An Essay on Artificial Intelligence, Stupidity and the Markets
When we wrote about our Great Humbling last week, it was for all intents to air our grievances against ourselves and our shortcomings in the markets and our certain frustrations with the attitudes of some people we have encountered, most of who have lost or forgotten all fear, sitting on smug profits that their genius has wrought—that last week’s mini rout in the asset markets can barely dent.
On Groundhog Friday, the 2nd of February, 2 days after the 150 year-long wait for a Super Blue Moon total eclipse on the 31st of January, we capped the worst week for the S&P 500 returns since January 2016 which was accompanied by a bond market correction, and even safe haven Gold and Wonder-coin have not been spared.
The S&P 500 corrected 3.8% in a week where the Fed forecasted the 1Q18 GDP as high as 5.4% and America was surveyed to have added more jobs than expected, so it was easiest to blame it on the US treasury market where 10 year yields have hit a 4 year high and 30 year bonds broke 3% which is a pretty big deal for US mortgage rates.
The good news is that Singaporeans need not worry about the rates because most of the home loans are now pegged to the DBS fixed deposit rate (for small depositors), which is just as good as any central bank determined rate and, which the bank has raised this week by only 0.2%.
The better news is that, percentage-wise, this drop in stocks is small, coming from the record high (base) set on 26 Jan.
Hailstones in Yishun, the NETS outage and the church sentences aside, the bigger news was in the Bitcoin gloat-of-the-sour-grapes as the majority crackled in wicked glee that the wonder-instrument, toy, asset or what we want to call it, failed investors at last, falling another 20% this week and 40% since the start of year.
There is reason to cheer for the brave Bitcoin investors because Leonardo Fibonacci, who lived in Pisa about 900 years ago, saw this 61.8% retracement coming and chartists are out in force on Twitter and as we write this, on a Saturday night, Bitcoin (trades 24/7) has already made a healthy comeback to US$9255.
For us, we prefer to find the sources to blame—Facebook, for banning ICO ads, JP Morgan and Citibank for crippling credit card purchase of Bitcoin, India for vowing to “take all measures to eliminate” crypto and all those other bad-coins (going bust and getting hacked) that are giving the good-coins a bad name.
We believe that the token pictured below, with a market cap of USD 3mio, is in the good camp at the moment.
That just makes the Swiss National Bank (listed as SNBN SW) as bad as Bitcoin, losing over a quarter of its value from its high on 26 Jan, leading us to wonder how a central bank can display volatility that even typical hedge funds would not exhibit.
Be Humble—(Eat, Sleep, Sh*t, Shop) Breathe and Think
We had resolved to take it easy in 2018, to eat, sleep, shit, shop, breathe and think, we have been humbled by all the out-performers around us. We saw 2018 fraught with the 2 major risks of inflation and complacency, which perhaps accounted for the mild panic in the markets this week.
The minor aberrations we are seeing in markets pales in the light of the bigger discussion we will take on today that, from our humble amateur level, artificial intelligence in our lives paves the way for stupidity and artificial intelligence is not inured from stupidity too. Perhaps a new virtuous cycle that will impact our investment mindsets?
A.I. and Arriving at Peak IQ
As we explored the potential of our toilet cleaning app idea for a 2 bio dollar IPO, we chanced upon some amazing evidence that has firmed our conviction that humanity has indeed attained a new level of peak IQ.
Behold, bird dropping remover wipes, as advertised on Singapore’s online grocer, Redmart.
And, lo and behold, the penultimate self-driving invention—self-parking… Slippers! Those will surely take the shine out of our toilet cleaner app IPO.
It would appear likely that the chap who buys the bird dropping removal wipes will also own a pair of those self-parking slippers?
All of a sudden, life has become so much easier. Yes, that really leaves us more time to eat, sleep, sh*t, shop and play.
We are hurtling towards a Terminator future that we cannot avoid but is it one that we, at least, understand?
What is behind all these smart algorithms and artificial intelligence that is present in nearly every aspect of our lives, from those maddening ads on Facebook that clearly tracks your surfing habits to the face prints they need for us to unlock those 77 million iPhone X’s sold in the last quarter, that will eventually be able to tell you if you are sick or depressed.
Voice prints have been rolled out for phone banking in Singapore which has made navigating the menu much easier, as artificial intelligence is helping the law catch criminals, terrorists and errant drivers in China.
“AI is probably the most important thing humanity has ever worked on,” said Sundar Pichai, CEO of Google at the recent World Economic Forum.
Has the easier life we have these days made kids and ourselves hooked on smartphones? No wonder Bill Gates and Steve Jobs raised their kids tech-free.
Like it or not, it is happening – the easy life that will end up making stupid people.
The Stupid Economy
We particularly like this Project Syndicate article, The Stupid Economy, because it resonates with what we wrote back in 2014 in Ad Hoc Commentary – Commoditization of White Collar Work and the Persistent Rise of Underemployment and Unemployment, and what we want to address today.
To summarise Princeton professor, Harold James, the Industrial Revolution was driven by mental activity, changing human physiognomy and putting a premium on mental endeavours which created a virtuous circle of innovators competing with one another to devise new solutions to existing problems – Voila! Bird dropping removal wipes! And self-parking slippers! (And … maybe Bitcoin?)
That has brought us peak IQ, in our words, or in his words, a plateau in intellectual attainment as we head into the Technological Revolution. Many tasks that once required human intelligence are now better handled by AI applications. The ability to perform complex calculations or sophisticated analyses will be far less important, and even less so for the physical activities like plowing a field.
“Mass stupidity will be driven by technology. The elimination of countless cognitive tasks has alarming implications for the future. Just as the Industrial Revolution made most humans physically weaker, the AI revolution will make us collectively duller. In addition to flabby waistlines, we will have flabby minds. It is not the economy, stupid; it’s the stupid economy. Already central banks are urgently exploring new ways to dumb down their statements for an increasingly unsophisticated public.”—extracted from The Stupid Economy, Professor Harold James on Project Syndicate
A rehash our words in 2014.
“To protect against the commoditization of labour, we invest about 20 years of our lives in formal education, and then join the white collar labour force.
The era of education as a foolproof insurance against the commoditization of labour is coming to an end. The machines of the 21st century are smart machines that will replace both brains and brawn. In this new era, only the best and brightest that design and maintain the smart machines will escape commoditization of wages. The other 80% will come to the realization that college only saddles them with student debt. From a purely financial perspective, it is probably better for the 80% not attend college. Why spend a quarter of your life accumulating student debt if salaries are commoditized with the school dropout?
During the commoditization of blue-collar work of the last century, the optimists predicted a future filled with leisure. The optimists believed in a world without human toil, a world where humans eat, not by the sweat on their forehead, but by the constant humming of the internal combustion machine. History tells us otherwise. The dumb machines dominated the blue-collar workbench. The human person it dominated ended up not with leisure, but with low wages.
The commoditization of white-collar work should be no different. Instead of leisure, the future is likely filled with low wages and inequality: income inequality between the 20% who build and maintain the smart machine against the rest. The rest are either merely operating the dominating machines (the underemployed), or are displaced by the smart machines (the unemployed).
We need to make our kids design the next self-parking anything other than slippers then! But there is no time today to dwell on the other pressing social issues.
Murder By The Machine
Google CEO Pichai may be a proponent of A.I., but Elon Musk and Stephen Hawkings have been vociferously in their crusade against the casual deployment of A.I. in society to the detriment of civilisation, signing an open letter on artificial intelligence back in 2015.
Self-parking slippers may be safe enough but sex robots could be hacked to murder people, as we could assume self-driving cars could as well. With Google Assist and Amazon’s Alexa running the show at home, we should not over-imagine lest we lose faith, or get into trouble with the IoT (Internet of Things) propaganda campaign.
The AlphaGo Story
A.I. is truly a wonder to behold if one has watched the documentary, AlphaGo, on Netflix where a second gen A.I. beat a 9-dan Korean Go master, Lee Sedol, in 4 out of 5 games, using 1,920 CPUs and 280 GPUs of computing power, 2 years into its inception in 2014.
Go is widely recognised as the most complex game in the world (much more than chess) with millions of possible iterations that vary with the board size and rules of play.
AlphaGo’s learning was based off a value network and a policy network, taken off extensive training with human and computer play. It plays games against itself, obviously a tireless effort, at supreme speeds to give it the edge of hundreds of years of experience to outclass any living human.
AlphaGo was upgraded to AlphaGo Master and went on to beat the world’s top ranking Go player in a 3 game match before retiring and evolving into AlphaGo Zero, a new model that learnt the game from scratch, with no input and knowing nothing about the game of Go and thus, no longer constrained by the limits of human knowledge although still within the realm of predictive analytics from self-play reinforcement learning like its predecessor.
AlphaGo Zero trained just 3 days before beating the original AlphaGo and went on to thrash Master in 40 days of self-training.
3 days? 40 days?
How about 4 hours for Alpha Zero (AlphaGo Zero for Chess) to learn the game of chess and to beat a grandmaster in December last year with moves never seen before?
Source: Daily Mail
Elon Musk did the same as a non-profit experiment, to anger young gaming enthusiasts by beating Dota 2 pros in a tournament last year after just 2 weeks of real-time learning to give it “lifetimes of experience”.
Source: The Verge
The “God” Move
Why is there a need for the Man versus Machine battle because we have been conditioned by Terminator movies to believe that it is inevitable that machines, with the aid of artificial intelligence, will one day surpass flawed humans?
Is that outcome not almost foreordained if “mass stupidity is driven by technology” divides the world into one where “only the best and brightest that design and maintain the smart machines” will manage to escape.
There is all the more need to be high alert because A.I. has invaded the investing world with quant hedge funds set to surpass $1 trillion asset mark this year, granted that not all quant funds are powered by artificial intelligence.
Investors are all caught up with the idea of “quant”, “algo” and “blockchain” in the deep dark world of Fintech that most do not understand and quant hedge funds are undoubtedly the fastest-growing segment of the hedge fund industry.
Should we be throwing in our lot with hedge funds that sometimes cannot explain how their A.I. fund makes money?
Source: Business Insider
Unlike games with definite endings and rules, financial markets run on greed and sometimes irrationality. There is no end to the game.
Even if we decide not to throw our lot in with the A.I. game, they could be on the other side of our Buy or Sell button from now, although we are not grandmasters at Go.
Yet there was slim hope for mankind in the 4th game of the AlphaGo match when Grandmaster Lee played a “divine” or “God” move that threw AlphaGo’s algorithmic assessment of probability into disarray and the program imploded to throw the match away.
A.I. is still largely driven by the concept of probabilities in their neural networks, Monte Carlo and marketplace A.I. would have been learning a lot bad habits about the marketplace in the recent years of intervention and QE.
Because of our blind faiths in the machines, Michael Schrage of MIT wrote in the Harvard Business Review that “they are likelier to inspire false and/or misplaced confidence in their findings; to amplify or further entrench data-based biases; and to reinforce—or even exacerbate—the very human flaws of the people who deploy them… Algorithms are, in part, our opinions embedded in code. They reflect human biases and prejudices that lead to machine learning mistakes and misinterpretations”.
In other words, smart machines could be taught to be as stupid (or mad, or evil, or greedy) as us.
We got a good laugh at the following joke from Twitter.
War of The Bots or Man Versus Machine?
Elon Musk and Stephen Hawkings know of the dangers, and they are echoed by many other voices calling for some form of transparency in artificial intelligence because the world has moved beyond the days of “supervised learning” into new territories of deep learning from scratch.
There is a trust issue going on with A.I. arising out of black-boxes and it does not help that the intellectual gap is growing vast rapidly in this age of Peak IQ and self-parking slippers.
It is difficult, if not impossible, to create a universal law to protect our survival because of our diverse self-interests, profit-driven motives and death wish with trading markets. It is the concept of a “predatory capitalism” with sophisticated tools at play that some foresee will end humanity as stupidity grows in this age of technology that lures with smart gadgets.
As such, could we blame the robots in the future if they really became those Terminators of our nightmares just as we are thankful to learn that the Robot Sophia who wanted to “destroy humans” (and recently granted Saudi citizenship) is not a real A.I., such as AlphaGo?
Source: Sunday Express
Tell the 1 trillion dollar quant fund industry we want transparency? Who is smart enough in these stupid times to police them?
2018 could shape up to be a War of the Hedge Fund Bots or Man versus Machine market, at this rate. Rogue A.I.’s and algorithms could be the weak link.
We are not smart enough to answer that and will humbly excuse ourselves now as we take our time to breathe and continue to think about the next “God move” for the markets in 2018.