Bubbles - AI set to take a bath

Are we living in an Artificial Intelligence (AI) bubble which is about to burst? It is a question many are asking so it seems like a good time to take a step back from all the hype and reflect on lessons from the past, and the part human nature plays in all this. 

The first bubble I covered as a journalist was the 1987 stock market crash when the Dow Jones Industrial Average collapsed 22 pct in a single October day, which is still the largest ever single-day percentage drop.

There have been many bubbles including the Tulip Bubble in the 1630s, the Railway Bubble of the 1840s, the debt-fuelled Asian Crisis bubble of 1997, the Dot-com Bubble around 2000, the Cryptocurrency Bubble of 2017 and the Non-fungible Tokens (NFTs) bubble which burst in 2022. 

This is by no means an exhaustive list, but bubbles basically form as a result of two simple factors. Fear and Greed. Fear of losing out and greed in what is perceived to be an easy route to riches. Human nature takes over as rationally flies out of the window.  

There have been other market collapses such as the 2008 crisis which don’t strictly meet the criteria of a bursting bubble. Although even then you could argue that it was a bubble in terms of everyone wanting to invest in property which was, so they thought, an asset that would always go up in value. 

I was in Singapore when the Dot-com Bubble burst. I was heavily invested in tech-funds and riding the huge rally. My main fund was up 137 percent in just over a year but I took my profits at the end of 1999, much to the chagrin of my broker who said I was mad not to stay in the market. The market collapsed soon after with many analysts urging people to buy the dip. This was appalling advice as the tech-heavy NASDAQ continued to decline for the next two years. It only got back to pre-crash levels in 2014 - more than 10 years later.    

I got out because obviously unreasonable hype was circulating about what the internet could do. Claims that ignored not just reality, but human nature. There was a widespread belief that internet growth would be limitless, and that companies involved would expand exponentially.

What tipped the balance of my decision to exit came when I attended an in-house seminar and the presenter told me that I would soon be buying a refrigerator with a screen on the door connected to the internet which would automatically order butter or milk via the internet when I was running low. Apparently this smart fridge was coming to stores within a couple of years and this would help make my life perfect.   

I stuck my hand up and asked why that morning I had been sitting in a high-tech newsroom surrounded by the latest technology, my internet feed had been down for hours. This kind of ridiculous hype made no sense and I decided to take profits. Twenty-five years on, my current fridge can still only cool things down, freeze stuff and make ice.   

The internet did transform the way we live and many aspects of life, but when claims get exaggerated it is time to prepare for the bubble to burst. It is easy to see how those with a vested interest will claim almost anything to keep the investment money rolling in. This is why you can’t open a news feed without seeing Sam Alman, Mark Zuckerburg and their ilk pontificating on how AI is the absolute future for everything. (Ok Mark, we will quietly forget about the Metaverse disaster which was ‘the next big thing’.)

So we have an almost perfect bubble scenario with a new innovation which will make a material difference to the way we live, which has already created a huge flow of investment money on the back of fear of losing out and greed seeking easy large returns. All this is compounded by huge hype and unrealistic claims by those with a vested interest of what Artificial Intelligence can achieve in reality. 

Smart Money/Dumb Money

This where another financial market truth kicks in - Smart Money vs Dumb Money. Banks and large investors typify smart money, retail investors typify dumb money. Smart money has access to experts, analysts and market instruments that retail investors do not. They are pretty much fully invested by the time large fund managers get hold of the research and then their money helps keep the market rising.

Never lose sight of the fact that financial markets will only rise if money is flowing in, the rally stops when the liquidity flow stops. Rallies in any market run on liquidy, very often in all the hype, investors lose sight of that very simple fact. 

The end of a rally, such as that we are seeing now with AI, comes when the retail investors are sucked in with stories of untold riches mainly in the media and speeches by those with a vested interest in attracting more money to keep the rally going. 

Many years ago I was told by a trader “If you land in Hong Kong and your taxi driver tells you to buy a certain stock. When you get into the office, the first thing you should do is short-it.”

Which raises the issue that professional investors are able to play both sides of a market while, generally speaking, retail investors tend to only buy and hold as they don’t have access to short-selling instruments. It means retail investors are playing in the market with one hand tied behind their backs. It is an unfair fight and why money tends to flow away from retail investors into the hands of professional investors.  

Professional inventors are also able to track liquidity flows in the market and can see the end of any rally coming as the last vestiges of retail interest dries up. This when they aggressively take profits and short-sell, which explains why rallies are slowly paced and declines savage and steep.

The other issue which accelerates a market decline is margin calls, where investors who have borrowed money to buy stock turn into distressed sellers who are forced to sell stock to cover margin payments on the loans they took out. They have no option and have to sell. 

So such corrections can be extremely rapid and severe. Investors need to keep a close eye on their portfolios and not forget the Hong Kong taxi driver - if you are reading in a newspaper that xxx company is a good stock/bond to buy, you are already too late. Or, if you are invested, and the newspaper says buy, you should likely take profits.  

UBS, widely considered to be the largest and most sophisticated global investment banks, estimates that companies will spend $375 billion globally this year on A.I. infrastructure, and that is projected to rise to $500 billion next year. You have to question whether that level of investment is sustainable and also ask - where is that money going to come from? 

By their nature bubbles burst quickly. One day everything is fine, the next day it is not. The warning signs come slowly and are usually clearly observable, but fear of losing out and greed for profits clouds any rational viewpoint and the wagon keeps rolling. Until it doesn’t. 

The 1997 Asia Crisis was a classic in this regard. I was in Reuters Singapore newsroom on July 2 1997 when I received a call from the Bangkok bureau who read me a fax from the Thai Central Bank saying they had decided to devalue the baht. This came after months of speculative pressure on the fixed currency and precipitated the entire Asian Currency/debt/economic/financial crisis. On July 1 things were under pressure, but fine, by the next day all hell broke loose. These things happen quickly when the turning point comes.    

I believe the AI bubble will burst and when it does, it will be a fast and extremely rapid correction for the reasons outlined above. Retail investors beware and please keep a close eye on things. Wild claims about AI are not unlike Elon Musk’s claims of colonising Mars, at best, a long way off.  

It took years for the internet to really settle down and turn into something that was universally useful and helped companies become more productive. It was also years before the dark side of the internet revealed itself in terms of hacking, the potential harm social media can do, cyber bullying, spreading lies and conspiracy theories. 

Inflated expectations

It is possible to see a picture of this. It is called the Gartner Hype Cycle and while it has its critics, it demonstrates nicely what happened during the Dot-com bubble in the run up to it bursting in the early 2000s.

It has five stages -

New Technology - Research and development kicks things off and initial proof-of-concept stories find their way into the media and the hype starts to build. After all, everyone loves technology, right? Commercial viability is yet to be proved but everyone starts to believe - this is going to be huge.

Inflated expectations - The publicity becomes such that wild claims start to be believed, fear and greed kick in, venture capital pours in and it becomes difficult to persuade investors to put their money into anything else. The bubble forms at this point.

Disillusionment - Not all those who invested make money as the inflated expectations are not met and the technology is not as super-fantastic as it was hyped up to be. This is the “oh yes, it’s useful and will help, but we need time to implement it and we need to find more money” stage. Some early adopters fail.

Enlightenment - There is more evidence of its usefulness and productivity gains start to become real. The technology settles down, is more widely understood and next generation products come on line. Companies remain cautious.

True level of productivity - The technology becomes mainstream as early issues are resolved and viability become commercially proven and companies improve margins because of it. The slope from here depends on whether there are other or new adaptations for the technology or whether it is too niche.

 It is easy to see how you can apply this picture to the current position of AI in the global collective psyche. It tends to be confined to technology developments and certainly works for the way the internet Dot-com era panned out. 

Likewise, if you think about electric vehicles (EVs), it is turning out to be a similar picture. Early huge excitement and massive investment with governments falling over themselves to hand out huge subsidies and tax breaks and even passing laws banning the sale of petrol fuelled vehicles by a certain date. The peak of expectations was very clear. 

Then all the issues about range anxiety kicked in and the lack of infrastructure (charging points) coupled with the apparently unseen massive expense of changing batteries and the collapsing second hand value of EVs led to many early adopters being disappointed and disillusioned.

So having gone through the cycle, it would seem EVs are gradually approaching the Plateau of Productivity and things are settling down. There are still some big issues to be sorted out, but the world is getting there in terms of wider adoption of electric vehicles and governments are rowing back on dates for banning the sale of petrol vehicles. 

Elon Musk is a perfect example of a vested interest ramping up hype to ridiculous levels. “It’s financially insane to buy anything other than a Tesla,” he told a Palo Alto audience in 2019. 

“It would be like owning a horse in three years. I mean, fine if you want to own a horse. But you should go into it with that expectation.”

Like most of Musk’s shameless boasts and predictions, it has not aged well. But it perfectly demonstrates how those with vested interests are willing to say almost anything if it is to their benefit. 

For AI, it would appear clear that CEOs are believing the hype and the predictions of vast swathes of professions will disappear as a robot brian takes over their workforce's jobs. Free workers, what CEO could resist? (At this point I would be happy to argue with anyone who thinks just because someone reaches the heights of becoming a CEO, it automatically makes them a business guru. Look up the story of car hire company Hertz and the absolute shambles they made of piling into electric cars at the height of the Inflated Expectation curve.)

OpenAI CEO Sam Altman is telling us that 40 percent of jobs will be replaced by AI and predicted super intelligent AI in five years. Microsoft founder Bill Gates has confidently predicted that in ten years you won’t be able to see a doctor when you get sick as AI will have replaced him or her. He has also talked about teachers being unnecessary in this new world and AI will be doing that too. He even said humans will not be needed “for most things” in ten years time.

I'm cynical and don’t expect any of this to be a reality until I’m long dead. Sure it’s possible but is this going to happen as quickly as many predict? I doubt it very much. Like Musk’s almost constant ‘by the end of the year/next year this is possible’ predictions about several things - a lot of this is wishful thinking or just so much hot air.

Futurists

I remember well in the 1970s, so-called futurists were making a good living in predicting what the world would be like for us youngsters now at college. They were happy to go on TV and pontificate about robots doing all the manual labour and computers all the office jobs. When I was at college these people were telling me my biggest problem in life would be all the spare time I would have on my hands, as the huge productivity gains from free labour via robots and computers would mean I’d be paid for doing nothing. 

I remember thinking at the time…this is so cool, why the hell am I bothering with collage and exams? But my life has not been neverending leisure but slogs of 12 hour days and unpaid overtime for decade after decade.  

What they failed to predict was that while computers and robots have indeed made us all hugely productive with 100 car assembly workers replaced by a dozen robots with a single engineer keeping an eye on them. Salaries have stagnated in many countries while many companies have become hugely profitable on the back of increased productivity and created a world where billionaires now control governments.

Young people are already struggling to find work and salaries have stagnated to a level where buying a place to live is way out of reach for those exiting the education system. Many governments are so debt-ridden they can’t afford to help those who are not working. If Gates is right and humans will not be needed “for most things” he would appear to be predicting something close to the collapse of societies around the world.   

Yes, AI will be transformative and will affect the workplace but amid all the hype of the current bubble, the time for some more rational thinking must be close and the more crazy predictions will unwind. 

That will very likely lead to a shockwave through the financial markets. The trouble with that is it will drag the entire market down and impact everyone whether they are invested in AI or not.

Tinkerty Tonk