Krittamet Saehan/iStock via Getty Images
When’s the last time you changed your mind about a big decision?
Late this summer, we learned that my daughter’s JV field hockey coach had quit, just two weeks before tryouts. The team needed a coach or twenty girls would lose the chance to play. A good friend reached out to see if I might be interested. My gut reaction was no. I knew very little about the sport, and I could not imagine finding the time given a busy work and family calendar.
Fortunately, I reconsidered. I knew I’d have fun coaching with my friend, who was willing to shoulder the additional responsibility of being head coach. I also liked the idea of giving a group of girls the same opportunity my daughter had been given the year before. Most importantly, it meant more time with Caroline. I’m so thankful I ultimately changed my mind.
It has turned out to be a priceless experience. I have learned a new sport, gotten to know a terrific group of girls, and enjoyed helping a team get the most out of its individual talents. Watching these young ladies have fun, improve, and in some cases discover a new favorite activity has been incredibly rewarding. None of it was on my bingo card for 2026. Yet now, I have a hard time imagining my life without it.
That experience has me thinking about a broader question: when is changing your mind the right move?
There is enormous value in consistency. In investing, constantly changing your mind in response to the latest price movement, headline or popular narrative is usually a terrible strategy. Some of the best investments require doing exactly the opposite: remaining patient when others will not.
But there is another danger too. A process that never changes in response to new information is not discipline. Eventually, it becomes rigidity.
I have been thinking about this quite a bit as my views on artificial intelligence have evolved.
When I first tried ChatGPT, I found it interesting but not especially useful. It would confidently suggest restaurants that had been closed for years or provide answers that sounded plausible but were plainly wrong. I did not view it as a game changer.
That began to change as the technology improved. I started using AI more frequently to edit things I had written, and by 2025 I was using it much more heavily for investment research.
Earlier this year, I experimented with building simple tools in Claude Cowork. One project was a family calendar system that scanned our work and family calendars as well as emails to prepare us for upcoming activities. It was not as good as a human administrative assistant, but it helped and I could see the potential. However, the process was frustrating. I would spend meaningful time getting something to work, only to have the model glitch, lose track of what it was doing, or repeatedly ask for permissions I had already granted.
What is striking is how quickly that experience has changed.
More recently, I have been using personal AI agents such as GrokBot and Muse more frequently in both work and everyday life. Tasks similar to the family calendar project can now be handled far more simply and effectively. The systems require less prompting, understand intent better, and increasingly complete multi-step tasks rather than simply answer questions. Often, the AI itself suggests the next useful application.
I am certainly not a Silicon Valley power user or software developer. But compared with many friends across different professions, I use these tools extensively. What has caught my attention most is not simply where the technology is today, but how quickly it keeps improving while adoption still feels relatively early.
That raises an obvious question. What happens if both the rate of improvement and the rate of adoption continue?
I do not know the answer. But my own experience has caused me to assign a greater probability to AI becoming one of the most consequential technological changes of my investing career and perhaps our lifetimes. I am also assigning greater probability to AI having a meaningful positive or negative impact on the economics of individual businesses.
How this relates to our investments
Many of the companies we own continue to grow earnings and I expect will continue to create value. Their competitive advantages have not disappeared simply because artificial intelligence has captured investors’ attention.
At the same time, several are not viewed as obvious AI beneficiaries, and some could eventually face real disruption.
Meanwhile, much of stock market performance has come from companies perceived as direct beneficiaries of AI. Our direct exposure to those businesses beyond Alphabet (GOOGL) has been relatively limited. At times, I have even caught myself rooting for AI-related investments to perform poorly.
That was revealing. I do not want to chase short-term performance. Strong recent stock returns are not a sufficient reason to make an investment.
But rooting against a technology that I have become increasingly enthusiastic about seems inconsistent. That does not mean the answer is to buy whatever the market calls an “AI winner.” It does mean I should continue to reassess how AI could affect both the businesses we own and the opportunity set ahead.
There is a plausible world in which AI meaningfully increases productivity, changes how knowledge work is performed, accelerates scientific discovery, and transforms how software is created and used.
That does not mean every AI-related investment will be successful. One of the most important distinctions in my thinking is between being right about the technology and being right about an investment.
Enormous demand does not guarantee attractive investment returns. A business can experience extraordinary revenue growth while producing disappointing returns on the capital required to generate that growth.
What am I doing about it?
As I assign a greater probability to AI having a meaningful impact, I expect the portfolio may evolve in three ways.
First, I will be more mindful of companies where AI could weaken the underlying business and where the valuation does not adequately reflect that risk.
Second, I am open to owning more companies that could benefit if AI meets or exceeds current expectations, particularly businesses with durable competitive positions that do not depend on correctly predicting one specific technology or model.
Third, I am increasingly interested in companies that use AI to strengthen their existing competitive advantages. Some of the biggest long-term beneficiaries may ultimately be the companies that use the technology best, rather than those building it. We may be extremely early in this process.
None of this changes the core investment process. I will continue to rely on deep fundamental research, focusing on competitive advantages, management, reinvestment opportunities, balance sheets, valuation and long-term earnings power.
I will remain patient where appropriate. Companies won’t be sold simply because they aren’t perceived as AI winners. But if AI changes the economics of a business, I will change my mind.
The portfolio will reflect my evolving view of how the world is changing without abandoning the investment principles that got me here. I am excited about the opportunities ahead.
Best wishes for an amazing holiday season,
Jeremy Kokemor
Editor’s Note: The summary bullets for this article were chosen by Seeking Alpha editors.
