,
25 tweets,
5 min read
Read on Twitter
1) Super interesting and thoughtful work, although I disagree with some of the conclusions. Carlotta Perez’ framework is core to my thinking, but I don’t think the automobile era is a good analog for the present environment. A few thoughts.
2) James Watt’s invention of the rotary steam engine (1763-1775) was a foundational innovation that kicked off the entire industrial revolution, which encapsulates the first 3 technological revolutions that are highlighted (Industrial Revolution, Steam/Railway, Steel/Electricity)
3) Peak social/technological/economic change due to this foundational innovation was roughly 60-75 years later. I see the 1959 invention of the microchip (integrated circuit) as a similarly foundational innovation that also kicked off 3 related technological revolutions
4) These three related technological revolutions – PC/internet, smartphone, AI - are happening in closer succession and overlapping to a much greater degree than the three industrial revolution related epochs and I think we are just entering the peak period of change.
5) First technological revolution as a result of the Microchip was the PC era and the internet, which began in the late 1980s and had a turning point in 2000. Internet is analogous to railroads – both were essential for the next age.
6) The iPhone was the next technological revolution circa 2007 – location, identity and presence aware supercomputers in everyone’s pocket – and it began a new installation phase with much of the frenzy around apps enabled by smartphones taking place in private markets.
7) The emergence of 3b edge devices also coincided with the advent of cloud computing, which made them infinitely more useful and was quite a paradigm shift for the enterprise. No cloud, no Uber, etc.
8) Sidenote: Funny to remember googling "cloud computing" in 2009 - remember because I first heard the term during Q1 earnings season that year. I believe Eric Schmidt recommended a Youtube video by Urs Holzle.
Time flies! 😀
Time flies! 😀
9) We *may* be at a turning point for the Smartphone/Cloud technological revolution, but are already well into the installation phase for artificial intelligence and ambient computing.
10) Just as the internet was necessary for the Smartphone/Cloud era, smartphones and cloud computing is essential for the AI era.
11) As Jeff Bezos and many others have noted, there have not really been any advances in AI algorithms (beginning to change with GANs and CNNs), we just simply had vastly more data available as a result of smartphones and more compute power as a result of cloud computing.
12) Using relatively old algorithms with much more data and much more compute produced astonishing results in many domains. Understanding the importance of (relevant) data quantity to AI quality is critical to technology investing today.
13) I think the turning point for the AI technological epoch is a decade(ish) away and the disruption that autonomous vehicles (the most visible instantiation) will cause will be significantly larger than that caused by either the PC/Internet epoch or the Smartphone/Cloud epoch.
14) Sidenote - this AI epoch will likely be accompanied by a UI revolution as well around augmented reality and ambient, always on computing.
15) Second sidenote: this AI revolution *may* be accompanied be some pain in stocks that benefited from the smartphone/Cloud "frenzy" and are negatively impacted by AI.
16) Net-net, I think today’s technological revolutions are much more akin to the three earlier revolutions which collectively lasted much longer than they are to the automobile/mass production era – a relatively discrete, short technological epoch.
17) This difference along with the increasing overlap of these cycles vs. prior cycles makes the regime analysis in this paper less relevant, IMHO.
18) Would be supercool if @chrismeredith23 could stitch together data from the 19th century - I love this kind of analysis and would be happy to contribute in any way that I could.
19) And beyond the aptness of various historical analogies, I think value investing has been negatively impacted by other factors, not least of which is its increased popularity.
20) Other factors: ROIC are less mean reverting because scale is *much* more important online than offline, we are at a historically high point for regulatory capture and a huge % of the market is run quantitatively on value principles.
21) As highlighted in the paper, regulation *may* change the first two although will believe it when I see it given the extent of regulatory capture. However, quant investing and its effects are here to stay.
22) Simple, easy optical value has largely been arbitraged away by quantitative investing. It used to take fortitude to buy stocks at 52 week lows with terrible fundamentals – it was lonely, which was why it worked. Quantitative programs have no issues doing this.
23) And more broadly, everyone loves Buffet. Even after a long bullish run for growth, very few people are proud to call themselves “growth” investors and the smartmoney invite only investing communities are dominated by “value” investors.
24) This quote from Charlie Munger from a January 25, 2009 @jasonzweigwsj article was interesting - and of course, Berkshire now owns many stocks that are not traditional value stocks. 
25) And rather than say it’s never different, I would say that it's always different - history rhymes rather than repeats. Overall, though I loved the spirit of the paper even if I disagreed with some of the conclusions. Thank you for sharing.
