IA vs AI

At the last Performance Firestarters on The Power of Feeds at Google, our speakers talked about the use of client and third party APIs and data feeds to dynamically update advertising messaging (beyond pricing and stock levels) to add new levels of contextual relevance (location, behaviour, weather, exposure to other media and so on).

It was a fascinating session and it seems that we're only scratching the surface of where this might go. Visar Shabi, CTO at the super-smart BrainLabs, spoke about applying a layer of machine learning (or even utilising prediction APIs) to the input of data so that we might develop a model that continuously updates itself. Alistair Dent of iProspect talked about how APIs bring scalability and speed to managing real-time changes in communications, and all the different ways in which we might use structured data to do this in better ways. Kris Tait of digital agency Croud (who themselves have a really interesting operating model) spoke of how they had used data feeds to dynamically update advertising messages for their client Netflix, making 297,000 changes over a six month period (the equivalent of 50,000 changes a month). 

Kris also described the process of using feeds as being one of augmenting human capability. In doing so he mentioned a really interesting delineation between AI (Artificial Intelligence) and IA (Intelligence Augmentation, or Amplification), which led me to this recent Andreessen Horowitz  podcast (embedded above) featuring the NYT's John Markoff who's just written a book, Machines of Loving Grace, about the increasingly complex relationship between humans and machines.

In the interview Markoff talks about how two communities had grown up in Silicon Valley - one focused on Artificial Intelligence and how technology might replicate or replace humans, and the other all about how technology might augment human capability. He talks about the McLuhanism of how we shape our tools and then our tools shape us (but we do shape our tools), why we'll end up talking to machines but not just through a conversational UI, whether robots really will replace human jobs, what’s really happening with Moore’s Law, and why we’re not seeing an equivalent Moore’s Law-type impact on productivity growth. A useful distinction and a fascinating listen.

The Future of Agencies


I've been working on a major new piece of Econsultancy research over the past couple of months, the output of which is a report that I've written on The Future of Agencies. As well as a quant survey, I interviewed a broad range of senior agency-side practitioners around the world to gain a perspective on how agency models were shifting, and their view on the where things are going.

The results were pretty compelling. When I last did a similar piece of research back in 2012, many agency people I interviewed then were talking about change but there was little evidence that the fundamentals were actually evolving. This time felt very different. There was evidence of major shifts underway in the competitive landscape, operating models, development of capability, client engagements and working processes. It was truly fascinating. I'll be picking out a few of the themes to talk about further here but you can access the full report via Econsultancy if you're a member there or if not via the Adobe Partners (who sponsored the research) site. I'll also be doing the second and third (the first was a few weeks back) in a series of webinars on the findings, the next one of which is on Thurs 24th. You can register for that here. Feedback on the findings, as always, is welcome.

Open Office Hours


I like the 'open office hours' idea expounded here. Apparently instituted by Marissa Mayer whilst at Google, she cleared an hour and a half of her diary at the end of each day and staff could reserve a chunk of that time by putting their name on a board outside her office. This enabled her to reportedly fit a large number of very short meetings into a block of time where staff could come and talk to her about anything. Meetings which apparently surfaced interesting product ideas including Google News. A better option perhaps than filling too much time up with the half hour/one hour blocks that managers tend to segment their calendars into, or to keeping an entirely open door policy which might lead to overly frequent interruption.

The Extensible Self


I'm somewhat biased (since I curated it) but last Friday's Dots Conference, run by the smart folk at Brilliant Noise, was excellent. One of my favourites out of a whole bunch of inspiring and insightful talks came from Stuart Turner, founder of Robots and Cake, who talked about the transformational power of robotics. 

Stuart began losing function in his arms and legs over ten years ago and became a complete quadriplegic when he was 31. But he refused to accept the shrinking world that seemed to present itself to him and hacked together a wide variety of different technologies to help him use his computer, take pictures, create a smart-home and even control a drone remotely through micro-movements. He made a lovely point about how APIs and making technology scriptable and hackable had left the door open just wide enough for him to be adapt technologies to transform his entire life. Technologies that weren't designed with disabled people in mind but that had allowed enough wiggle room, just enough of a crack to let the light in.

Stuart described how flying drones had started to change the way he thought about technology, giving him the feeling of space, control and movement that his body denies him. Something he called the 'extensible self'. The idea of physically interacting with the world remotely. The fact that visiting the National Museum of Australia by using their robot can be easier than going to the end of his road. Disabled people, he said, have grown used to living as disclocated selves, but the idea of interacting remotely with the world, extending ourselves through robots, is something that far more of us will need to get used to in the future. It's a fascinating thought.