ChatGPT at Ezra
I’ve been playing around with ChatGPT, and it’s truly stunning how powerful these models are. We started paying $20 per month at Ezra to experiment with ChatGPT-4 and get a sense of how it can help us in research, analysis, creative work, product development, and everything in between. Here are three ways I’ve used in the last two days:
- Summarizing articles and other long-form writing into short summaries (very skilled)
- Taking a first draft at this Insights copy (poor performance; scrapped everything it wrote and started fresh)
- Feeding it investment memos and asking if it suggests making an investment or not accurate, but not nuanced)
The promise of AI
Artificial intelligence has the potential to disrupt everything from industries to labor markets to even our own understanding of what is intelligent and original.
- Goldman Sachs economists predicted that AI could automate 18% of the global labor force.
- For startups, generative AI has the potential to reduce payroll expenditures, increase productivity, automate processes, and make teams more efficient.
The new baseline
Effectively using AI will become new baseline skills for startup leaders and employees. Six months ago, the ability to summarize a 10-page memo into three paragraphs required careful reading and synthesis. Today, it requires 1) a thoughtful prompt of what you’re looking for, and 2) the ability to refine and enhance the algorithm’s work-product. There are a number of resources aimed at how to better leverage AI models; I’m highlighting a few below.
Artificial Intelligence vs. Human Wisdom
But it’s a mistake to think that artificial intelligence is a substitute for human wisdom. AI models might help generate analyses, make recommendations, and summarize complex ideas, but in my experience as a CEO and coaching entrepreneurs over the years, the biggest drivers of a company’s success can’t be automated away by technology.
There is no technological substitute for wise, effective leadership. From applying good judgment to maintaining sharp focus to deeply listening to the people on a team, the quality of a company’s leadership will always be a far bigger predictor of success than their adoption of technology.
If I had to choose between getting better at prioritizing or getting better at productivity, I would choose prioritizing every single time. AI models unlock major constraints, but they don’t unlock the constraint of applying good judgment into complex situations and leading courageously.
How to apply good judgment as a startup leader
- Carve out the space for good judgment. Enemy #1 of applying good judgment is the speed at which you’re moving. Slow down and create intellectual spaciousness. Otherwise, as our Chief Investment Officer, Shujaat Khan, says, you’re just applying “advanced common sense” to the situation.
- Recognize trade-offs: Every yes in one direction is a no in a dozen others. One way to improve the quality of your decision-making is to raise your awareness of the trade-offs in every choice. The tradeoff is usually easier to see when you're saying "no" to something than when you're saying "yes". But it's magical thinking to believe we can do it all.
- Kill your darlings: If it’s not absolutely essential, consider ways to de-prioritize. The leadership author John Maxwell says, “You cannot overestimate the unimportance of practically everything.”
- Zoom-in / zoom-out: One of the hardest parts about being a leader is needing to switch between the 30,000-foot vision and the 3-foot details. But this change in altitude also allows leaders not to get stuck obsessing over something unimportant. Use this perspective shift to make better decisions.
- Ask better questions: The quality of your questions dictates the quality of the answers. Slow down and consider ways to ask better questions to the people on your team. One of my favorite questions as a leader is to ask my direct reports: “What is one way I can be more effective for you?”
- Name your biases: You can increase the quality of your decision-making by acknowledging the biases you're (probably) carrying. Two I look out for: confirmation biases and recency biases (where I am the product of my last conversation).
- The reflective leadership question: My coach Jerry Colonna has a beautiful question for leaders, which AI will struggle to answer: "How am I complicit in creating the conditions I say I don't want?"