The Geek’s Reading List – Week of April 21 2017
Welcome to the Geek’s Reading List. These articles and the commentary are not intended to be taken as investment advice, nor should they today. That being said, investors need to understand crucial trends and developments in the industries in which they invest. Therefore, I believe these comments may actually help investors with a longer time horizon. Not to mention they might come in handy for consumers, CEOs, IT managers … or just about anybody, come to think of it. Technology isn’t just a niche area of interest to geeks these days: it impacts almost every part of our economy. I guess, in a way, we are all geeks now.
Please feel free to pass this newsletter on. Of course, if you find any articles you think should be included please send them on to me. Or feel free to email me to discuss any of these topics in more depth: the sentence or two I write before each topic is usually only a fraction of my highly opinionated views on the subject!
This edition of the Geeks List, and all back issues, can be found at www.thegeeksreadinglist.com.
1) VMware joins Cisco, HPE and Verizon in shifting course on public cloud
Cloud computing consists of two components: commodity computing and a whole bunch of additional services. The largest players (Amazon, Microsoft, and Google) developed the data centers and additional services in the course of their normal business and that gave them a tremendous leg up on lesser players. Frankly I think it would be hard for the like of HP, let alone a telephone company, to make a case to use their cloud services. At its core that is why they are “shifting course”.
“Successful cloud companies don’t need to be public cloud companies. Confused? Well, if you think the public cloud is the be-all, end-all of cloud computing, that’s understandable. Story after story focuses on Amazon Web Services (AWS), Azure, Google Compute and IBM Cloud. But, in the background several major companies are abandoning their public clouds. First, HPE dropped its Helion public cloud in 2015. Then, Verizon dropped its public cloud offering in 2016. Later that year, Cisco got out of the public cloud business. Now, VMware has sold its Infrastructure-as-a-Service (IaaS) public cloud, vCloud Air, to OVH, a European cloud provider. What’s going on here? Are these companies fleeing the AWS Goliath? That may be part of it, but certainly not the whole story.”
2) IBM Sales Miss Shows Return to Growth Not Without Roadblocks
IBM is a company which specializes in financial rather than actual engineering. As this article notes, revenue has been falling for 5 years despite the company spending like drunken sailors on acquisitions. (It seems they’d rather enrich the shareholders of other companies rather than their own). They appear to have bet the farm on things like Watson at a time when AI is becoming a useful commodity which cloud service providers will offer at nominal cost. In the press release the company noted “our cognitive solutions again grew strongly”, because, apparently at IBM an increase of 2.1% in the segment you bet the farm on is “strongly”. Meanwhile the stock is actually up year/year even though it dropped (only) a few percent on this news.
“IBM’s quarterly sales fell short of analysts’ estimates for the first time in a year on struggles in some overseas markets and delays in service contracts, reminding investors that obstacles remain on the path back to growth. The results announced Tuesday marked a 20th consecutive quarterly revenue decline. IBM also reported gross margin that contracted from last year, sending shares down as much as 5.5 percent in extended trading. The stock had gained 11 percent in the past year.”
3) Silicon Valley’s $400 Juicer May Be Feeling the Squeeze
We covered Juicero as an example of a pathological environment for venture capital in the tech space. In summary this is an Internet connected juice machine which, long story short, squeezed an expensive bag of DRM protected mush. It turns out that people realized they could use their hands to squeeze a bag of mush as well, bringing into question the $400 purchase price of the machine. No worries: in a stroke of brilliance the company issued this http://www.theverge.com/2017/4/20/15375940/juicero-full-refund-customers-ceo-jeff-dunn. Frankly all this sage shows is that you should ignore who invested in a company if you think it is a stupid idea. Stupid ideas don’t become good ideas because Google put money into then. Thanks to Marcel Valentin for this item.
“The experiment found that squeezing the bag yields nearly the same amount of juice just as quickly—and in some cases, faster—than using the device. Juicero declined to comment. A person close to the company said Juicero is aware the packs can be squeezed by hand but that most people would prefer to use the machine because the process is more consistent and less messy. The device also reads a QR code printed on the back of each produce pack and checks the source against an online database to ensure the contents haven’t expired or been recalled, the person said. The expiration date is also printed on the pack. … Kleiner Perkins Caufield & Byers joined Alphabet Inc. and others in funding Juicero. Evans’s subscription model had hit on a sweet spot for venture capitalists, said Brian Frank, who invests in food-tech companies through his FTW Ventures fund.”
4) Toyota is testing heavy-duty hydrogen trucks at the Port of Long Beach
Fuel cells have their uses and the Port of Long Beach (near LA) and the Port of Los Angeles (right next it) may be one of them. These trucks are used in drayage (short haul service) and diesel trucks are a pollution nightmare when used for drayage applications. Having a massive concentration of them in a small area (Long Beach/LA) is a big problem. Plus, there is ample hydrogen production in the area so the usual challenge of fuel supply is not an issue in those particular locations. This does not make fuel cells a viable solution for long haul (though more viable than an electric truck).
“Two years ago, Toyota began secretly testing a hydrogen fuel cell system alternative to the conventional diesel powertrain for heavy Class-8 trucks. Called “Project Portal,” this system is intended for drayage (short-haul movements), shuttling shipping containers between Los Angeles and Long Beach ports plus other freight depots. Toyota is the first major car company to dip a toe in the fuel cell trucking waters, and it could eventually market the powertrains to various truck manufacturers nationwide or through its own Hino truck division. (Toyota used a Kenworth to demonstrate the powertrain; however, Hino does not make a Class-8 rig.)”
5) Parents In Germany Face $26,500 Fine If They Don’t Destroy Controversial ‘My Friend Cayla’ Dolls
Actually it is not clear if the fine would actually every be levied. The idea is simply that the doll is essentially a listening device which sends data in violation of German privacy laws. That should be enough to destroy the things.
“After researchers found that My Friend Cayla dolls were recording users’ and sending this information out to a third party specializing in voice-recognition for police and military forces, officials in Germany told parents to get rid of the toys. In case families didn’t take that request seriously, the country’s telecommunications regulator has since clarified that parents who don’t destroy their Cayla dolls could face more than $25,000 in fines.”
6) Edible CRISPR Could Replace Antibiotics
I continue to consider the CRISPR gene editing system to one of the greatest discoveries in history. This approach applies CRISPR to a sort of antibiotic application: use bacteriophages (viruses which attack specific bacteria) to deliver a CRISPR self-destruct mechanism. Since the bacteria already have CRISPR as part of their “immune system” it is unlikely they will develop a resistance – assuming the approach works as expected. The good news is, with CRISPR we’ll know soon enough.
“Van Pijkeren’s idea is to use bacteriophage to send a false message to C. difficile, one that instead causes the bacteria to make lethal cuts to its own DNA. To do it, Van Pijkeren lab is developing bacteriophage capable of carrying a customized CRISPR message. On their own, the bacteriophage would quickly get broken down by stomach acid. So to get the viruses into a person, Van Pijkeren plans to add them to a cocktail of innocuous bacteria, or probiotic, that a person could swallow as a pill or a liquid.”
7) Does Google’s TPU Investment Make Sense Going Forward?
We covered Google’s TPU announcement last week. If you recall, the device showed substantially better performance than contemporary GPUs and CPUs, which is very bad news for those who believe the future of AI is GPUs, and, in particular, NVidia’s stock price. This article seems to be an effort at damage control but it misses the point: inference is done thousands or millions of times more often than AI training. Most likely, the TPU chip itself cost less than $100, and the TPU board less than $500, or a fraction the cost of a GPU. Combine substantially lower costs, lower power consumption, and the fact it is proprietary and you have a winner.
“The question one has to ask is what did Google pay to develop the TPU and then have it manufactured. Yes, the TPU allowed it to save massive amounts of money not doing inference on CPUs, which are clearly not as good at it as specialized processors are. But so what? If a P4 accelerator card costs $2,100 and if the P40 accelerator costs $4,700, which we think are the approximate street prices for these devices, then Google has to be able to make its own chip cost no more than this on a cost per watt per unit of performance basis for the TPU to make economic sense – and it has to cost less than this, presumably. If Nvidia can double the performance of machine learning inference with the future “Volta” GPUs that will presumably be announced at the GPU Technical Conference in May and possibly shipping later this year for selected HPC customers and maybe for AI customers, then Nvidia V4 and V40 accelerators will be in the same league as a TPU gussied up with GDDR5 memory and moved to a slightly more aggressive process shrink to 20 nanometers.”
8) Facebook Announces “Typing-by-Brain” Project
Brain-computer interfaces appear to have galvanized the attention of billionaires recently. Musk and Zuckerberg appear to have decided this is the next big thing. Unfortunately, actual specialists in the field seem to think what they want to do is, well, not likely, but who is gonna argue with a billionaire – they know everything, right? Of course another issue is, assuming somebody could, actually, make something which could read brainwaves, what healthy person would let Facebook see their thoughts?
“At its developer conference, Facebook executive Regina Dugan promised that this brain-computer interface will decode signals from the brain’s speech center at the remarkable rate of 100 words per minute. … The promise of 100 words per minute represents quite a leap from the current speed record. In February, Stanford researchers enabled a paralyzed patient to type 8 words per minute—and that was using a device implanted in his brain. In that experiment the implant was placed in the patient’s motor cortex, and he imagined moving a cursor over a screen to select letters.”
9) A Chinese warehouse reportedly cut its labor costs in half with a fleet of tiny robots
The sounds pretty frightening if you are a warehouse worker, but it is a lot less scary if you look into it. Essentially in this case the robots are taking the place of a second and third tier sorting conveyor belt (see a simple example https://www.youtube.com/watch?v=5mMignhvXpc) . Presumably the labor costs savings are relative to doing it all by hand. Nonetheless the video is awesome.
“Shentong Express, a Chinese shipping company, showed off a mildly-dystopian automated warehouse last week that reportedly cut its labor costs by half, according to the South China Morning Post. In a video, tiny orange robots made by Hikvision ferry packages around an eastern China warehouse, taking each parcel from a human worker, driving under a scanner, and then dumping the package down a specific chute for it to be shipped.”
10) Uber lost $2.8 billion last year
I don’t get the part about reporting gross revenue since the company only gets their cut, or the net revenue. Even so Uber’s investors are subsidizing the service to the tune of $2B per year, meaning that Uber’s growth may continue only as long as it pays customers to use the service. More likely, the services itself else a tiny return on investment while the rest of the money is blown on stupid projects such as self-driving cars which meant to drive up the value of the shares. The company is a car service and car services are not inherently profitable. Nevertheless access to an unending supply of capital means the company is not run for profit’s sake.
“Uber’s gross bookings for 2016 hit $20 billion, more than doubling from the year prior, according to financial figures the company provided to Bloomberg. Its net revenue, after drivers took their cut, totaled $6.5 billion for the year. But that rapid growth came at a cost. Uber says it lost $2.8 billion in 2016, excluding the China business it sold midway through the year. Uber’s CEO had previously said it was losing $1 billion a year in China, prior to selling its China business to rival Didi Chuxing in August. … “We’re fortunate to have a healthy and growing business, giving us the room to make the changes we know are needed on management and accountability, our culture and organization, and our relationship with drivers,” Rachel Holt, Uber’s regional general manager for U.S. and Canada, said in a statement provided to CNNTech.”