Categories
Articles Thoughts

The Apple Experience

At this point, people don’t need to upgrade their phones every two years. Phones are fast enough and the bump from the last generation A10 fusion chip to the latest A11 bionic really isn’t that important. Apple has even started added some fancy name to the end to uphold the experience of getting a new, more powerful phone. As a result, the deliberate slowdown was seen as user hostile to deceptively increase user delight when upgrading to a new phone and artificially enhancing the “this is so much smoother than my old phone” feeling. If the last iPhone started at 100% performance and degraded to 75%, the jump to 125% feels more significant.

From A Message to Our Customers about iPhone Batteries and Performance

It should go without saying that we think sudden, unexpected shutdowns are unacceptable. We don’t want any of our users to lose a call, miss taking a picture or have any other part of their iPhone experience interrupted if we can avoid it.

Apple mentions there are three contributions to battery life and performance:

  • a normal, temporary performance impact when upgrading the operating system as iPhone installs new software and updates apps
  • minor bugs in the initial release which have since been fixed
  • continued chemical aging of the batteries in older iPhone 6 and iPhone 6s devices, many of which are still running on their original batteries.

As always, our team is working on ways to make the user experience even better, including improving how we manage performance and avoid unexpected shutdowns as batteries age.

As they should. Apple has always been the experience company. The Apple walled garden is carefully designed in the ethos that people don’t know what they want until you show it to them. Maybe we need a little more clarity into how Apple creates people’s preferences.

Categories
Articles Thoughts

Cosmic Brain

Remember this video? In it you start with a person lying on the ground looking up at the sky. The camera zooms out exponentially into space showing the immense scale of the galaxies. Then, we quickly zoom back in to molecular scale

Notice the similarity of the Cosmic Web at 1 billion light years and Quarks at 1 femtometer

I immediately recalled the Cosmic eye video when I read this headline:

The Strange Similarity of Neuron and Galaxy Networks

The article is written by an astrophysicist and a neuroscientist on the similar complexities and structures of the brain and the cosmic web.

The task of comparing brains and clusters of galaxies is a difficult one. For one thing it requires dealing with data obtained in drastically different ways: telescopes and numerical simulations on the one hand, electron microscopy, immunohistochemistry, and functional magnetic resonance on the other.

It also requires us to consider enormously different scales: The entirety of the cosmic web—the large-scale structure traced out by all of the universe’s galaxies—extends over at least a few tens of billions of light-years. This is 27 orders of magnitude larger than the human brain. Plus, one of these galaxies is home to billions of actual brains. If the cosmic web is at least as complex as any of its constituent parts, we might naively conclude that it must be at least as complex as the brain.

Complexity of the brain and cosmos “can be quantified by counting how many bits of information are necessary for building the smallest possible computer program that can … predict its behavior.” We do this using an equally fascinating measurement tool: Computers. “independent studies have concluded that the total memory capacity of the adult human brain should be around 2.5 petabytes, not far from the 1-10 petabyte range estimated for the cosmic web!” Petabytes are huge.  This means the amount of information needed to predict the behavior of a human brain is roughly equivalent to the amount of data required to stream the first two seasons of Stranger Things 50 thousand times.

Does this fact tell us something profound about the physics of emergent phenomena in the two systems? Maybe. But we must take these findings with a grain of salt. Our analysis has been limited to small samples taken with very different measurement techniques.

But it’s fun to think about.

Categories
Articles Technology

Catching up on Stratechery

My Instapaper reading list was piling up. Nearly half of the articles were from Stratechery, so I decided to knock them all out at once (well, over the course of a day or two).

https://stratechery.com/2017/goodbye-gatekeepers/

From Weinstein and movies to the NYTimes and YouTube

In a world where the default news source is the Facebook News Feed, the New York Times is breaking out of the inevitable modularization and commodification entailed in supplying the “news” to the feed. That, in turn, requires building a direct relationship with customers: they are the ones in charge, not the gatekeepers of old — even they must now go direct.

YouTube produces an astounding amount of fame.

YouTube represents something else that is just as important: the complete lack of gatekeepers. Google CEO Sundar Pichai said on an earnings’ call earlier this year that “Every single day, over 1,000 creators reached the milestone of having 1,000 channel subscribers.” That is an astounding number in its own right; what is even more remarkable is that while Hollywood has only ~3,500 acting slots a year (including all movies, not just major studios), YouTube creates 100 times as many “stars” over the same time period.

https://stratechery.com/2017/tech-goes-to-washington/

Did he say 330 million?

https://stratechery.com/2017/why-facebook-shouldnt-be-allowed-to-buy-tbh/

Requiring Facebook to offer its social graph to any would-be competitor as a condition of acquiring tbh would be a good outcome; unfortunately, it is perhaps the most unlikely, given the FTC’s commitment to unfettered privacy (without a consideration of the impact on competition).

https://stratechery.com/2017/stitch-fix-and-the-senate/

Negative churn

existing customers were increasing spend by more than the revenue lost by those leaving

https://stratechery.com/2017/pro-neutrality-anti-title-ii/

The most famous example of an ISP acting badly was a company called Madison River Communication which, in 2005, blocked ports used for Voice over Internet Protocol (VoIP) services, presumably to prop up their own alternative; it remains the canonical violation of net neutrality. It was also a short-lived one: Vonage quickly complained to the FCC, which quickly obtained a consent decree that included a nominal fine and guarantee from Madison River Communications that they would not block such services again. They did not, and no other ISP has tried to do the same; the reasoning is straightforward: foreclosing a service that competes with an ISP’s own service is a clear antitrust violation. In other words, there are already regulations in place to deal with this behavior, and the limited evidence we have suggests it works.

https://stratechery.com/2017/free-daily-update-light-touch-cable-and-dsl-the-broadband-tradeoff-the-importance-of-antitrust/

The equation is straightforward: there is wide consensus amongst economists of all political stripes that regulation imposes costs on both innovation and society through regulatory capture; I would prefer to avoid bearing that cost until we are certain it is necessary, particularly since the evidence to date suggests after-the-fact regulation is working.

The question that must be grappled with, though, is whether or not the Internet is “done.” By that I mean that today’s bandwidth is all we all never need, which means we can risk chilling investment through prophylactic regulation and the elimination of price signals that may spur infrastructure build-out (that being the elimination of paid prioritization).

If we are “done”, then the potential harm of a Title II reclassification is much lower; sure, ISPs will have to do more paperwork, but honestly, they’re just a bunch of mean monopolists anyways, right? Best to get laws in place to preserve what we have.

But what if we aren’t done? What if virtual reality with dual 8k displays actually becomes something meaningful? What if those imagined remote medicine applications are actually developed? What if the Internet of Things moves beyond this messy experimentation phase and into real-time value generation, not just in the home but in all kinds of unimagined commercial applications? I certainly hope we will have the bandwidth to support all of that!

The problem with regulating broadband in this way, though, is that the definition of acceptable broadband is much more of a moving target. As Marc Andreessen memorably put it on Twitter:
@mattyglesias @binarybits Because sewers and electricity are far more static markets than broadband. You don’t shit 10x as much every 3 yrs.
— Marc Andreessen (@pmarca) February 23, 2014

https://stratechery.com/2017/the-pollyannish-assumption/

Documenting why and how these platforms have power has, in many respects, been the ultimate theme of Stratechery over the last four-and-a-half year: this is a call to exercise it, in part, and a request to not, in another. There is a line: what is broadly deemed unacceptable, and what is still under dispute; the responsibility of these new powers that be is to actively search out the former, and keep their hands — and algorithms and policies — off the latter. Said French Revolution offers hints at fates if this all goes wrong.

https://stratechery.com/2017/disney-and-fox/

This is a remarkable look at how Disney could leverage 21st Century Fox to compete against Netflix in the years ahead. One of the most insightful articles with a clear line of how we could get to a future where Netflix and Disney are massive content aggregators.

The best sort of acquisitions, though, are best described by the famous Wayne Gretzky admonition, “Skate to where the puck is going, not where it has been”; these are acquisitions that don’t necessarily make perfect sense in the present but place the acquirer in a far better position going forward: think Google and YouTube, Facebook and Instagram, or Disney’s own acquisition of Capital Cities (which included ESPN).

 

The problem now is obvious: Netflix wasn’t simply a customer for Disney’s content, the company was also a competitor for Disney’s far more important and lucrative customer — cable TV. And, over the next five years, as more and more cable TV customers either cut the cord or, more critically, never got cable in the first place, happy to let Netflix fulfill their TV needs, Disney was facing declines in a business it assumed would grow forever.

 

… differentiated content is Disney’s core competency, as demonstrated by its ability to extract profits from cable companies.

 

   Consider the comparison in terms of BATNA (Best Alternative to a Negotiated Agreement): for distributors the alternative to not carrying ESPN was losing a huge number of customers who cared about seeing live sports; that’s not much of an alternative! Netflix, on the other hand, can — and is! — going straight to creators for content that viewers can watch instead of whatever Disney may choose to withhold if Netflix’s price is unsatisfactory.
Clearly it’s working: Netflix isn’t simply adding customers, it is raising prices at the same time, the surest sign of market power.
Therefore, the only way for Disney to avoid commoditization is to itself go vertical and connect directly with customers

Will it go through?

If one starts with a static view of the world as it is at the end of 2017, then there may be some minor antitrust concerns, but probably nothing that would stop the deal. Disney might have to divest a cable channel or two (the company’s power over distributors would be even stronger; basically the opposite of the some of the concerns that halted the Comcast acquisition of Time Warner), and potentially be limited in its ability to make operational decisions about Hulu (Disney would have a controlling stake after the merger; Comcast was similarly restricted after acquiring NBC Universal, but there the concern was more about Comcast’s conflict of interest with regards to its cable TV business competing with Hulu). The Hulu point is interesting in its own right: Disney could choose to focus its streaming efforts there instead of building its own service, but I suspect it would rather own it all.

That’s it for now. Keep reading. Keep connecting.

Categories
Technology Thoughts

Netflix and Coinbase Engineering Cultures

I watched a few QCon videos on the InfoQ YouTube channel. The presentations are from conferences over the last year or so, but the videos were all uploaded within the last month.

A blueprint for organizational success- case study Netflix

I’m pretty sure this is a re-upload, but it caught my eye again. (Here’s another great session by Josh Evans Mastering Chaos – A Netflix Guide to Microservices) Josh Evans begins the talk with with focus on how the organizational structure at Netflix led to internal struggle and dictated the engineering process. Tribalism and the expression of Conway’s law meant the way Netflix shipped code mirrored the team hierarchy. It wasn’t until upper management got involved that the teams sorted out their differences and began to put the architecture before the org structure.

These quotes stood out to me:

Organizational Scalability: The ability for an organization to easily add people and domain responsibilities in response to increased work and complexity. The ease with which an organization or team can adapt to shifts in business strategy

For an organization to grow, the culture must be able to adapt to changes and fluctuations in daily tasks. Netflix grew from prioritizing DVD through the mail to online streaming, two starkly different business models. Looking back at the change, it is easy to place explainations on how things went, but what seemed to be constant is the engineering culture. As Evans mentions:

We have a culture of creativity and self discipline, freedom and responsibility.

Once defined, culture is not easily changed, but setting the right culture from the beginning is crucial to and validated by success.

If you get a chance, be sure to watch (or at least listen) to the video.

Crushing Tech Debt Through Automation at Coinbase

This session with Rob Witoff at Coinbase from March 6, 2017 details how the startup is growing its technology with the interest in cryptocurrencies. The nine month distance in timing gives more perspective in light of recent events. Things are constantly evolving with Bitcoin, but just look at the comments relevant to the week the video was published on November 30, 2017. While cryptocurrency is its own fascinating discussion, the engineering culture at Coinbase is of paramount importance to the company and worth investigating.

Engineering Velocity requires tools and guardrails to empower engineers to work without fear.

And what is the Coinbase engineering velocity? Devs deploy an average of 4 times per week and 16 times a month. This rate of code movement requires heavy investment in testing and tools to ensure changes are good. Not every deploy succeeds (by design) and each failure is an opportunity to improve the product. Bugs in successful deployments is an opportunity to improve the deployment pipeline, and catch similar errors in the future.

Coinbase avoids a culture of blame to ensure people have the freedom to learn and grow. The engineering systems support this ideal and allow the company to scale. People like to compare market caps of Bitcoin and publicly traded companies, and we’ll have to see if the culture at Coinbase allows it to scale similar to Netflix

Again, I won’t spoil it all, as it’s worth the watch.

Post note

The InfoQ youtube channel used to be NewCircle. NewCircle specialized in tech training videos, and InfoQ has a similar focus. I originally subscribed to the NewCircle channel, and when the channel switched over to InfoQ I was happily surprised by the new QCon session videos.