Unconventional wisdom: Investing remains a craft - even in the age of AI
What a $19 million lunch with Warren Buffett reveals about judgement, process, and machines.
Conventional wisdom is a byproduct of groupthink that presents solutions good enough for the average person while simultaneously not being right for any individual. You follow it at your peril. Each Monday I will challenge the investing norms that just may be holding you back from living the life you want.
Unconventional wisdom: Investing remains a craft even in the age of AI
“Computers don’t have insights. People do. And collaborative efforts are only effective when individuals are valued.”
- Erik Larson
In 2022 someone paid $19 million US to have lunch with Warren Buffett at NY steak house Smith & Wollensky. Proceeds went to a San Franciso charity named the Glide Foundation which combats poverty, hunger, and homelessness.
I’ve been to Smith & Wollensky and it is safe to say the anonymous donor wasn’t paying for the meal alone. What the donor paid to pick the brain of the world’s greatest investor is quantifiable. The value of the lunch was up to the anonymous donor.
A valuable conversation with Buffett requires a base level of knowledge and some thought about how best to use this opportunity. Asking vague questions isn’t a good way to profit from such a knowledgeable and experienced investor.
If you aren’t asking the right questions you can’t get much from any conversation. Most investors don’t have access to the oracle of Omaha. Instead, many are turning to artificial intelligence to guide their financial moves.
What AI can do
In the late 18th century three inventions revolutionised the textile industry. The flying shuttle, spinning jenny and water frame made cotton cloth cheaper to produce. This helped people buying cloth but devastated the lives of textile workers.
The workers responded by forming the Luddite movement in the early 19th century. The Luddites conducted military style drills outside of industrial towns in England and invaded the mills to smash the machinery.
The machines came out on top against the Luddites. But the term stuck and is now applied to anyone who is resistant to new technology. Despite the views of many of my younger colleagues I am not a Luddite.
I use AI as a tool to help with investing. But I’m cognisant of what it can and can’t do. The most obvious investment application of AI is in quantitative investing. In a sense it is more evolutionary than revolutionary. Quantitative investing uses data points to identify shares to buy and sell. The approach has been around for almost 75 years.
Pulling insights from large datasets originally involved feeding punchcards into giant mainframes. Computing power increased and machine learning crunched the numbers more efficiently. Large language models are the next step in that evolution.
For individual investors AI can be an effective way to identify companies to research further – but only if you know what attributes you are looking for. Much like lunch with Buffett asking, “what share should I buy” is a fruitless way of getting what you want. Garbage in, garbage out is a maxim that very much applies to AI.
What AI can’t do for you
There will be another bear market and if history tells us anything investors will respond poorly and take actions that lower their long-term returns. AI will not be able to regulate your emotional response to this inevitable event. It can’t make you patient and thoughtful.
When your portfolio is down 30% AI won’t be able to tell you what to say when your partner asks if everything will be ok. AI can’t relieve the pressure of being responsible for your family’s financial future and won’t be there for you when doubt takes over your thoughts late at night.
AI can be a tool to find investments that meet the criteria needed to achieve your goals but it can’t define them. AI can’t put structure around your decision-making process. Automation and efficiency are helpful, but it isn’t a substitute for the process of investing which is far more important to your outcomes than the investments in your portfolio.
Taking AI for a spin
I asked Claude for the best US and Aussie shares. I received a list of 12 shares and 2 ETFs – Palantir, CoreWeave, Taiwan Semi, Western Digital, Cloudflare, Eli Lilly, Pro Medicus, Xero, WiseTech, Goodman Group, NextDC, Telix Pharmaceuticals, NDQ and FANG. Nothing on this list is in my portfolio.
I did dismiss several high-level follow-up questions that were attempting to add specificity to my question. The follow-up questions from AI are generally good at narrowing choices but require some knowledge to navigate and a clear view of what you are looking for.
I put my investment strategy and share selection criteria into Claude and asked which US and Aussie shares I should consider. I was given a list of 15 US listed shares and 9 ASX listed shares. I own 6 of them – Proctor & Gamble, PepsiCo, Johnson & Johnson, ADP, CSL and Transurban. There was nothing on the list that I wouldn’t consider for additional research.
My investment criteria isn’t as strictly defined as a quantitative model. A series of financial metrics alone doesn’t provide the context needed to make a judgement call on a particular share.
Quantitative strategies acknowledge and attempt to counteract this truth by buying vast amounts of shares. A metric like a high return on equity alone doesn’t tell you much about an individual company but if you buy every share with a high return on equity it may lead to better returns. Tools are only as effective as the job you are asking them to do.
Final thoughts
An American couple recently purchased intimate apparel producer La Perla for $29 million US – about a lunch and a half with Buffett. People pay upwards of $600 for La Perla bras and in no small part they are paying for the handicraft of talented Italian seamstresses who hand loom each article of clothing.
Those prized seamstresses have won the battle against the machines the Luddites zealously destroyed. There is a place in this world for efficiency and a place for craftsmanship stemming from experience and knowledge.
Investing remains a craft – even in the age of AI. The efficiency gains from AI matter but success comes down to judgement and regulating the emotional pull of markets. Your outcomes are a result of knowing what matters to you and not the collective ‘wisdom’ that AI puts at your fingertips.
If you aren’t using AI to help with investing give it a try. Just make sure you go in knowing what it can and can’t do.
Email me at mark.lamonica1@morningstar.com and let me know how you use AI to help with investing.
Get your finances on track with Invest Your Way
Our book Invest Your Way is available in 206 bookstores across Australia. Kindle and audiobook versions can also be purchased.
Invest Your Way is a personal finance book that combines foundational investing theory, real-world application and our own experiences. It is designed to help readers create a financial plan and investing strategy that is tailored to their unique goals and circumstances.
Get Mark’s insights in your inbox
Read more of Mark’s articles
Read previous editions of Unconventional wisdom
What I’ve been eating
After World War II Japanese soldiers returning from Manchuria tried to re-create a Chinese dumpling called jiaozi. Gyoza was born. Why it took the Japanese until after World War II to add a dumpling to their cuisine is a mystery. Almost every culture has a dumpling like dish, and the Chinese have been eating them since 150 AD. Japan may be insular but this is a step too far.
Gyoza Hohei is a 20 seat restaurant in Gion, Kyoto. There are two choices – garlic & leek and ginger. I preferred the garlic & leek but try some of each and get multiple orders. They are cooked in the Japanese style meaning they are steamed and fried in the same pan. Gyoza Hohei is popular but show up early and get a number out of a vending machine and you will be in the first seating when it opens. Throw in several beers and you are on your way to great night in Kyoto.
