Pages in this blog

Sunday, December 23, 2018

Dan Wang on Moore's Law [beyond stagnation], with an addendum on entertainment

From Dan Wang, What I Learned in 2018:
Moore’s Law turned from a neat backwards-looking observation into an obligation for the entire chips industry to keep improving. One description that I like of it is that it’s a “clock that has become a chaotic attractor for innovation.” I don’t think there are many other technologies in which exponential growth in performance over decades is possible. But maybe there’s a handful more that are, and they await a nice label that will concentrate minds, mobilize capital, and attract talent to keep improving. Coming up with that label might be a kind of low-hanging fruit that would encourage greater growth.

I’d like this exponential progress to come to other fields, especially industrial technologies. Semiconductors are upstream of all electronics, which is a sector that has been vibrantly innovative over the last few decades. If we had exponential progress in a few more upstream technologies, we may be able to enjoy faster innovation in fields beyond computers, software, and the internet. Silicon Valley is rightly celebrated as a driver of innovation and wealth creation. But I’m not sure to what extent that Silicon Valley companies have yet promoted dynamism in the broader non-tech world. Companies there are very good at building software on top of and abstracted from the physical world. The tech companies we hear most often about tend to be capital-light, beautifully-scalable businesses that earn the most handsome returns for investors.

We’re excited about companies like Airbnb and Uber, which match consumers with underutilized assets. Better matching of supply and demand is valuable, but I’m looking for something more ambitious. Focusing on industrial technologies is more like taking a firm hold of the supply curve and pulling it downwards; that process can unconstrain the growth of many downstream companies. For example, energy is upstream of everything in the economy; think about how much more room AI would have to play with if energy costs were measured in cents rather than dollars. [...]

Instead of being enamored with downstream, consumer-facing internet companies, I wish more people could be excited about upstream, industrial technology companies. It’s easy to love smartphones, the internet, and all the apps we use without thinking about how semiconductor improvements have made a lot of these things possible. Furthermore, I wish that more of these industrial components can improve at the pace of Moore’s Law. We haven’t had quite as much progress in energy, space, chemicals, and medicine that we were expecting decades ago.

To some extent, Moore’s Law is an irrational commitment by the chips industry. It’s expressly driven by an engineering benchmark, i.e. to keep doubling transistor density, which is not necessarily a market- or customer-driven demand. This is a triumph of scientists and engineers over financial types, who would question why an abstract scientific challenge should be invoked for capital allocation decisions. In my view, this sort of irrationality is not a bad thing.
Read this in the context of two recent posts:
Addendum:
When I saw the Berlin Philharmonic this year, I thought about the complex system required to produce music of this quality. The few dozen musicians on stage are extraordinarily talented. They’re so good because the world, and especially Germany, has developed a superb pipeline of talent to staff this orchestra and others. The program was Schoenberg and Mozart, and I thought about how long it took to develop such a deep repertory of pieces that would include these two composers. I can also bring up the technologies required for the orchestra. The Berlin audience was sophisticated, and that takes time to develop too. That is a wonderful system that has gotten a lot of things right.

I had the same type of thought watching UFC as I did when I listened to the Berlin Philharmonic. UFC is an amazing spectacle, and so many things had to be developed before something of that quality could be produced. Think of Vegas, first of all, a remote city in the desert that has managed to attract people, not just for this fight, but year-round for entertainment. Second, consider all the accouterments around the fight: the seamless transitions; the interludes from Joe Rogan and Bruce Buffer, who are both talented announcers; the special effects of lighting, fog, and music, all of which combine to marvelous effect. Third, think about what it takes to market this type of event. And finally the fighters themselves, who know what they have to do to provide a good show.

It was then I felt that I grasped how outstanding the US is at producing entertainment. This is a valuable cultural competence. I don’t think there are any other countries that can develop an audience and put on so many types of high-quality shows.
H/t Tyler Cowen.

Addendum, early Monday morning, 12.24.18:

Notice the trajectory of the essay (one comment for each section, which Wang separates with asterisks):
  1. Moore's Law and exponential growth.
  2. Music: "This year, my tastes in music veered towards the more adventurous. That means I made a conscious decision to dwell less on Beethoven and Wagner." And so he tells us of Mozart, Shostakovitch, and Verdi. Fine music all, but by what metric is it adventurous?  I digress.
  3. "I’m spending most of my time studying Chinese industrial policy and the country’s technology upgrading process. This was a busy year for work given the escalating trade war."
  4. "Enough highlights; let’s get on to self-criticism." Wang vows to write more, but "I’m more aggrieved by my lack of movie and television consumption in 2018."
  5. On to books, fiction and non-fiction.
  6. And the, at last, to television:  "The most novel thing I watched this year was the UFC fight between McGregor and Nurmagomedov. I found it a deep experience..."
  7.  "I failed this year when it comes to the most important type of learning activity that I do: playing enough sessions of Avalon, my favorite board game. Kevin suggests that he’d like to promote Avalon to become the golf of tech. I want that as well, so I’ll take this chance to evangelize the game." 
  8. Finally: "I have several questions."
"The golf of tech"–the game where deals are made? See my addendum to On the deep utility of parties as sites of transformation.

So, we start with serious business stuff, Moore's Law & the economy in general, and then to highlights in the expressive domain, and then back to business. That's a bit of a ring-form, no?

Then to self-criticism. Call that the center.

And then to books, fiction as well as non-fiction, followed by TV with a ritual climax (a UFC fight), and then to a game, an expressive activity where he is an active participant rather than a consumer.

Call the final section a coda, a pendant. It's outside the main form.

If so, that makes the post look like a ring-composition, rhetorically speaking, where the pivot comes with self-criticism. So, before the center he's full-composed, looking out on the world: business-art-business. Then he looks inward. And comes back out for another trio, with that televised UFC ritual in the center.

No comments:

Post a Comment