sell your shares (IPO, M&A, etc) so interim hype isn't much help. So the jury is still very much out on whether this spate of AI-frenzied investments will yield stellar returns.

However, what I missed, was the overflow of all that heat-seeking money.

The so-called "obvious" trade. Not just the infrastructure. But all those traders, professional and individual, who had no more momentum to ride in Web3 (except in Bitcoin and ETH), and were seeing massive interest, heat, and money going into AI.

First NVDA was crushed because Web3 "bitcoin mining" demand was declining. Now they are the darling of AI, running calculations each time you prompt ChatGPT. Who sells to NVDA? AMAT. And MSFT is a majority owner of OpenAI, ChatGPT's creator, and ChatGPT runs on MSFT cloud.

Pretty soon, I was getting emailed lists of AI-adjacent public companies.

It is quite far-fetched that AMAT is an AI company. In fact, they are more likely to get hurt by the US-China squabbles. But money is heat-seeking, momentum is self-fulfilling, and markets are voting machines in the short term, so those were great investments off the ChatGPT release last December.

Reminders to self: