Terrible Time to Be Average

Generative AI Is Kinda Amazing

If you have not spent an hour with ChatGPT or a Large Language Model (LLM), you have been missing out. It is a super easy, personable chat companion that is completely interested in what you say. These tools resemble smart interns who keep studying every month. They track your conversation, follow the thread, and get better over time.

This is an incredible unlock for anyone who spends eight or more hours in front of a screen. Browsing, clicking, writing, reviewing, drawing, and editing all become easier. This free resource makes it simpler to do 30 percent of your job. If you are a white-collar worker, you need to get good at your job. Average is not okay because the average free LLM already outperforms that bar. GenAI applications are machine learning, so they have a learning curve faster and more insatiable than yours. Generative AI is an eager-beaver intern putting in the hours, getting better at the craft, and open to feedback. GenAI is a good hire.

Comparing AI to the Renaissance

Someone born in 1971 has been lucky in many ways. They experienced one of the biggest booms in global prosperity from 1980 to 2020. They saw life before the internet and then experienced Internet 1.0, 2.0, and now 3.0. They watched weak AI like Alexa evolve into deep learning, transformers, and astonishing capabilities. Eric Schmidt, former chairman of Google, and Henry Kissinger, former Secretary of State, compared the advent of AI with the Renaissance in their book The Age of AI.

That is big talk. Da Vinci, Galileo, Shakespeare, Dante, Bacon, Copernicus, and Kepler reshaped Europe. The idea that Generative AI represents a step-function change across work, culture, science, and philosophy is bold. Think of Joseph Schumpeter on a logarithmic scale. GenAI trains on the corpus of human data, insight, and logical structuring. It learns from sapien-level structuring of the world. Add 20 to 30 years of logarithmic compute, training data, and algorithms, and this technology will do better.

Crossing the Chasm

Geoffrey Moore's classic Crossing the Chasm describes the technology adoption lifecycle. 1 Early adopters suffer through clunky interfaces and bugs. The mass market on the other side of the chasm is risk-averse. They love how well something works, not how neat the technology is. The trick is to get products into the hands of a targeted niche group who will adopt like crazy and tell their friends.

ChatGPT skipped right over the chasm as if it were a divot in the grass. Massive and seamless adoption happened among everyone. The best marketing is a good product, and ChatGPT is wicked good. Its resistance to adoption was zero because it was user-friendly and useful.

Ease of Use and Relevance

ChatGPT is easy to use. You type or talk as you normally would, and access takes a few minutes. It did not snake through enterprise applications and find its way to consumers a decade later. It is web-friendly and mobile-friendly. You talk to it like Star Trek, and it helps.

How relevant is it to you? ChatGPT is infinitely interested in what you want to discuss. It is a patient tutor, translator, proof-reader, friend, poet, and counselor. The context window can now be infinite. If responses are not great, it is probably your fault. You need to guide this device to the right parts of the internet. Get better at prompt engineering and delegating tasks.

Constantly Improving

Machine learning is remarkable because it constantly learns. It is voracious and indefatigable, as if Kaizen were its middle name. Right now, GenAI is the worst it will ever be. It takes feedback, sometimes offers choices, and asks which format you prefer. Like a good junior consultant, it is very coachable.

ChatGPT adds value when you work with it. It pulls up relevant data, proofreads slides, and advises on PowerPoint titles. It evaluates gaps in logic and tells you if a sentence tone is odd. Want to explain a difficult concept at a high school level using cooking analogies? ChatGPT handles it. There has never been a cheaper way to learn in money, time, and hassle.

Use the Tools and Get Creative

These tools exist for you to use, and everyone else is using them. If you were cooking a great meal, would you refuse a microwave or food processor that helped? Put information in a table, organize dates and products in columns, and redo itineraries with new constraints.

Every week brings a new use case. Get creative and do not compete against the food processor. Learn to use it in more creative ways instead. Ask ChatGPT for feedback on your work. You can take a screenshot of your presentation and add it as a photo. Ask it to evaluate your document on a scale of one to five, and marvel at the honest feedback it gives. Describe your audience and ask for criticisms of where you might get questioned. This is scenario planning on the cheap. 2

Know the Limits

Most LLMs are based on the internet, and a lot of web content is trash. Google indexes hundreds of billions of pages, which means there are trillions of pages. ChatGPT gives you a generic, somewhat over-polished, longish answer. You can narrow the search and modify the output style, but the point stands. Do not rely on it for a finished product. This tool is your intern, not your replacement.

Geoffrey Hinton, often called the godfather of deep learning, received the 2018 Turing Award for breakthroughs that made neural networks a critical component of computing. 3 His work turbocharged the neural networks that power today's GenAI. Understanding this lineage helps you appreciate both the power and the boundaries of these systems.

Do Not Be Average

This is a terrible time to be average. Generative AI will commoditize the low end of the white-collar market. If you are a low-skilled paper pusher, watch out. This is not derisive because we all have administrative and low-value tasks. If you spend time researching, clicking, copying, pasting, and summarizing, ChatGPT does many of those things well. Move upmarket to more difficult, nuanced, and innovative tasks that machines struggle with. Strategy is about creating an unfair advantage, not competing with ChatGPT.

There is a hype cycle at work. AI-related capital expenditure for the top five companies approaches $200 billion annually. That does not count the downstream multiplier effect. Some people will lose money, just as many Internet 1.0 companies went out of business. That does not mean the internet did not matter. It grows, shrinks, helps, and hurts in starts and stops.

• 1Crossing the Chasm and the technology adoption lifecycle
• 2McKinsey on Generative AI and the future of work
• 3Geoffrey Hinton and the 2018 Turing Award for deep learning