Saxo Bank CIO and Chief economist Steen Jakobsen made a presentation yesterday following the Fed rate hike announcement. The macro presentation Goldilocks or Frankenstein was on the Technology Winter, Monopolies and FED policy mistakes.
Via email, Jakobsen writes …
Dear All,
This is presentation I did yesterday at an event. The FOMC meeting to me indicates peak in expectations has been reached, confirming our long held view that the global credit impulse (which explains 60% of next 9-12 month activity) is kicking in negatively.
Monopoly operators like FANGS will be under pressure:
I did interview with Bloomberg yesterday which partly explains our reasoning: Zero Appetite for Tech Stocks Right Now.
We think there is risk of technology winter similar to what we saw in the 1970s and 1980s – the perception vs. reality of what technology can do is widest in history – there is an extreme naïve believe in the “benign” nature of it. That doesn’t mean technology will not do good or be productive but the expectations and what can be done and when is out of sync with reality. Furthermore the recent Facebook event also tells story of how these monopolies are out of control – intentionally or not.
Peter Garnry phrased it excellently in internal email to our team:
“….But even before this event self-driving car technology was over-hyped. Recent consensus at NIPS (the biggest AI conference) was that deep learning etc. where only deemed good at narrow applications and Hinton, the father of backpropagation igniting deep neural networks, has even hinted that self-driving cars are not around the corner.
Based on yesterday’s event I’m revising up the probability that we will see a new AI winter like the one in the 1970s and 1980s.”
Finally going into a FOMC meeting where markets positioning and rhetoric wish’ for four hikes instead of three – I think the biggest value – except agriculture is in long US rates – ie. long fixed income……
Enjoy,
Med venlig hilsen | Best regards
Steen Jakobsen | Chief Investment Officer
Goldilocks or Frankenstein?
I disagree with opinions that self-driving cars are not around the corner. But we need to define “around the corner”. For the markets, 2022 is almost an eternity away.
Facebook problems are right here right now.
Look at the trailing 12-month Price-to-Earnings ratio of some technology companies: Amazon (P/E 250), Netflix (P/E 245) , Google (P/E 58), Facebook (P/E 31).
Trade Wars
Today, Trump announced $50 billion in Tariffs. He says China is “Out of Control”.
Excuse me for asking, but who is it that’s “out of control”?
Policy Mistakes
Policy mistakes are visible everywhere. Yet few will see them until we have a huge decline in the stock market. It’s coming.
Mind the Rear View Mirror
Nearly everyone is on the lookout for inflation.
Not me. I strongly suggest Inflation is in the Rear View Mirror.
Mike “Mish” Shedlock
I have to smile when Trump claims the US has been economically raped when since WW11 it is the US that raped the world. Churchill said of breton woods that it was the biggest defeat the British empire had ever suffered outside war. It is difficult to see what is worse, the rise in interest rates or the rise in stupid.
Realist, I don’t think the reference to a Tech Winter implies that tech won’t continue to be invented, or that it won’t continue to progress. I think that, rather, it’s a prediction that investing in tech will go somewhat out of fashion. Many people have made a fortune in the last twenty years by investing in stocks like Amazon, where prices are not in any rational way tied to actual earnings, but rather to the hope that in the future they can earn enough to justify it. If people stop projecting unlimited profitability expansion, that alone will lead to a “winter” for technology stocks.
The mention of deep learning only being applicable for narrow domains, is very much true. And very much lost on the hypesters. Constrict the environment sufficiently, and AI can work well. Siri can answer a call, and dial a series of digits spoken in a language she has been trained to understand. But ask her to learn Swahili and get back to you, and she’ll be out of battery, and likely replaced by someone younger and hotter, before even understanding the question.
Ditto for self-driving cars. Make the environment narrow, and the goals/rating of various actions sufficiently unambiguous; and the bots can do well. But leave it as open ended as the totality of all things cars are used for, and you’ll be disappointed. Those throwing money at them indiscriminately, should at a minimum be aware that the military has a huge incentive to replace soldiers and APCs with non-human versions. Yet still rely on fragile meatsacks to navigate urban, and even wide open, battlefields. Simple because the bots are too easily tripped up, in all but the most narrowly confined environments. The bots can technically both see and engage the enemy earlier/faster and more precisely than human soldiers. They just cannot reliably make kill/don’t kill; engage or retreat etc. decisions. Even in an environment as clear cut and simplified as on an open battlefield. Pretending they will somehow be able to do so efficiently in an opportunistic-human filled Manhattan, is just farcical.
So, for all the hype; while the technology and algos slowly improve; the biggest gains in AV space will come from reducing the complexity of the environment in which they are tasked with operating. Otto’s initial focus on trucks driving on limited access, lightly trafficked freeways, is a good one. Self-parking, hence allowing for densepacking urban lots, or even just allowing for lots too big to comfortably walk from one end to another, are other low hanging fruits.
Running gun battles through Downtown Manhattan rush hour, is decidedly not. Yet while the latter is (so far) rare, it’s one of those situations in which you really, really need your car to be at least not completely hopeless, for those rare occasions when you do find yourself in one. Ask anyone in an express-kidnap heavy Latin city whether he would want to send his kids to school in a robocar with no human override through a rough neighborhood where every thug knew the robot’s only response to being crowded was to cease and freeze…. A human driver would be able to distinguish a thug band from random pedestrians. And may just run them over, with or without opening fire on them in the process. Try coding, or backpropagating that one into the logic of a robot, and get it past the guys in Legal…..
Nope, no robot doing our dishes, but I’ve got an iPhone stuffed with politically correct emojis.
I’m glad I’m not in the stock market right now. Stock are a fool’s paradise.
link to sovereignman.com
FTSE is a leading indicator and “sell in May and go away” has been brought forward. Return St Ledger Day will still stand but perhaps just to sell some more.
Well. After a BOE interest hike, Trump’s trade war announcement has finally wiped out all of the FTSE gains in 2017. Down -10% YTD.
https://uk.news.yahoo.com/banks-send-britains-ftse-15-month-low-reckitt-102413080–sector.html
In other words, those unicorns are actually just donkeys with strap-on plastic horns.
How about Siri, Alexis, etc? Incredible AI achievement in widespread use.
Agree completely. I haven’t seen a real product created by AI yet. In fact, I haven’t seen anything of value, tech-wise, created in the last decade. Anybody buying 3d printers? How’s the Internet of Things doing? Anybody have a robot doing their dishes? I’m a firm believer that the computer/internet revolution is firmly in the rear view mirror. And there are no more revolutions coming.