SITALWeek #184 3-17-19

SITALWeek #184

Stuff I thought about last week 3-17-19

Greetings – music streaming as a loss leader and how it fits into the big Internet platforms; further signs of a major reset at Facebook continue to widen the range of outcomes for the business; Amazon dives into the dark world of economics, and more in this week’s note. As always reply back with any thoughts.

Click here to sign up for SITALWeek

[Please read important disclaimers at the bottom of this post]

Stuff about Innovation and Technology

Music is always likely to be a loss leader as part of a broader bundle – currently it makes the most sense for music distribution to be part of Apple, Google, Amazon, Tencent, etc. However, in order to balance the power law based, network effects of the big Internet platforms, it’s possible music should belong to a different type of bundle, like transportation as a service with Lyft or video with Hulu (or both!). Here are lot more of my thoughts on music streaming for anyone interested.

Following my comments and concerns last week on Facebook’s pivot to secure, private messaging (note: I think this is ultimately good for users, but bad for FB’s revenues) the news this week of more key departures, particularly product lead Chris Cox, was remarkable not only because it indicated intense disagreement internally, but more so because investors completely shrugged it off. Zuckerberg began signaling a huge shift in Facebook’s business in January of 2018. Since then, attempts to fix the newsfeed and engagement (the biggest driver of FB revenues and profits globally) have failed, so Zuckerberg is doing the right thing for the product by overhauling the user experience. However, FB has no particular advantage in secure, private messaging, and there is no reason messaging needs to be a monetizable business. Short term, FB business is probably fine, or better than fine, but, long term, it may go through a significant reset, which has a wide range of outcomes not priced into in the current share price.

Video games are going through a major re-platforming from consoles to phones and streaming. Microsoft made some announcements last week and Google has a big event planned. Amazon is also really well positioned with their AWS and Twitch assets, and there are great independent platforms like Unity. It’s going to be a major race to get to a power law winner for gaming, which should be huge for gamers and game makers as it happens over the next 5-10 years.

Walmart has seen great outcomes from leveraging specialized centers of excellence for specific medical issues for employees: “At the heart of the program is an idea that health care incentives need to change. Rather than paying health providers for performing procedures and tests, Walmart rewards them for overall health outcomes.” And it's going well, so much so that Walmart wants to encourage other employers to follow its lead. Geisinger Medical Center said it wants to make the program available to smaller employers, including mid-sized companies.

With Amazon now employing 150 economists, 2nd only to the Fed in the US, the rise of economists at tech companies is a double-edged sword – they can be quite useful in the small number of cases where economic theory applies (like auction dynamics for search advertising, or housing forecasts at Zillow), but, for the most, part classical economic theory is typically useless. One of the problems with having a team of economists around is that you start to become confident you can make accurate predictions about the future. However, this is overwhelmingly not the case, as the economy is an inherently unpredictable complex adaptive system. If this topic interests you and you’ve got a little tolerance for mathematics, I recommend Ole Peters’ lecture notes on ergodicity.

In other Amazon news, the company pulled 2 books this week that had dangerous and inaccurate information in them that had the potential to do harm to people who took their advice. While, in this particular case, I agree with Amazon’s actions, it’s another echo of the power of Internet platforms and their role in controlling what people learn or don’t learn on various subjects. To draw a line in the ground and say these books can do damage and these books cannot is perhaps a little more difficult than it appears on the surface - most of the self help and diet book industry is filled with inaccuracies and potentially harmful advice. And speaking of economists, we might as well ban all classical economics books so that the politicians don’t get their hands on them! As Mark Twain said: “Be careful about reading health books. You may die of a misprint.”

The Boeing 737 Max tragic crash and grounding reiterates the increasing risks and complexity of technology integration – hardware, sensors, software and user interfaces – in every facet of business. Increasingly, every company is becoming a tech company as this trend continues. Aviation, like terrestrial transportation, should also see exciting innovation in propulsion and power systems over the next couple of decades, which will likely reduce risks and become increasingly necessary given the impacts of burning jet fuel at high elevations on the environment.

After NVIDIA’s $7B acquisition of Mellanox this week, to wrap up SITALWeek’s innovation and tech section, here is a brief analysis from Brinton Johns about some interesting developments for data center chips – we think investors are missing a lot of megatrends in semis, including this issue of speeding up computing in the cloud caused by the Intel bottleneck:

This week saw NVIDIA buy Mellanox and Facebook purchased Sonics. To understand possible conclusions, we start with a quick primer on system speedup. There are three primary ways to speed up a system: 1) faster processing 2) closer proximity between chips, especially memory and 3) faster interconnect fabric.

Faster processing: In the past, this speed up has been taken for granted due to the consistency of Moore's law. However, Intel is now four years behind. Over the past 3-5 years, we've seen a massive shift to parallel computing aided by graphics processors made by NVIDIA and programmable chips such as the ones made by Xilinx. This takes out the bottleneck posed by serial X86 processors.  We expect this to continue to shift to RISC based processors, notably RISC-V processors could make major headway over the coming five years as we elaborated here.

Closer proximity between chips: There has been a major evolution in chip packaging over the past decade.  The result is that chips (especially memory) sits closer than ever to the processor.  Taiwan Semiconductor, the largest outsourced maker of chips pioneered a process known as chip on wafer on substrate (COWOS) that has lead to outsized system speedups.

Faster Interconnect: Up to this point, interconnect has been sort of the red-headed stepchild of system speedup. Interconnect speedup can be accomplished through faster physical interconnects and faster communication protocols.  It seems that the acquisitions from NVIDIA and the much smaller acquisition from FB this week are focused on the communication protocol side.  It gets technical very quick in this space, but the high level is that NVIDIA is clearly pushing harder into Intel's fortress (the data center) and Facebook is following in the footsteps of Google and others by placing additional emphasis on developing their own internal silicon.  

What conclusions can we draw? In the post-Moore's law world, companies are seeking out new ways to speed up systems outside of faster Intel X86 chips. It's early days, but what seems clear to us through our recent work on RISC-V, the events we're seeing in interconnect, our long history with Taiwan Semiconductor, programmable chips, and graphics processors is that Intel's fortress is being taken apart brick by brick while this profit pool is increasingly becoming open to new entrants.  Another clear implication is the opening of additional streams of income to the software tools players Cadence and Synopsis.

Lastly on the topic of semis, here is a good history and overview of the RISC-V market from ZDNet.

Miscellaneous Stuff

Farewell to Johnny Thompson, one of the greatest magicians ever. He worked later in his career behinds the scenes to perfect some of the greatest tricks.

Your head is older than your feet – it’s true! Here is a great little piece on time worth reading.

Sharp rise in Arctic temps now unavoidable: “Scientists fear Arctic heating could trigger a climate “tipping point” as melting permafrost releases the powerful greenhouse gas methane into the atmosphere, which in turn could create a runaway warming effect.”

The college admission scam news out this week was neither surprising nor newsworthy, but in many cases was certainly hilarious (and sad of course, but also really funny). I do think it might be in part related to the culture of the fake, phony and surreal that social networking (especially Instagram of late) has brought us, and, in general, it’s a black eye on Americans that I’d hope would cause a tiny rethink of values for folks...but probably not – people seem quite committed to their Instagram-induced alternate version of reality.

The potential permanent loss of the gut microbiome might take humans with it.

Stuff about Geopolitics, Economics, and the Finance Industry

China will continue to support the VIE structure for Chinese Internet companies – this is somewhat arcane, but represents an important positive for US investors in Chinese companies, like Alibaba. However, the new rules might require 300,000 JVs with western companies to be renegotiated.

Power laws seem inevitable as we’ve written about before, but, in the asset management industry, especially quant- and algorithm-driven investing, power laws will only collapse in spectacular ways as we saw in the 2008 financial crisis, so this concentration of quant hedge fund assets is concerning.


Nothing in this newsletter should be construed as investment advice. Nothing contained in this newsletter is an offer to sell or solicit any investment services or securities. I may own long or short positions in stocks discussed in this newsletter. This newsletter is simply an informal gathering of topics I’ve recently read and thought about. It generally covers topics related to the digitization of the global economy, technology and innovation, macro and geopolitics, as well as scientific progress especially in the fields of cosmology and the brain. I will frequently state things in the newsletter that contradict my own views in order to be provocative. I may change my opinions without updating them in the newsletter. Lastly, often I try to make jokes, and they aren’t very funny sorry.