Blog by Paul Golding

Within the Futures and Innovations group at O2, I currently lead the technical "Platform Strategy," which is all about finding new ways to enable open innovation and co-creation of services that utilise underlying O2 assets for a number of ends. I also lead the O2 Incubator scheme (more experiment at this stage), which is thus far going well.

There is a job opening within the newly formed platform team.

However, this isn't just about Network-as-a-Service (listen to my podcast interview "Open Mobile" with Andreas Constantinou). Much of the role is about looking at new platform opportunities - i.e. creating new capabilities and new network effects that combine with the substantial O2 (and Telefonica) base. There is a lot of thinking and experimenting to be done in the area of low-friction interaction between telco services and Web. Much of the platform work involves re-thinking what a "connected service" is in the age of the "Social Web." Lots of telcos talk about this stuff. I'm trying to do stuff, not just talk about it.

Under my technical direction, we have begun a series of experiments, such as the creation of #Blue, which is an evolution of the original Bluebook service, but built entirely using lean-web methods and atop of modern scalable storage (e.g. MongoDB). Whereas the original Bluebook was (and is) positioned as a "back-up", which is a useful and highly requested service, its delivery is relatively closed, unable to exploit the potential for other services using the data. 

One of the innovation questions for #Blue is how might interactivity with texts on the Web create new types of user experience, somewhat uniquely because of the network ability to carbon copy text messages in mid-flight. One way to find out is to try new ideas, which is easier when using a lean-web set-up. The other is to let innovators dream up their own ideas and build them on top of the #Blue API, which is what we did at the WarbleCamp hackday, leading to Adam Burmister's interesting smsowl.com hack, which copies P2P texts to Twitter if the sender includes a #owl tag.

Although I'm told that the use of the term Ninja is now un-hip, the role was originally penned as "Platform Ninja." Of course, I should explain the term... It's a multi-tasking role that requires a fair whack of strategic thinking supported by an acute knowledge of Web trends (social and technological) and Web technologies, combined with the ability to code. The coding skills need to be proficient hands-on and able to work or play in a range of languages and frameworks in order to build demos and try out ideas, or perhaps just to sniff around an interesting open source project or API to see what's up. The role requires the combination of all these skills to their maximum combined effect, hence "Ninja."

As a strategist and technologist, you will need to demonstrate clear thinking about the future and then be able to evangelise it. However, this isn't about creating slide decks. As a coder, you will need to show the next step ... to go build something, or find a way to get it built, thereby demonstrating or proving that the strategy makes sense.

If the role sounds fun and exciting to you - then waste no time in applying.

I thought that I'd just give a summary post of what I've been up to and thinking of late in my many wireless wanders.

As many of my associates know, I spend a lot of time consulting at O2. I'm now responsible for platform strategy. I was already playing a "tech push" role, evangelizing various Web and start-up methods into the enterprise as a catalyst for new types of thinking and innovation. This is how I founded the O2 Incubator. I can report that the winning teams and ideas, which I hope to share with Debi Jones on the new Telefonica Developer blog, are showing exceptional promise. The "incubator experiment" appears to be working. One idea is in the area of business events aggregation and the other is in sentiment analysis of UGC sources. The teams are fantastic and I really get a buzz out of watching them develop their ideas.

(p.s. for those of you interested in working in the O2 Platform team - tech visionaries with the ability to still think in code when necessary - I'm looking for a couple of "platform ninjas" to join the team.)

As for platform strategy, my focus is mostly on trying to put the right ingredients in place to enable open innovation. This isn't just about NaaS and APIs. There are many ways to think about platforms. The issue is trying to get the telco mindset to adopt a different method of working where the final result of a project isn't known in advance, unlike most projects in a mature operating environment. In this sense, much of the education and evangelizing internally is about trying to encourage "start-up" thinking.

We successfully used start-up-think with the #Blue project, a new take on the existing (poor) Bluebook service. It was an experiment in pursuing various "Lean start-up" and Web-venture methods, that I won't elaborate here, but which proved to be very useful. The project will continue into a live service beyond the current prototype. There are some very exciting developments in the pipeline, many of them radically different to the telco "business as usual" approach.

I am on the fifth chapter of my new book (working title "Connected Services.") I'm currently writing a chapter about the "Big Data" developments on the Web. It's a wide set of technologies dealing with data sets big enough to require distributed storage and processing. This rapidly leads to the exciting area of schema-less storage solutions, like Hadoop, MongoDB and Cassandra. These are all interesting, but I'm much more interested in the "meaning of data" - analysis, statistics, regression, prediction etc.

I like what Hal Varian, Google’s Chief Economist, said in the Jan 2009 McKinsey Quarterly:

“The *sexy job* in the next ten years will be statisticians… The ability to take data—to be able to understand it, to process it, to extract value from it, to visualize it, to communicate it—that’s going to be a hugely important skill.”

I got excited because my wife has a statistics degree and I spent most of my early career working with signal processing, pattern analysis, even neural networks. Perhaps a new career as a "Big Data Guru" is in order. At least I know many of the tools for "thinking" about data. (That said, when I read my 1997 paper about Fuzzy Clustering recently, fuzzy is what it was - I couldn't think how I'd ever written it!)

Ultimately, the analysis of data is where the action is ultimately going to be. Big Data is really about getting value from the processing of unthinkably huge amounts of data. For some interesting provocations, I suggest reading the book Super Crunchers.

By the way, I recommend the Kumon maths system for those of you with kids. It's not about stats, but about really getting to grip with mathematics and numbers, based on encouraging independent learning.

And talking of kids learning, some of you will know that I have had various attempts at teaching my own kids computing and programming. I still like the approach of tools like Alice and Scratch. However, the bridge between these and a "grown up" system is very large. Also, I really want something that works on mobiles.

UPDATE [14-July-2010] - Google has the App Inventor, which actually uses the same graphic UI framework as Scratch.

What I'm really excited about is producing my own programming environment. It's merely a thought-experiment at the moment, called Spawn. The essence of the idea is to create a new programming model that is inherently distributed from the start and can run on desk, embedded or mobile platforms. The model is heavily influenced by the idea of modelling processes as organisms. The default UI paradigm being considered is also "virtual/augmented reality." I've written many times before about how inappropriate I find current UI paradigms are for teaching kids about computers - they don't really know of the history of disks, folders, icons, threads and all that stuff. In this regard, this is the genius of the iPad - it's lke a computer without being a computer.

Spawn might sound a bit grand, but the kernel of the idea - and the objective - is very simple: to allow kids to program immediately and to learn through doing, progressing in the same way a child does with reading. I recently met with a programming guru from the Telefonica R&D lab in Barcelona, who had similar ideas. I suggested we just go build it! I'd be interested to hear other views on the topic of making programming accessible to kids.

I recently gave the opening keynote at the Eduserv Symposium 2010 - The Mobile University. (Slides and video below.)

Andy Powell did a fantastic job of chairing the event and of posting a useful summary (back channel - #esym10), which includes links to other participants' thoughts. One recurring theme that I heard was how most lecturers were a long way behind their students in adopting new technologies, mobile included. Whilst not surprising, this is nonetheless a worrying prospect for the future of education, which I expressed as fantastic opportunity for educators and innovators in the UK.

No point in rehashing all the coverage of Chirp, as I have been tweeting live from the event - what else? You can read my stream, or scan the coverage on the #chirp stream (or the #chirpqa back channel).

Just a note here to convey my impressions of the event and to describe my hack.

During the warm-up music, they played Pearl Jam - some old school rock. To me, it summed up the whole event at an emotional level. I recall seeing Pearl Jam backing in their early days. It was their first London gig at the Brixton Academy. The "early days" is the sentiment here. It was one of those gigs that you could say (years later) - "I was there, at their first gig." It was during the early history of grunge about to go mainstream. (The pre-history was lost on UK-ers who couldn't attend the earlier gigs of the Sonic Youth era.)

Ditto the feeling for Chirp, the first ever Twitter developers event. It feels as though this is a special event at the beginning of history. What history is that? The transition of the web to the real-time web, or the "streaming web."

I've been a Twitter user (on and off) since it's earliest days on the radar. I picked it up via a friend who knew that I was working at the time on a site called Thumbcrowd, which was intended to be a text-message group-share service, Twitter-like. With no funding and not a chance, I ditched it.

I can only look with excitement and awe of what these guys are doing. The platform and the ecosystem around Twitter is, to use the fave adjective of the day "awesome." (Although the new verb of the day, which must surely adorn any "streaming web" biz plan is 'Curate.' I'll leave you to discover its meaning in this context.)

This is surely an ecosystem that is here to stay - and about to explode. It will evolve at a rate of knots. Some of the new API announcements from Twitter, like @anywhere and the User Streams are going to blow an even bigger hole in the net. Check out the new Twitter developer site for details.

The announcement, at last, of a business model, is also interesting. Lots of questions about promoted tweets and whether or not the concept of "tweet resonance" is the new secret sauce of search. Who knows? The Twitter execs certainly didn't seem to know. But it seems they're in no rush. Their ambitions are for 1 BILLION users. With that kind of ambition and the talent that they seem to possess, it's not an unlikely target.

Moving on to my hack...

It's a simple idea, but something I had been dying to try. 

I have a US number that I rent from Twilio. For the hack, I connected it to a backend that scans my tweets and uses them to control the call. I can DM a tweet to the Twitter account associated with the number and the content of the tweet will be used as the voice announcement upon answering (using text-to-speech).

If I append the hashtag #call, then it means that I'm available to take the call, in which case the announcement will play out, followed by "connect you..." and then a call forward to my mobile. If I'm not available (leave off the hashtag), then it forwards to a voicemail (to email) service (which also includes speech to text).

It seems like the obvious thing to do. After all, if Twitter is all about "my status," then that's what I want to use to control comms. It just seems more natural.

If I miss a call, I get a direct message to tell me the caller ID, so I could phone back if I wish. I had also been thinking to use this to initiate an immediate IM chat online (like one of those Livechat services), but that's for another time.

I'm around at day 2 for anyone who wants a demo.

The phrase "Smart Pipe" is an oxymoron, which is a word I love because it says "moron" on the tin - and it's generally morons who promote such nonsensical ideas.

I mean, give me a break. A pipe is a pipe is a pipe. In the real world, what would a smart pipe look like? What would it do? Is it like a hose pipe with a high IQ?

When people use this term, it's just a lame attempt to validate something that doesn't actually exist. If it did exist, then prove it! Show me, Mr Smarty Pipe what you can do....

I'm waiting (last x years)....

Didn't think so.

I've had a GSM phone since they were launched. Heck, I helped design and launch parts of GSM and have the patents to prove it. I was a customer of Vodafone for eons. I would have probably remained so, had it not been for the iPhone. Oh - because Vodafone are so bloody great? Nope. Because who cares?....

They didn't. When I left, not a peep - after 10 years as a loyal customer - going up to 15 contracts with them back in the day of running my own mobile software company before it got screwed between one big company and another over a patent dispute.

Let me boil it down a bit.

I have been running the O2 Incubator program. It was my idea. It's not about smart pipes. It's about smart people. One clutch of bright developers met with me in O2 Media's swanky new Soho offices (nice enough that I might just switch to advertising and use the word "edgy" as my uber-adjective - and thanks to my brother Vince for making up the word "disedginess," which he warned me to avoid.) 

One of the developers told me how many hours I slept and when they were likely to be. In other words, he knew my sleep patterns. Pretty cool, although he was slightly off. But when we discussed why, we figured there was probably a way to improve the algorithm. (I would have been even more impressed if it had figured out that I suffer from insomnia, but I think that's do-able too.)

"And how long did it take to write the code to figure out sleep patterns from the net?" I asked.

"I think I started at 5...," turning to his mate, "Was it five?"
Mate: "Yeah, I think it was about 5..."
"And I had finished by about 10..."

Initially, I thought he must have meant "the 5th" and finished on "the 10th," as in "of March," or something. Five days. Not shabby. Then again, maybe he meant his 10th Red Bull. He didn't. He meant it took him five hours. Within five hours he had written some software to go figure out people's sleep patterns based on what they do on the web.

Smart people can do smart things in incredibly small amounts of time.

And that's what matters. Smart pipes, if they do indeed exist (which they don't) are not going to emerge from dumb pipes. Generally speaking, operators fail to escape the laws of entropy. So, dumb isn't about to morph into smart any time soon. The opportunities are there: The brands, the customers, the assets, the golden "billing relationship," and all that other stuff.... except, that is, the people.

It's not that the people aren't smart. They're just the wrong sort of smart.
And I'm probably no smarter...

But those guys with the code and the sleep patterns... they are smart. And there are others like them, like the smart guys I'm working with to create the #Blue service for O2. I didn't even give them a spec. Heck, they knew what to build without me even telling them (did they use that algorithm to figure out my app habits?)

And that's the issue with all this "Web 2.0" stuff and BIG DATA, and every other such meme, pattern and trend on the internet. It's a language, a way of thinking - and even a way of being. There are certain cultural norms in the internet world that just don't exist elsewhere, least of all in a business predicated on building a pipe. Not to underestimate the skill and resources required to build infrastructure, but it's the cathedral, not the self-organizing bazaar.

Marc Andreesen said it well, as featured in my visual tale of BIG DATA and operators, when he said:



The smart bit of the smart pipe, which doesn't exist, is rapidly building up on in the internet, which is what I was trying to say in my pitch at MWC -- skip straight to the last slide for this point. The new "infrastructure" of connected services is self-assembling on the internet. To operate in this world, operators need to become internet companies. OK, the "become" bit is the problem. They can't.

People who build pipes don't get this stuff. It's not their fault. They're good at what they do. Web platform people are good at what they do too - and this is the new infrastructure. It's a software one. It's being built as operators watch.

Get some smart internet people. Figure out how to standardize some of this "connected services - or contextual - infrastructure" stuff and take the power back. It's never too late to innovate.

There are an increasing number of ecosystems designed to assist start-ups: mentoring schemes, incubators, developer communities and so on. I have been involved in all manner of these, inside and outside. It is great to be running the incubator scheme for O2. We have had some brilliant applications from lots of great teams. This is just the start - there are many ideas and schemes in the planning.

I'm really excited about a few projects just now.

Following on from the announcement by O2 of their O2 incubator program for start-ups, the program has been upgraded to include the possibility that multiple start-ups can be funded in the initial phase. This will enable more than one idea to be run at the same time, so long as the ideas are good enough to impress the O2 SME team, who are the initial sponsors of this campaign.

One of the exciting projects for 2010 for me is #Blue, which is a project that I am leading for O2. It involves the innovative use of mobile network features blended with web 2.0.

Launched only last Thursday (17th Dec), the O2 Incubator program has already attracted a number of applicants and potential applicants, and not just from the traditional mobile developer circles. The project has piqued the interest of some web dudes who probably haven't heard of Mobile Monday and don't do app stores. This is great news and I'm excited to be leading this initiative for O2.