What happens when venture capital fuels AI hype?
Liberation capital, the power law and blitzscaling
On its face, venture capital seems simple. Investors raise money, put it into promising startups and take some equity in the hope that they realise their potential.
However, the way in which this is done highlights the important role VC plays in modern technological development. This role could be broken down into three key elements:
Liberation capital
The power law
Blitzscaling
Liberation capital is a long-standing concept with origins from the 1950s when VC as we currently know it started to take shape. It refers to how VC provides an opportunity for budding entrepreneurs whose ideas are too risky for traditional investors.
Accordingly, VC liberates those entrepreneurs. It gives them an opportunity to develop their ideas into groundbreaking inventions that shape industries and define cultures.
But VCs know that developing such spectacular ideas is not commonplace. Among the various ventures that they back, only a few will survive and thrive.
This is the power law. This rule dictates that the success of a venture fund depends on the immense success of only a few companies.
In other words, VCs bet on the success of select startups and hope that they are so successful that they will more than make up for the losses of other failing ventures. They rely on the long tails, the farthest ends in the distribution of returns.
To help ensure these outcomes, VCs encourage blitzscaling. Startups must grow as fast as possible, they must move fast and break things in the pursuit of profits.
Put these elements together, and you have corporate financing system responsible for some of the richest companies in the world. Some would even argue that VCs have helped deliver more modern technological progress than any other vehicle for change.
But even if this were true, such a system is not without its drawbacks. In particular, VCs can be seen as instigators of wild disrupters with little regard for what or who is disrupted.
By liberating the ideas with potentially massive impacts and fuelling them with growth-seeking motives, maybe this ends up being not such a great thing. Maybe this causes negative externalities and unintended consequences that are actually deleterious to society.
What happens when this system is applied to current developments in AI? The sequence of events might be something like the following:
Lots of hype will be generated whereby developers make great promises with their models and VCs will entertain a lot of it.
Most of these AI startups and the models they develop will probably fail and not amount to anything significant.
The ones that do manage to succeed will be the ones that grow the fastest, and in turn could have the greatest impact.
The last part is the scariest, especially when it is being facilitated by the VC system. This is because that system is only concerned with the impacts to the extent it makes investors money, and therefore not concerned with whether those impacts are net positives or net negatives.
This something crudely explained by Meredith Whittaker, CEO of Signal, in the clip below:
When VC fuels AI, there are positives and negatives. But hopefully in the end the latter does not outweigh the former.