US technology stocks have added almost $5-trillion (R95-trillion) in market value in2023 on the rise of artificial intelligence (AI). Is this another bubble or the real deal?
I suspect it’s both. The field is moving fast, and uncertainty remains high. Predicting second- and third-order effects are mostly educated guesswork at this stage.
So, what is the hype about?
Large language models (LLMs) have reached the ability to pass human-level tests in a variety of subjects. If we project recent growth rates into the future, it’s plausible that artificial general intelligence (AGI) — AI equal to human intelligence across all fields — could be achieved in years to come.
If the hype turns into reality, then productivity could see a significant boost over time. Productivity growth becomes more important with ageing demographics, as GDP growth is a function of population growth and productivity growth. As populations stop growing in the West and China, these countries need productivity growth, or their economies will stagnate.
Where factory automation has been a threat to blue-collar workers, AI is more likely to be a threat to white-collar workers, which are more abundant in developed markets.
AI will initially replace some of our tasks and could eventually replace entire jobs. This creates significant fear about broad-based job losses and is indeed a possibility. However, technology optimists do point to previous technological disruptions, which grew the net number of jobs in society despite the global population growing from 2-billion to 8-billion since 1930. To be fair, if not for technological advancement, we would all still be subsistence farmers.
Views are divided on how AI will affect society.
Marc Andreessen, a Silicon Valley venture capitalist, believes AI will save the world, driving productivity and solving problems in healthcare, education and elsewhere.
Fake news and video will destroy our trust in truth. Social media, and Facebook and Cambridge Analytica specifically, provided us with Brexit and Trump.
Benedict Evans, an independent technology analyst, makes a strong case that job losses will take a long time, and at the same time new jobs will be created in fields we cannot yet fathom.
The philosopher Yuval Harari believes that truth will die and along with it democracy, as LLMs will be better storytellers. Fake news and video will destroy our trust in truth. Social media, and Facebook and Cambridge Analytica specifically, provided us with Brexit and Trump, so I take his point.
Henry Kissinger believes AI will supercharge the rivalry between the US and China and is hastening the need for diplomacy and pragmatism. I suspect this might be partially behind recent White House efforts to de-risk the US-China relationship to some extent.
Eric Schmidt, the former CEO of Google, fears the invention of autonomous weapons with no home country. AI is dual use, has destructive potential, and spreads easily via open platform models such as OpenAI. This is unlike nuclear tech, which is also dual use, has destructive potential and is strictly controlled.
Elon Musk believes it will either be the best, or worst thing ever, for humanity.
Prof Aswath Damodaran, a valuation authority at NYU Stern, believes AI will be a zero-sum game for most companies, as it is a general purpose technology. If you can use it, so can your competitor.
I disagree to some extent as I think companies with unique and regulated data sets could harvest greater returns from their data with the help of AI.
Discovery and Life Healthcare come to mind and are two such potential beneficiaries. Africa is a vastly underserved market from a healthcare perspective, and the ability to use AI to expand service reach is significant.
Life Healthcare has been acquiring radiology practices, which could provide it with unique data sets over time, and with the help of AI could significantly broaden the potential pool of profitable patients.
Another innovative SA company, Envisionit Deep AI, has succeeded in diagnosing and importantly validating imaging diagnosis in real-time at low cost, using AI. The company has a mission of broadening diagnostic healthcare. I was surprised to learn that Africa, a continent with 650-million children under the age of 18, has fewer than 25 paediatric radiologists, or one radiologist per 26-million children.
Technologically scalable solutions are required if the SA government, or the continent, wants to expand healthcare reach to its citizens, without bankrupting itself in the process.
This brings me to our next takeaway: AI has the potential to significantly benefit emerging markets that suffer skills shortages, such as SA. AI makes intelligence basically free, which could deal with our ever-growing brain drain.
Entities such as Khan Academy, a free education channel on YouTube, has collaborated with OpenAI to create an AI tutor. It is now being tested in the US and could become available globally over time, at the low cost of an internet connection. This has the potential to democratise education in the true sense of the word.
So, is AI a bubble?
There is no doubt that Nvidia, the pick-and-shovel supplier to this gold rush, which added $190bn to its market capitalisation overnight after a guidance upgrade, is most likely a bubble at 25 times forward sales. A point on which Damoradan agrees.
Companies with unique and well-labelled data sets might be a more fruitful area for stock picking.
We are entering a brave new world with much uncertainty, but also opportunity to broaden access to services to a larger set of the global population.
• Van Rensburg is a strategy consultant to All Weather Capital.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.