Alphabet (NASDAQ:GOOGL) has a reputation as masters of science fiction that’s on the verge of becoming real, from driverless cars to global broadband internet delivered by balloons. The company has come down to earth somewhat since the arrival of Ruth Porat, a Wall Street veteran, as the company’s chief financial officer in May, 2015. She was instrumental in the separation of GOOG’s main business -- advertising and search -- from all the “other bets” into which the company has invested billions.
Lo and behold, it has turned out that GOOG is indeed, as one analyst put it, “an advertising company with a lot of hobbies.” They have navigated significant industry changes very ably -- for example, the shift from desktop to mobile -- while maintaining growth in their core business. But still, the timeframe on real monetization for most of their “other bets” is long. Ms Porat has served the company well in bringing a dose of no-nonsense realism to temper the flights of futurism to which the company’s founders have been forgivably prone.
In January last year, we wrote an article for this newsletter outlining the case that “strong AI,” or artificial intelligence -- i.e., fully autonomous thinking computers -- could well remain in the realm of science fiction forever. AI may remain the ostensible goal, but the real benefits will come from augmented human intelligence, not artificial machine intelligence -- that is, from the assistance to human intelligence that machines can provide.
Terry Winograd, an academic mentor of GOOG’s founders, is one of the godfathers of “augmented intelligence.” His orientation is towards the “human-computer interface” -- to use computers for the tasks they are best at, such as brute-force exhaustive analysis of huge data sets, and to let humans do what they’re best at -- interpreting vague statements, intuitively sizing up a situation based on input that never becomes explicit, or immediately selecting relevant sensory phenomena without exhaustively analyzing all the inputs.
In short, skeptics are probably right that GOOG is never going to create “Hal” from 2001: A Space Odyssey. And there is always going to be work for humans to do -- but they will have more and more powerful tools to help them do it.
This is where GOOG’s AI ambitions come in. In 2014 they acquired a British company called DeepMind Technologies, which made headlines last year when one of its programs managed to beat the world champion of the ancient Asian strategy game called “go.” The was a step beyond the victory of IBM's (NYSE:IBM) Deep Blue computer over world chess champion Garry Kasparov in 1997, because go is a much more complex game than chess -- more complex because there are so many more possible games (millions of times as many, in fact). Even though the “AlphaGo” program learned to play from analyzing hundreds of thousands of human games, and playing millions of games against itself, the processes are still distinct and algorithmic, unlike human intelligence. What makes AlphaGo’s victory more important than the IBM chess victory is that while Deep Blue was built just to play chess, the learning system created for AlphaGo will have many applications in areas as diverse as helping physicians diagnose disease.
GOOG didn’t acquire DeepMind because its founders are unrealistic dreamers, in short. It acquired DeepMind first, to keep talent and prestige out of the hands of rivals -- particularly Facebook (NASDAQ:FB) , Microsoft (NASDAQ:MSFT) and Amazon (NASDAQ:AMZN) . Second, the acquisition keeps GOOG in the forefront not of a mythical push for superhuman sentient robots, but of peak power in real-life data crunching that can help human decision-makers be more effective. It has already helped GOOG tweak power usage in its data centers, once DeepMind got access to GOOG’s huge set of data on the company’s energy consumption habits. And this is the third, and maybe the biggest benefit of owning DeepMind -- every such acquisition is another inlet valve for the masses of data that the world’s great IT companies are so hungry for.
Investment implications: Despite popular press, GOOG’s ambition in buying DeepMind Technologies, the artificial intelligence company that beat the world go champion last year, was not to create a sentient and autonomous supercomputer. Rather, GOOG’s focus is on augmenting human intelligence rather than replacing it. However the company has shown its determination to aggressively collect the best talent from the world of artificial intelligence -- as well as the vast data troves that will eventually make data analytics services indispensable to businesses in almost every industry.
Market Summary
We remain bullish on U.S. stocks, but we are careful to point out that certain sectors will fare much better than the market as a whole. We favor those industries which have underperformed over the last few months or years, including medical technology, regional banks, oil production, and small and mid-sized companies with high tax rates and solid growth profiles which will benefit from lower U.S. corporate taxes.
We suggest avoiding bonds, interest sensitive areas such as utilities, and companies who have a large part of their sales to the U.S. government (President-elect Trump has made it clear that he wants U.S. taxpayers to pay low prices for whatever they buy). This will hurt the profit margins of companies who sell to the U.S. government.
Avoid retailers who do not have a strong internet presence (brick-and-mortar retail will suffer as online retail takes market share). Avoid restaurants as higher costs of food and labor will put pressure on profit margins.
In Europe, we are bullish on the German stock market and suggest that investors hedge their euro investments in U.S. dollars.
We are modestly bullish on gold, and much more bullish on gold mining shares. Gold miners have cut costs, closed unprofitable operations, and rationalized their operations to maximize profits. We believe that a 20–25% rise in the price of gold could create a gain of 40% or more in selected gold mining shares.