IEX #166: Understanding Hyper-scale Networks...
Why Facebook and other Social Networks have much to offer.
Good morning. I’m exploring options for the best time and day of the week to send the newsletter out. What do we think about Thursday mornings?
On to hyper-scale networks…
We all know now that despite a number of noteworthy advances, we find it hard to comprehend the complexity of the human brain, and impossible to recreate it, although we have a sense of the scale of the network, of the billions of neurons and trillions of connections that it holds. On the other hand, willingly or unwillingly we have created other networks, which also operate at hyper-scale - which means that it they are also now too complex for us to control actions and responses.
When facebook went down last week it brought home both the scale and complexity of the network. Yes, a lot of people may have realised that they didn't really need Facebook but, let's take a step back and look at the numbers first. Facebook has 2.89 billion active monthly users. Which means about 37% of the worlds population. WhatApp has 2 billion monthly users. Instagram - 1 billion. Or consider Twitter which has 192m active daily users. LinkedIn has 750m active users. We know that Facebook has been central to elections and the battleground of information vs misinformation over many subjects. But Facebook (and these other hyper-scale networks) also have some common characteristics. Some are obvious, while others we may need to think harder about.
Let's start with the most obvious one - these hyper-scale networks are extra-national. By that I don't just mean that they go across the borders of multiple nations, but they are also bigger than most nations. FB is bigger than India and China's combined population. This means in the simplest terms that FB matters. What people see and think on FB, and the opinions they form, even 1% of them, is a very significant number. So FB algorithms, matter. And FB content policies too.
Hyper-scale networks have inbuilt network effects. What we call going viral is really a power law distribution of re-posting and mentions, for example. And nobody can yet predict the 'virality' of a topic or subject. It's hard to stop a post that's going viral as well.
What we also learnt from FB's outage was that for some people, it's a part of the critical infrastructure. And no matter how much you and I believe we can do without it, for many people, FB is critical - not just physically but socially as well.
Many of these hyper-scale networks have demonstrated exponential growth - creating entirely new economic models, which make a mockery of 'productivity' conversations. Azeem Azhar's book calls this divergence of exponential growth and traditional organisation and governance models the exponential gap.
Although we think of many of them as evil, or direct our ire at Zuckerberg, the fact is that there is no human at the centre of the network making decisions, reviewing data, deciding what ads you should be seen. Yes, there are moderators, but in the bigger scheme of the business model, moderators are peripheral, and might be a transient technology before automation takes over.
And lastly these hyper-scale networks are increasingly critical to identity management. You can already register with any number of services using your FB, Google, or Microsoft id. Do we see a time when they might become the dominant id providers? It is likely that FB can validate you much more effectively and faster than your bank can. The flip side of identity is anonymity and increasingly this comes at a price. Twitter's big play might well be to offer one of the worlds largest anonymity networks. An anti-identity model that trolls and activists both gravitate towards.
The reason that this is important for us to understand is that social networks aren't the only hyper-scale networks. Other much more impactful ones are also around us. Pandemics are examples of hyper-scale networks where our ability to contain virality may be hugely valuable. Global trade and finance is now a hyper-scale network with the kind of complexity and criticality that would probably dwarf facebook - our inability to foresee or prevent shortages in microchips until they are already happening, or our inability to offset the impact of events like the Ever Given blockage of the Suez Canal suggest that we still struggle to deal with the levels of complexity and uncertainty. Or even, as in the case of the FB outage, the vulnerabilities. So while modelling the brain might still be out of reach for us, we may do well to understand how to exploit and manage the hyper-scale networks which already surround us. And Facebook & co give us a great model to work with.
Reading This Week
Innovation Model: Heres’s an interesting blog post on why corporate venturing may be the idea model for innovation success in the enterprise. The difference between R&D and the path to market success is a key aspect of this. (AimForTheMoon)
Hybrid Work: I enjoyed this piece and the approach to flexible working - keeping in mind schedule, place, continuity, workload, and mode. (HBR)
Future Cities: Despite NYPD’s spiralling overtime costs, the city isn’t getting safer. Why? (Bloomberg)
AI Future: A thought provoking paper on creating an AI workforce in the US. (Georgetown.edu)
New Products: Event based camera chips, from Sony. A 7 year journey towards lower cost, low energy, and more flexibility. (IEEE Spectrum)
Have a great Thursday!