Bulletin 23 November 2018: You can’t program serendipity

I’ll be heading off to AWS Re:Invent on Sunday. As well as thinking once again about the paradox of having to go half way round the world to meet tech industry people face to face, I’m reminded of the serendipities of doing so, such as meeting multiple people called Dave

The very nature, and indeed joy of serendipity is that it follows a non-algorithmic pattern that we can nonetheless follow. For example, if I stay at home, I will be far less likely to bump into someone on a station platform who might offer me an opportunity, or make an introduction, or, well, whatever!

Strangely, this goes a long way to explaining the AWS business model. I’m not referring to all that HQ-in-Luxembourg shenanigans but the nature of creating an all-encompassing platform. I genuinely don’t think the AWS powers that be know what it is for, as that isn’t the job; rather, the goal is to ensure that whatever people are using right now is incorporated. 

This goes some way to explaining the astonishing, and continued announcements of new features, functions and packages supported by the AWS platform. The company has done a great job of catching little waves, not differentiating or caring about whether they are important but allowing them to happen, making them available. Let serendipity decide could have been the underlying mantra…

…until more recently. I’ve noticed the rhetoric around cloud providers evolve, not only from its original “everything to the cloud” towards “everything to the cloud but the stuff you don’t want in the cloud”, but also towards “which cloud should I use for what?” As providers mature, they are offering more evaluate-able services, for example around artificial intelligence capabilities and IoT models. 

The result is that genuine reasons are emerging to choose one over another, that is beyond “we have everything you might need” kinds of arguments. It has been difficult to keep up with it all, to be frank, but as the debate moves up a level or two, it will get both easier to understand and therefore harder for providers to differentiate. 

To put it in concrete terms (and to go up even higher in the stack), “We have great virtual machines, fab storage and every open source package you might need,” is a very different proposition to, “We have the best solution for healthcare/retail/<insert industry here>.” Right now we are somewhere between the two. 

Ultimately, cloud providers have been deriving value from that old chestnut of disintermediation, a.k.a. giving people what they want more easily than the traditional alternative. All well and good, one might say, as it has got them to where they are; now this source of differentiation is being used up and as a result of inevitable market evolution, providers are looking to differentiate in other ways. 

We continue to live out what has been termed the platform economy, which has been based on operating build-it-and-they-will-come models at a most strategic level. The future will continue to lie in building on platforms, but as lower levels commoditise, so will battles be fought in terms of the solutions, not the platforms, on offer. Proactive serendipity may have had its day. 

In other news, here’s an article for this week. 


On Value Stream Management in DevOps, and Seeing Problems as Solutions

A frequent habit in tech is to describe something that isn’t working in a way that suggests an answer: “Data Leakage Protection” is one of the more obvious. In business circles meanwhile, the whole notion of efficiency relates to the fact that much of what we call work is unproductive and pointless. I know, I know. So, Value Stream Management is both an answer and a set of questions — what’s not working, and how can it be fixed? When applied to DevOps, it’s also an indicator that it doesn’t deliver pots of gold without a bit more forethought. 


Extra-curricular: Super-Awesome — the Musical continues apace

Sometimes, things move from a nice idea to a “blimey, this might actually happen” state almost by stealth. I started writing a musical on a whim in 2010, and completed a first draft earlier this year: suddenly I have collaborators, songs and, well, if you’ve never heard something you’ve written sung by a professional singer, I highly recommend it. Now planning a reading early next year, followed by a workshop and then, perhaps a Kickstarter but a long way to go before that happens! Keep watching this space…


That’s all for this week, thanks for reading!

Cheers, Jon

Bulletin 23 November 2018: You can’t program serendipity

Leave a Reply

Your email address will not be published. Required fields are marked *