Published in the San Diego Union-Tribune, February 5, 2024
by Neil Senturia
The average adult makes about 35,000 remotely conscious decisions each day. No wonder you are exhausted by the end of the day.
All right, how many of those did you get wrong or did you regret? The tofu salad was clearly a bad choice.
Can artificial intelligence help you get it right more frequently than not? The chatbot tells all, but you know you can never go wrong with a double cheeseburger from In-N-Out. No brain damage there.
Decisions. Do we make them with certainty or simply put our finger in the air? Leaders are supposed to inspire confidence in their decision-making. They have reviewed the landscape, surveyed the defensive alignment, and are ready to go forward. Oorah.
However, volume is not the same as vision. That person with the loudest voice, who speaks with the surest sense of confidence — what are the odds that is the person with the best decision?
When we are sure we are right, when we have high confidence, do we really, or are we fooling ourselves? To paraphrase Descartes, I speak loudly, therefore I am right. This is often followed shortly after by, wow, I never saw that coming.
Harvard professors Thomas Graeber and Benjamin Enke studied 70,000 decisions and found that “self-assurance does not necessarily reflect actual skill.” Understanding this is critical in an organization that depends on strategic decision-making and innovation.
My personal view is that bombast usually indicates exactly the opposite — lack of self-confidence in the decision in that moment. The level of certainty probably decreases as the decibel volume increases.
Graeber says that good decision-making is closely tied to “whether the right people are confident or not.” When unpleasant surprises show up in your organization, that means the “confidence calibration” needs to be fine-tuned and reset. Is the company listening to the loudest voice or the wisest? And the fancy title won’t save you. It’s what you don’t know that you don’t know that will kill you.
Now, to the topic du jour for 2024 — artificial intelligence. The question is how does AI intersect with confident and correct decision-making. How loud will its voice be?
The consensus seems to be that the best answers will be created by “collective intelligence” that will blend human ingenuity with AI’s efficiency and analytical power.
Sounds right but let’s ask some experts.
Wharton professors Eric Bradlow and Christian Terwiesch have identified the four basic food groups (for venture investing). Opportunities and risks, strategy development, organizational structure and value capture. Can generative AI solve any of these problems? Speaking as a politician, the answer is yes and no.
Bradlow is clear that market research is the most critical area in the venture racket. Always step one, does anyone care? For this, the entrepreneur needs to leave the cubicle, pound the pavement and talk to the customer. And “large language” models do not have shoe leather.
But AI’s best strength is in generating better opportunities. Terwiesch ran a rigorous competition — GPT-4 vs. his MBA students. After comparing thousands of ideas in the Wharton database, his study shows that “large language models are better at creating high variance innovations (the ones that can change the world) than the MBA-generated ideas.”
The AI bot beat the kids 35-5. It didn’t just generate an idea, it evaluated it as well. You might want to ask ChatGPT if that $140,000 tuition is truly best spent in the halls of academia.
Where generative AI fails is in predicting the “commercial viability” of an idea. It is not yet built to predict the odds of success. It lacks the ability to factor in the incalculable human element. In the startup game, success cannot be reduced to an algorithm.
So, OK, go to business school after all.
Ending on a serious note, AI is the new frontier. It will change simple things like writing a patent, and it will inform massive changes like solving treatments for cancer. There will be prospectors, and there will be folks who make the picks and shovels. Both are looking for gold.
Generative AI cannot yet “think” — so for a while that gives the entrepreneur at least a running start. But the wave is coming, and your best hope is to get aboard and ride it to shore.
Rule No. 792: AI surfs the 100-foot wave.