Bulletin February 9 2018. AI and the theory of exponential linearity

You know that thing where, in a discussion, a concept comes seemingly out of nowhere, but then changes the whole dialogue? So it was in a recent conversation with my old friend and colleague Roger Davies, him who has written whole books on Value Management (kind of, what to do when you realise Balanced Scorecard can only take things so far) and who is currently working on his thesis. 

“Of course, despite all this innovation, our progress has only been linear,” I blurted, immediately finding myself in a position where I had to justify my remark. “Is that true?” asked Roger. “Yes,” I said, sounding much less confident than I felt. But as we explored the idea, it did seem to hold water — are we any better at communicating, for example, or really any more productive than we were two or three decades ago?

This question is writ large when it comes to Artificial Intelligence, whose impact, we are told, is to be transformative. I admit to being flummoxed in another conversation, this time just before Christmas at the Great Telco Debate, when I was told in very clear terms that AI would, indeed, change just about everything. He clearly interpreted my uncertainty as ignorance, as he undertook to explain (as one might to a child) why this should be. 

Interestingly, the example he cited was telecommunications complexity, what with Network Function Virtualisation (NFV), 5G and the like. I was almost immediately taken back to the early Noughties, when I presented on neural networks in enterprise management, on behalf of CA. Also on the bill was an Italian researcher with several PhDs — “Maybe we could share research,” he suggested. I politely declined, not wanting to show him up with my hastily collated notes and superficial thinking. 

The thing about all this fantastic innovation is that it has a fractal effect on the problem space. I know, that’s me getting pompous again, but it happens every time: we are as sorcerers’ apprentices, generating data as quickly as we work out new ways to harness it. The impact on most technological fields is that breakthroughs tend to be quite tightly scoped — at the moment, voice, image and other signal processing are seeing the most progress, and will rightly yield some great results alongside the dumb animations we can add to video chats. 

But nobody has yet created an algorithm that can somehow get ahead of the game. To my delight and relief, someone far smarter than me has captured this in an article, “The impossibility of intelligence explosion,” i.e. that point at which our own intellects get left behind. By extension, as long as they (our intellects) do not (get left behind), we are stuck with them in all their wonderful, dumb glory. In the meantime, AI will serve to augment, not replace capabilities including that litmus-test topic, radiology

Don’t get me wrong, these are transformative times. But even so, our relationship with technology remains that of a craftsperson and their tools — it is up to us to prioritise, to architect, to make things happen. We should not expect our environments, or relationships, our abilities or anxieties to be that much different in ten years’ time, compared to now. And that is a good thing. 

And so, to some articles. 

The digital world needs new lawmakers

</soapbox on>In case it isn’t clear enough from this article, I’m none too comfortable with the repeated, deal-with-problems-after-they-happen approach to lawmaking in the information age. Moves are being made to change this, notably in the financial industry, but here in the UK for example, our general legislative processes still take place over months or years. Not only is this too slow, but it is massively inefficient and open to exploitation by corporations and cybercriminals alike.

As Einstein said, “We can’t solve problems by using the same kind of thinking we used when we created them.” Word.</soapbox off>

On luxmobiles and flying unicorns: how diversification and proliferation will rule the routes

On a much lighter note, yes, I very much enjoyed writing this — but it has a serious point. If vehicles are no longer constrained by needing a driver, what will they become? I posited some examples as a staring point:

  • Luxmobiles and budget pods. Like Apple vs Android, the market tends to diverge
  • Pizza box scooters. Expect transport the shape and size of its load
  • Drone swarms. No doubt flying in some sponsored logo formation
  • Super Strings. I should really patent the coupling mechanism now…
  • Public Subs. What about underwater package delivery? Really, why not?
  • Vompods. Awkward but who wants to be the third drop-off after a late night out?
  • Flying Unicorns. For multi-functional transport, the aerial has to go somewhere

The last one was first postulated in our ‘techsplaining’ podcast, the ironic title of which seems strangely appropriate. 

Webinar: DevOps in the Real World

Yes, I’ve been that person who tries to get people to do things in different ways to previously. Indeed, in the course of this bulletin it was good to hook up again with Mary Lynn Manns, co-author of the book Fearless Change: patterns for introducing new ideas, to which I was delighted to contribute something back in 2004 (my proposed pattern was “Champion Skeptic” — who would have guessed!). I’m rambling, but the long and short is that I am well aware you can’t just tell people to ‘do’ agile, or DevOps, or anything else that makes them have to think differently. Even if it’s a really good idea. And even if you are a guru

That’s why I’m really looking forward to having a virtual fireside chat about it all with Nigel Kersten of Puppet Labs. If you’re interested you can sign up here



The whirlwind romance with piano continues – still two hours a night, as the latest video blog shows I’ve started to crack two hands though I’m a long way from being able to just launch into playing a piece. Fur Elise is getting a lot of play at the moment, (at least the first, easier bit). 

And meanwhile that august institution, the Society of Authors has seen fit to re-post my blog on “How Not To Be A Biographer. Which is nice. 

Finally, thank you to all my subscribers. Any questions or feedback, let me know.

Until next time, Jon

Bulletin February 9 2018. AI and the theory of exponential linearity

Leave a Reply

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