FRNTIER: VC: The Importance of Lindy
The longer a technology survives, the longer it can be expected to live.
|Dec Thomas||Jan 29|| 5|
FRNTIER does research for venture capital. You can read more about our approach here.
Consider liking, sharing and subscribing.
This essay aims to do two things.
First, determine the implications of Nassim Taleb’s Lindy Effect for VCs.
Second, propose a sister to Lindy and determine its implications for VCs.
Taleb’s negative forecast was this: “Old tech survives; the new is replaced by the newer… innovation is likely to attack the new and spare the old.”
This was a reference to Taleb’s Lindy Effect which we will consider below; but, first, it will be useful to have a common definition of technology.
What is technology?
Consider two quotes:
“Properly understood, any new and better way of doing things is technology.” (Peter Thiel); and
Taken together, these quotes provide clarity about what technology is. It is:
Anything which is a fundamentally new and better way of doing things; this can include:
Making a 10x improvement; and/or
Combining existing things in a new and better way.
What is Lindy?
Taleb’s Lindy Effect is this:
Two points require emphasis.
First, that Lindy applies to the non-perishable, i.e. have an informational nature.
“A single car is perishable, but the automobile as a technology has survived about a century (and we will speculate should survive another one).” (Antifragile, p. 317)
Second, that Lindy is “not about every technology, but about life expectancy, which is simply a probabilistically derived average” and that it “should work on average, not in every case” (Antifragile, p. 319), i.e. Lindy does not necessarily mean that a new(er) technology will not endure; only time will tell if it does endure*.
*There are patterns which allow VCs to speculate that a new technology/company will endure, e.g. whether the technology benefits from network effects. However, these patterns seem insufficient to guarantee durability.
The question becomes: what can(not) a VC do with Lindy?
What can(not) a VC do with Lindy?
A VC cannot use Lindy to make positive predictions about which new technologies/companies will arise. The reason is simple: Lindy is not a predictor of new technologies/companies arising; it is an estimator of life expectancy of existing technologies/companies.
A VC can use Lindy to determine the (anti-)fragility of an existing technology. The reason is simple: Lindy is an estimator of life expectancy of existing technologies/companies; it is not a predictor of new technologies/companies arising.
The question becomes: how is this useful to a VC who invests in new technologies/companies?
How is Lindy useful to a VC?
The answer to this can be found by superimposing our definition of technology on Taleb’s negative forecast that “old tech survives; the new is replaced by the newer”.
By doing so, we realise that VCs can back 5 types of company:
Creating something fundamentally new and better aimed at dis-/re-placing an old(er) technology; and/or
Creating something fundamentally new and better aimed at supplementing/augmenting an old(er) technology; and/or
Creating something fundamentally new and better aimed at dis-/re-placing a new(er) technology; and/or
Creating something fundamentally new and better aimed at supplementing a new(er) technology; and/or
[Creating something fundamentally new not aimed at dis-/re-placing or supplementing something existing. (It is not clear that this is possible, i.e. everything seems derivative to a greater or lesser extent.)]
In short, the combination of Lindy and our definition of technology provides a measure of (un)certainty for a VC. The following seems true:
There is greater certainty for a VC when backing a company which aims to augment rather than replace an old(er) technology; and
There is lesser certainty for a VC when backing a company which aims to augment rather than replace a new technology.
[To be clear, Lindy is not the exclusive consideration for a VC when making an investment decision.]
We can explore these options with an example: computer-mediated-communication (CMC); specifically, text-based professional CMC.
As with any technology category, text-based professional CMC tools follow a fat tail. The obvious fat tail technology in this instance is email. This will be our focus.
The old, Lindy-proof technology: email (invented in 1971; i.e. it has existed for approx. half a century so we can speculate that it will exist for another one).
The new technology aimed at replacing the old technology: Slack (2013; chat-based collaboration; aimed at replacing (internal) email; “Ultimately, Slack is more than email replacement. It is a new layer of technology that brings together people, applications, and data.” (Slack S-1)).
The newer technologies aimed at replacing the new technology:
The new technology aimed at making a 10x improvement to augment the old technology: Superhuman (2015; ‘the fastest email experience ever made’).
The new technology combining existing things to augment the old technology:
To apply our measures of (un)certainty:
A VC backing Slack* (new replacing old) would have had lesser certainty when making the investment decision, i.e. higher technology risk; therefore, would have needed a contrarian reason why it would be able to replace (internal) email (and a contrarian reason why it would not be replaced by a newer technology (at least until after a liquidation event)); and
*To be clear, Slack could be positioned as a 10x improvement on preceding chat-based professional CMC technologies (e.g. internet relay chat) so that our positioning it as a ‘new technology replacing email’ may not be fully accurate.
A VC backing a company aimed at augmenting email (e.g. Superhuman, Front, Consider) would have had greater certainty when making the investment decision, i.e. lower technology risk; but, would still need a contrarian reason why the company would not be replaced by a newer technology (at least until after a liquidation event).
To repeat: Lindy requires that a technology exists to be able to estimate its life expectancy; it does not tell us about the length of non-existence, i.e. how long a proposed technology will not come into existence.
Therefore, perhaps Lindy has a sister. We can call it the HS2/Van Ness Effect.
HS2 is a proposed high-speed railway in the UK. The project was officially launched in 2009. However, work has barely started and latest forecasts are that the first trains will not run before 2028 and final completion being between 2035-2040.
“San Francisco proposed a new bus lane on Van Ness in 2001. Its opening was recently delayed to 2021, yielding a project duration of around 7,300 days. “The project has been delayed due to an increase of wet weather since the project started,” said Paul Rose, a San Francisco Municipal Transportation Agency spokesperson. The project will cost $310 million, i.e. $100,000 per meter. The Alaska Highway, mentioned above, constructed across remote tundra, cost $793 per meter in 2019 dollars.”
We can state the HS2/Van Ness Effect as follows:
The longer a technology does not come into existence, the longer it can be expected not to come into existence.
It is true that HS2 and Van Ness are infrastructure projects so that a more accurate statement would be: The longer a technology is not implemented, the longer it can be expected not to be implemented.
[Note: In Antifragile (pp. 283-285; PROJECTS AND PREDICTION), Taleb provides us with the reason why projects are taking longer (and longer) to complete. A very crude summary of the reason: increased (i) uncertainty, (ii) complexity, and (iii) (negative) non-linearities (asymmetries, convexities) in the world.]
Perhaps HS2/Van Ness applies to technologies e.g. autonomous vehicles, AGI and nuclear (fusion) energy. These all were first proposed a substantial time ago, but remain non-existent and/or unimplemented.
Integrated Logistics Services@IntegratedLogi5Magna shutting off costly partnership with Lyft https://t.co/M5S63MMKAP via @freightwaves
It seems true that those things which are most HS2/Van Ness-prone face the greatest technology and regulatory risks (which only serves to compound the length of non-existence).
The question becomes: what are the implications of HS2/Van Ness for a VC?
What are the implications of HS2/Van Ness for a VC?
It seems that there are two main implications.
First, that the longer a technology does not come into existence, the more a VC needs a contrarian reason to back a specific company working to bring it into existence.
Second, that if a proposed technology is inevitable but is/has been vulnerable to HS2/Van Ness, a VC should maximize acceleration towards it. For example, consider Josh Wolfe and Lux Capital’s thesis on energy:
“The undeniable trend [is] more and more energy density per unit of raw material”, i.e. nuclear.
There is no greater acceleration than founding a company, i.e. Kurion (nuclear waste).
Technology can be defined as: anything which is a fundamentally new and better way of doing things; this can include making a 10x improvement and/or combining existing things in a new and better way;
Lindy is an estimator of life expectancy of existing technologies/companies; it is not a predictor of new technologies/companies arising;
By superimposing the definition of technology on Taleb’s negative forecast that “old tech survives; the new is replaced by the newer” we can understand how Lindy is useful to a VC; and
HS2/Van Ness is an estimator of non-birth expectancy.