Humor: Modern warfare

Modern warfare – another good one from the New Yorker.

The Internet of Eyes

Super interesting article in the New Yorker about the advances of facial recognition and its potential abuses.  Some points that caught my eye:

• There’s a start up (Cainthus) that apparently has cracked the code on providing facial recognition for cows. Yes cows. Hence the cow photo.
• I.B.M. has a system which is able to automatically generate highlight reels from raw sports footage by analyzing things like crowd cheers, announcer excitement, fist pumps, high fives, and the location of the three-point-shot line.  Wow.
• China has nearly two hundred million public-surveillance cameras, far more than any other country.

Definitely worth a read. Scary.

China leads in top papers for math and computing

Hardly a day goes by when’s there not an article or two about the increasing dominance of China in technology. It’s often instructive to think about the root causes for that. One of those has to be the increasing quality of China’s researchers. The Economist has a nicely written article on the rise of the top Chinese universities – particularly Tsinghua. A few points that caught my eye:

• In 2006-2009 Tsinghua was 66th in the ranking table for maths and computing research. Now it’s at the top. That’s a fast rise. 
• Chinese universities give our bonuses to their researchers for getting a paper published in a top journal equivalent to about 20x their annual salary. Wow – that’s in incentive. 
• Tsinghua and Peking universities are neighbors modeled about Oxford and Cambridge. 

Worth a read. Emphasizes the need for long range planning and investments. 

Internet’s next 20 years

Great presentation on the future of the internet by Benedict Evans of a16z. It’s a quick ~20′ but it covers a wide swath of the internet with simplicity, rigor and a lot of data.

His essential points seem to be that we are really at the beginning of a new phase of the internet with new building blocks (crypto, ML) rather than the end.  Well worth the watch. (Unfortunately, I can’t find a pdf of the presentation). I’ve also heard him on another good podcast on innovation.

The end of “seeing is believing”?

Great article in the New Yorker on the rise of advances in “synthetic video” and what that could imply for humanity. While Photoshop and similar editing tools have been around for a while, one has still needed a skilled person to produce a “fake”. But with the advance of ML and the easy availability of digital data that’s no longer true.

A couple of points that caught my eye:

• Modern digital cameras, meanwhile, often achieve higher resolutions by guessing about [~66% of] the light their sensors don’t catch
• Companies are using Blockchain to fight back – e.g. Truepic, a start up, uses Blockchain + attributes about the image / camera to verify whether a photo is original.

You can read the paper here; or see an example of an actor “impersonating” Obama here.

Great article — scary implications for our future.

Alibaba slows down US expansion

According to a new article from the Information, Alibaba, has slammed on the brakes for its US expansion plans. It was pretty recently (2015) that the CEO of Alibaba (Daniel Zhang) had outlined some ambitious plans:

The cloud business will be a very important sector for Alibaba. We hope to match or even surpass Amazon in 3 to 4 years.

Apparently they have run into headwinds and now plan to focus on “the narrower market of U.S. multinational companies that need cloud services in China”. Among the reasons are AWS’ lobbying power in DC given the trade climate and its inability to find the right sales personnel. It’s data centers in the US are running below capacity. But, it’s not all bleak: Alibaba Cloud overtook IBM to become the world’s fourth-largest public cloud provider during the first quarter of 2018.

Fastest unicorn ever

Nice article in the Economist about the rise of electric scooters + bikes. The part that caught my eye was economics. E.g. Bird (one of the startups) has a valuation of $2B and has now become the fastest startup to become a unicorn.

The fact that each scooter costs about $300 to $400, and with each earning about $20/day, it takes just 15-20 days to pay one off. In the US if 2M get deployed (across all providers, not just Bird), we are looking at earnings of about $15B a year. Wow.

(By the way: If you haven’t tried one, the electric scooters are fun to ride and a cinch to activate and drop off)

Unscaling

 

It has been “common” knowledge that in order to drive more business, make it harder for competitors to compete, for companies to drive down prices etc. they have to scale. A new, pretty well written book argues that those days may be over. It makes a pretty compelling argument that with all the “for rent” infrastructure out there (computation, manufacturing, potential consumers etc.) plus the advances in AI companies no longer need to build scale, but can rather rent it.

It’s not a academic book. The author is a VC and has personally invested in companies that illustrate this. Some examples / tidbits

• Voodoo Manufacturing wants to be the AWS of cloud manufacturing. It has a factory full of 3D printing machines available for “rent”
• Forward thinking utilities will evolve to become platforms essentially operating an energy version of the internet
• Mobility as a Service (MaaS) that allows one to pay for all types of journeys in one go (e.g. Whim)

And so on. Worth a quick read.

 

 

 

Apps market framing

When talking about ‘enterprise applications’ most of the discussion is in often what I would have called the traditional areas – HR, CRM, ERP, collaboration. And most of the discussion is really the hand wringing on how the SaaS players are (mostly) eating the legacy’s players’ lunch (e.g. SFDC vs. Oracle/SAP); but that doesn’t paint the whole picture. It was that mindset I found the following framing from UBS (Software 2.0, Nov 2017) valuable.

It makes the (cogent) argument there’s more than one way to play in this market. One is to go towards more simpler apps targeted at individuals / small groups that are less costly and simpler while the other is towards more strategic apps (think predictive maintenance) that are more expensive and complex.

Cloud perceptions favor Azure

Two charts from a recent report from UBS (Microsoft, Nov 2017) caught my eye. They point to the fact that Microsoft has a very positive perception from enterprise buyers on their cloud vision but lags on their understanding of the client’s business.

IBM leads the pack on this second dimension, presumably because of their services + consulting group.