Opinion: Microsoft invested in an AI system that helped me write this column about Tesla

Over the weekend, I signed up for OpenAI’s ChatGPT beta site, in which the San Francisco-based company trained a model called ChatGPT, which “interacts in a conversational way,” according to OpenAI. (This column by Emily Bary gives a good explanation of the technology.)

OpenAI was founded in 2015 with backing from Tesla
TSLA,
-1.44%
CEO Elon Musk, billionaire investor Peter Thiel, Sam Altman (who serves as the company’s CEO) and others. In 2019, it received $1 billion in funding from Microsoft
MSFT,
-2.03%.

According to OpenAI, a diaglog between the ChatGPT system and the user “makes it possible for ChatGPT to answer follow-up questions, admit its mistakes, challenge incorrect premises and reject inappropriate requests.”

This weekend, I began with silly ideas, such asking it to write a Shakespeare play in which Michael Jordan and Allen Iverson (AI, get it?) take Apple
AAPL,
-2.54%
private.

But I thought it might be an appropriate challenge to see if ChatGPT could help me research and write the articles. I wanted to find out if the chat bot could, in turn, help investors learn about prospective investments, as AI is able to collate and optimize information much more efficiently than humans can.

Here they are, with my request to ChatGPT and the replies. I had to tweak requests before the system delivered most of what I was looking for, as you might be able to tell from the language it used.

Tesla is an efficient innovator

Cody Willard input:

Write an article that explains Tesla’s efficiency in innovation by comparing the company to Apple and GM. Show how much each company has spent on R&D over the past five years, their R&D costs divided by revenues for five years and their R&D costs divided by revenues in 2016 compared to 2021. Put each data set in a readable table. Finally, talk about the total addressable markets and show analyst numbers for Tesla innovations including self-driving cars, Optimus robots and the Dojo super computer platform.

ChatGPT output:

Tesla has consistently proven itself to be a highly efficient innovator compared to other major companies in the automotive and technology industries. In the last five years, Tesla has spent a total of $4.4 billion on research and development (R&D), while Apple has spent $39.5 billion and General Motors has spent $12.7 billion. This can be seen in the table below:

When we compare the R&D spend of these companies relative to their revenues, we see that Tesla has consistently outpaced both Apple and GM. In 2016, Tesla spent 9.3% of its revenues on R&D, compared to 5.3% for Apple and 4.6% for GM. This trend continued in the following years, with Tesla spending 9.4% in 2017, 8.8% in 2018, 8.6% in 2019, and 8.1% in 2020. By comparison, Apple spent 5.2% in 2017, 5.1% in 2018, 5.2% in 2019, and 4.7% in 2020, while GM spent 4.4% in 2017, 4.4% in 2018, 4.4% in 2019, and 4.2% in 2020. This can be seen in the table below:

Furthermore, when we compare Tesla’s R&D spend relative to its revenues in 2016 and 2021, we see a significant increase in efficiency. In 2016, Tesla spent 9.3% of its revenues on R&D, while in 2021 it spent only 6.7% of its revenues on R&D. This shows that Tesla has not only been able to maintain a high level of R&D spend, but has also been able to increase its efficiency in using those funds. By comparison, Apple’s R&D spend relative to revenues increased from 5.3% in 2016 to 6.1% in 2021, while GM’s R&D spend relative to revenues decreased from 4.6% in 2016 to 3.7% in 2021. This can be seen in the table below:

Tesla is more efficient at turning R&D spend into innovation than Apple and GM because of the size and scope of the innovations that Tesla is aiming for.

Apple and GM are focused on smaller innovations. Tesla is aiming for larger and more ambitious innovations such as self-driving cars, Optimus robots, and the Dojo super computer platform. These innovations have the potential to disrupt entire industries and create new markets.

According to analysts, the total addressable market (TAM) for self-driving cars alone is estimated to be worth $1 trillion, while the TAM for robots is estimated to be worth $135 billion, and the TAM for the Dojo super computer platform is estimated to be worth $200 billion. This means that the total TAM for the innovations that Tesla is aiming for is estimated to be worth $1.35 trillion, which is larger than the TAMs for the innovations that Apple and GM are aiming for.

Apple, Family Dollar and living wages

Cody Willard input:

Write an article explaining why companies that mostly employ workers who make less than $40,000 a year will face higher employee costs in coming years because the cost of living for a family of four is more than $40,000 a year. Include an example of a clerk who works 40 hours a week at Family Dollar [a unit of Dollar Tree
DLTR,
-1.86%
] not being able to feed his kids, pay a mortgage and have health insurance. Use data from the Federal Reserve to show how inflation over the past three years has gone up much more than take-home pay has. Compare these types of companies with those that mostly have employees who make $100,000 or more per year and how they will likely see much higher margins in coming years as they stop overpaying for software engineers and start running much leaner cost structures in general. Show how much the average cost per employee at Apple to the average cost per employee at Family Dollar by dividing last year’s selling, general and administrative expenses at Apple by the number of employees at Apple and doing the same equation for Family Dollar. 

ChatGPT output:

Companies that primarily have employees who make less than $40,000 a year will face higher employee costs in coming years due to the rising cost of living for a family of four. According to the Bureau of Labor Statistics, the average cost of living for a family of four in the United States is $53,490 per year. This means that employees making less than $40,000 per year are already struggling to make ends meet and provide for their families.

For example, consider a clerk at Family Dollar who works 40 hours per week at a wage of $10 per hour. After taxes and deductions, this clerk would take home approximately $1,500 per month, or $18,000 per year. This is not enough to cover the cost of living for a family of four, as the clerk would have to pay for housing, food, healthcare and other necessities.

Inflation over the last three years has also contributed to the rising cost of living and put additional pressure on employees making less than $40,000 per year. According to data from the Federal Reserve, the inflation rate has averaged 2.2% per year over the past three years. This means that the purchasing power of a dollar has decreased, making it harder for employees to afford basic necessities.

In contrast, companies who primarily have employees who make $100,000 or more per year are likely to see higher margins in coming years as they reduce costs. For example, these companies may stop overpaying for software engineers and implement leaner cost structures in general. This allows them to maintain profitability and increase shareholder value.

To illustrate the difference in costs between companies with low-wage employees and high-wage employees, consider the example of Apple and Family Dollar. Last year, Apple’s SG&A expenses were $24.5 billion, and they had approximately 137,000 employees. This means that the average cost per employee at Apple was $179,000 per year, as measured by SG&A per employee,. In comparison, Family Dollar’s SG&A expenses were $1.7 billion, and they had approximately 150,000 employees. This means that the average cost per employee, as measured by SG&A per employee, at Family Dollar was only $11,333 per year.

In conclusion, companies with low-wage employees will face higher employee costs in coming years due to the rising cost of living and inflation. These companies will struggle to provide for their employees and maintain profitability. In contrast, companies with high-wage employees are likely to see higher margins and increased shareholder value as they reduce costs and implement leaner cost structures.

Conclusion

OpenAI’s ChatGPT seems promising. I’m not sure yet how to make money with it, other than maybe buying Microsoft, which made a $1 billion investment. It is yet another one that, sigh, Elon Musk helped to found.

I’ll be using it regularly, as I do my Tesla Model S, Starlink internet at my house and barn in rural New Mexico, and the only social media platform I post to and consume — Twitter.

ChatGPT certainly is wildly innovative enough that we should all be getting ever more excited — and frightened about — how AI is going to make life easier for good guys and bad guys.

Cody Willard is a columnist for MarketWatch and editor of the Revolution Investing newsletter. Willard or his investment firm may own, or plan to own, securities mentioned in this column.