Feted as one of the first computer scientists to be involved in the internet’s early days, Jon Crowcroft discusses why the UK is poised to deliver on its technology promise, as well as his latest work in AI.
Despite reeling under the effects of a stalled post-Covid rebound, post-Brexit uncertainty and the economic shocks of Russia’s incursion into Ukraine, “the UK is in a really good place at the moment”, says Jon Crowcroft. This is “because we have this distinctive ability to do fundamental research”.
More than that, says Crowcroft, who is professor of computer science at the University of Cambridge and co-founder of artificial intelligence company iKVA, “what we’re really good at in the UK is taking that research and actually going on to make the technology that’s based on the research”.
Crowcroft is commenting on a statement from the UK’s Digital Economy Council claiming the UK tech industry is now valued at £1tn: “A milestone only achieved by China and the USA so far.” The headline has all the hallmarks of a watershed success story for UK plc. But by these figures, even factoring the recent tumble in Apple’s market capitalisation from the heights of $3tn at the start of 2022, this one American company alone is worth almost double the UK’s entire tech sector.
It might only be a feelgood marketing milestone, but Crowcroft still thinks there’s plenty of room for optimism. Referring to the so-called ‘three ‘Rs’ that could put pressure on the future growth of the UK’s technology sector (Russia, recession and interest rates), he suggests that “it’s not clear how any of those things will play out long-term”, or whether their effects will be as significant as “wider global challenges such as global heating”.
The idea that British science is a linear process is attractive to Crowcroft, who has been serially feted as one of the first computer scientists in the UK to be involved with the internet – “when it wasn’t even the internet, it was the Arpanet” – whose work has had a profound influence on the evolution of broadband, multimedia and networking.
We are sitting in a busy common area of London’s Alan Turing Institute, drinking coffee dispensed from a wildly over-engineered system that inevitably contains digital automation, artificial intelligence and a touchscreen front end. But Crowcroft, whose presentations are legendary for their enthusiasm and energy, sees only good in this, explaining that investors expect to find real-world applications of the technology they’re funding when they come to visit.
There’s a serious point in all this, because “Alan Turing famously didn’t just come up with a bunch of theories. He drew pictures in notebooks with circuits and built the ACE computer. He went all the way from fundamental theory to making stuff.” The reason for meeting Crowcroft at the Turing, is simply that the 65-year-old guitar-playing scientist is currently the Institute’s Researcher at Large, an honorific he appears to enjoy: “Makes me sound like Paddington Bear, doesn’t it? Do you remember the book ‘Paddington at Large’? He goes around causing all kinds of accidents that turn out to be serendipitous. That’s exactly the idea here. It’s really just about talking to people and looking for interesting solutions.”
Crowcroft says this practical application of research heritage goes all the way back to the mid-17th century, to the founding of the Royal Society, to which he was elected fellow in 2013. His nomination highlighted his “many seminal contributions to the development of the internet”. Citations for other distinguished awards have contained phrases such as “endless enthusiasm and energy”. It is this long history of coming up with the goods, says Crowcroft, that will stand the UK technology sector in good stead. The Turing Institute currently has 300 open-source projects.
Meanwhile in Cambridge, “we have companies such as Arm making more chips under licence than anyone else. That’s interesting because there’s no big factory there to see: they sell the designs to people. They do all the clever stuff and not the grotty work. They make more chips than Intel, so why no British bank got involved and said ‘Hey, we’ll own you’ is surreal.”
Throughout Crowcroft’s long career, “most of what I do has been on the software side.” He states that while “the UK’s contribution to the global market is huge”, it’s not often visible because a lot of the development happens in “tiny organisations, just two people and their pet cat dotted around the place. If you go around Cambridge there are loads and loads of small companies, some of which stay as cottage industries, while others turn themselves into fancy bespoke design houses. Then occasionally you get something in software that goes big, like BAE Systems. The UK is also significant in designing satellite systems – so there’s a lot of stuff out there.”
This landscape of generating basic ideas and seeing them through to market is “unusual. We’re second only to the US in terms of making new ideas in terms of cool papers being published. Way ahead of China. And a lot of this happens in Cambridge.”
Much has been written on the Silicon Fen phenomenon, a large cluster of software, electronic and biotech companies that had its origins with the likes of Sinclair Research and Acorn computers and is now recognised for being home to global organisations such as AstraZeneca as well as countless tech start-ups. Crowcroft attributes the Fen’s success to “the intellectual atmosphere. One of the contributing factors that’s interesting is the extremely rapid free movement of people between organisations: the university, commercial companies and even non-profits. The only other place in the world that had this freedom was the original Silicon Valley.”
The success of Silicon Fen is also a self-fulfilling prophecy, he suggests, in which, if you get enough intelligence in one place “things start to happen”.
His favourite example of such informal collaboration is the Raspberry Pi single-board computer for use in education. “There was a group of us sitting in the pub and we thought, ‘Wouldn’t it be good to do this?’.”
Crowcroft is a co-founder of iKVA, an early-stage AI knowledge management company formed in 2018 out of the University of Cambridge’s Computer Laboratory and the Alan Turing Institute. The company’s mission is to create value from an organisation’s data using its Insight Engine, which can break down data silos and access previously unreachable knowledge in engineering, software development and legal document scanning applications.
Crowcroft explains how iKVA’s Insight Engine uses vector search technology, “which is a fundamentally different approach to keyword and ‘smart’ search”. Instead of trying to directly match a keyword or a string of keywords, which has to be user initiated, vector search takes a more contextual, richer input source, which could be a whole document, a URL, audio or video. It then converts it to a vector and matches it against other vectors within an index that iKVA calls a ‘knowledge pack’. “Vector search does not look for direct keyword matches, but instead for ‘nearest neighbours’ from the full document content, allowing it to return relevant and accurate results.”
Crowcroft read natural sciences at Trinity College Cambridge in the 1970s while listening to “loud and nasty punk music while all my friends from school were still listening to Prog Rock”. This may sound like an irrelevant nugget of local colour, but it’s a detail that demonstrates how half a century ago he was already, as his 2009 SIGCOMM Award citation states, thinking ‘outside the box’.
His interest in software predates his late teens by a decade when, taken to the Royal Free Hospital on a primary school field trip, he saw his first computer. “For some reason we were shown how to program in binary. I later recovered this memory while programming on an IBM 360 and got a master’s and a PhD, halfway through which someone came in and said to me, ‘Do you want to work on this project to do with the internet?’.”
A few years later, Crowcroft was hired as a lecturer at University College London without an interview, “which was really weird and could never happen now”. Then 20 years after that he received a call from the vice-chancellor of the University of Cambridge congratulating him on his new job, again following an application process that can best be described as informal. This unorthodox approach to his career is made obvious on his Computer Laboratory home page on the Cambridge University website, where his CV resembles a cross between lines of code and the modernist freeform poetry of e e cummings, who famously didn’t use upper case characters either. Embedded in the biographical detail is a (slightly modified) quotation from cyberpunk author William Gibson: “the future is already here, it’s just unfairly distributed.”
Regarding the Gibson reference, “nobody believes that he understands the technology that he writes about and yet he comes up with these great insights”. The idea of unfair distribution is a concept Crowcroft agrees with. “One of the things that the early community of people working on the internet thought was that it was going to be inclusive, super-cheap and would get you access to all kinds of things.” We saw some of this during lockdown, he says, when “thousands of people joined in teleconferencing events that they wouldn’t have otherwise been able to get to. But the politics here about unfair distribution is that originally it just reflected existing economics.”
Yet the commercial end of technology has headed in the opposite direction to the original aim of a democratised internet, says Crowcroft, because we now have “all this centralisation” around the hyperscale FAANG companies, “and that’s not necessarily a very good thing”. At this point, Crowcroft returns to the subject of Apple, who he reluctantly agrees are “quite clever” because the “actual manufacture of their product is distributed right across the global economy, but they scoop up all the profit from a centralised node in the States”. He follows this with forthright statements about the tech giants that can be summed up as the internet should be the opposite of what they’re doing.
Having seen the complete history of the internet from the inside, while being a serial technology entrepreneur, educator and innovator, Crowcroft thinks the digital future comes with a massive sustainability bug. We need “to tackle the intellectual challenge of solving the more difficult problem, which is to work out how to build a distributed computing system that has a hundred billion nodes on it”. He says the question is not so dissimilar to how you would construct a stable and sustainable society out of 10 billion people: “One of the things you wouldn’t want to do in that kind of world is send all the food to, say, New York, and then redistribute it from there.”
Yet that is what’s happening with our current computing network, in which “all the data is being sent into giant data centres instead of processing it locally. It’s very lazy to just send it to some central place and then do some humungous computation in there and push the result out in the form of targeted advertising or whatever it happens to be.”
We already have the technology in our pocket to bring about a distributed future, he says, “so there’s no point moving all this data into those giant buildings. Why don’t we just keep it where it is and do interesting things with it there?” Furthermore, if you want to create better models for, say, healthcare data, then “all you need to do is share the model and not the data itself. This gives you the two wins of privacy and less energy consumption. It is much more sustainable to do things this way.”
Sensing that this could be interpreted as naïve, Crowcroft defends the vision, saying: “You can still make money out of it. You could still,” he says with an air of bowing to the inevitable, “push advertising onto people, providing they’ve ticked the box that says they’re OK with that.”
Sign up to the E&T News e-mail to get great stories like this delivered to your inbox every day.