National purpose on AI: Lovelace Institutes.
, .
The most recent paper in a series on National Purpose by Tony Blair’s Institute for Global Change is on the new institutions we need to make science and technology the foundation of a Britain that returns to being among the richest countries in the world. All stuff that I’ve worked on and written about for well over a decade. So I was excited to read it.
But first I did what I always do with papers since AI got good. I put the paper into Gemini, asked it to extract the main people and institutions mentioned, look up where the people live and work, and where the institutions exist, and put that in a table.
The split within the UK, with everyone and everything in the South East of England — where just a third of Britons live — and almost nothing outside of it is typical of most “national” reports in Britain. The inclusion of the United States is a feature that has become more common in recent years as Britain’s institutions in the capital city become ever more disinterested in the rest of their country and ever more jealous of an American dynamism that they feel we should import as well as emulate.
In a huff, I refused to read it.
It took the kind invitation by the University of Strathclyde to a fascinating, motivational, and slightly concerning dinner in London with AI Minister Kanishka Narayan to be seated next to the first author of the report Laura Ryan, and realise that I should get over myself and give her paper a try.
In doing so I learned, thankfully, that the AI is slightly wrong. The report mentions Peter Mansfield’s Nobel Prize winning work on magnetic resonance imaging and thus the University of Nottingham. But I spotted no other omissions.
Free money
The report wants free money for scientists to do interesting things, and it wants the UK government to provide that money with fewer of the strings attached that make current scientific research unproductive. It’s a tough sell, but over the course of the 124 pages I came to agree with much of the pitch.
British universities are great at what universities are good at, but they are responsible for too large a part of the UK’s public sector R&D spending. Alternatives such as Catapult Centres have helped to fill the gap, but even strong supporters like me of most of the manufacturing-focused ones wouldn’t claim that they solve the problem. The report could usefully have mentioned them more, but it was already 124 pages long.
Catapult Centres could go further, but we also need other institutions to go in different directions. The report has ideas on that.
I learned about Bind, a focused research organisation (FRO), something new happening in a similar direction to what the paper proposes.
Bind is funded by the UK government and thus, of course, it is based in London.
I also got to learn how the proposed Lovelace Institutes (the provisional name of disruptive invention labs feels unlikely to stick so I’m using a shorter name for them) would supplement ARIA, another new institution going in a new and promising direction but not providing the whole answer to the problem.
ARIA is funded by the UK government and thus, of course, also based in London.
The case for public money was convincing, especially since it needn’t be extra money, just a small slice of the growing R&D funding provided by our government. The universities will fight of course, but it’s a fight that we probably need to have.
In the absence of the kind of industrial monopolies that powered Bell Labs or Xerox Parc, and in the absence of the scale of philanthropic giving that is more common in the USA, and in the absence of British equivalents to Google, Apple, Microsoft, OpenAI, Meta, and Elon Musk’s varied endeavours who can fund huge amounts of industrial R&D at the frontier of technology, we would need the UK government to step up.
National narratives of innovation and AI success.
Back at our dinner we were discussing the need for a compelling narrative about the opportunities that AI was bringing to the country, and the pitch for investment in AI here. The suggestions at a dinner in London were unsurprisingly focused on London, while claiming to be national. Me being the lone voice in pointing this out is partly why the University of Strathclyde invited me.
I mentioned of course that the UK’s most successful tech company, Bet365, is based in Stoke-on-Trent and increasingly in Manchester. It is a company which has achieved, especially with its implementation of in-play betting, about as much of the “global influence and centrality within a transformative new field” that the Lovelace Institute paper claims is unique in Britain to Deepmind.
I wrote recently that Stoke-on-Trent’s tech sector is now larger than Oxford, Cambridge, Brighton, and Milton Keynes. I celebrate that the company’s CEO, the first in her family to go to University, a University notably outside of South East England, has long generated big profits, large tax revenues, and a huge trade surplus for Britain.
Sports, and gambling on sports, both increasingly incorporating AI, are industries in which the UK is truly world-leading. Importantly, both our history and our current global success in these areas are concentrated in North England and the Midlands. I am not sure whether the ethical and moral issues of the industries or them not being centred in the South East explain more of why they are so frequently underrated in London.
My company The Data City classify and track every company in the UK so I know that companies like Bet365 in its early days are common across our country, especially in our dense urban agglomerations. At the dinner I told the story of one, Sewtec, who have long been our top showcase of companies applying AI to manufacturing in the UK.
Sewtec Automation of Wakefield were founded, initially in Dewsbury, by former R&D department employees of the largest Singer sewing machine factory in the world near Glasgow after it closed in 1980. Through innovation they have grown to great success, and last year they were acquired by Automated Industrial Robotics of the USA.
Sewtec’s success is a great story for Britain and the British economy. A company growing out of a declining industry into a new industry in a city without a university in a part of the country with very low public sector R&D funding and no major public R&D institutions shows that British industry can succeed on its own. And the ability of the company to turn foreign investment into domestic wealth and economic growth is to be celebrated.
But there should be a burning question in the world of British AI and innovation policy thinkers about whether Sewtec could have innovated even quicker, grown even faster, and been the acquiring partner and not the acquired partner in different circumstances.
Are our national institutions trying so hard to stimulate and back new startups in places and sectors that London, Oxford, and Cambridge care about that they continue to ignore, as they always have, companies like Sewtec, in sectors that Sewtec leads in, in places like where Sewtec are?
As I read the Lovelace Institute report the next day, I wondered how such an institute might have changed the story of an innovative British company grown out of the textile industry into the future of manufacturing. Which also got me thinking about textiles.
The myths we tell ourselves.
Among American examples I know less about, the Lovelace Institute paper picks a British model for success. It chooses the MRC Laboratory of Molecular Biology in Cambridge. The LMB.
This is the perfect example to choose to set me up to share my concerns about the proposal, my concerns about the direction of British science and innovation thinking more generally (as highlighted by the dinner), and my proposal for how we could try and do better this time.
I couldn’t have picked any better myself and I thank the authors for their fortuitous decision.
The section on the LMB starts “The genesis of the institute can be traced to a bold decision by the MRC to fund an unorthodox unit of researchers led by Max Perutz” and “The LMB’s founders believed that biology could become a fundamental science only by rooting itself in the physical sciences: physics, chemistry and crystallography – which transformed biology from a largely descriptive field into a mechanistic one. This laid the foundations for molecular biology, biotechnology and much of modern medicine.”
I though it safest for my stress levels at this point to continue reading in the pub, with a beer.
This is the South East English telling of the story. It is not completely wrong, but it is also a long way from right, and the gap matters enormously to any attempt to use innovation policy to build a truly national purpose.
Let me tell the Northern version of the story, the true foundation for molecular biology, and the unorthodox unit of researchers led by William Astbury who were snubbed.
In 1915, Lawrence and William Bragg at the University of Leeds shared the Nobel Prize in Physics for their work inventing the X-ray spectrometer and pioneering its use to analyse crystal structures. Leeds was the place to do X-ray spectrometry.
In 1928 William Astbury, born in Stoke-on-Trent, was appointed Lecturer in Textile Physics at the University of Leeds. He began using X-ray spectrometry to study the structure of wool. He worked on wool largely because his work was heavily funded by the local wool industry who had been founders and funders of the university and its predecessors and whose interests were now substantially represented by The Wool Industries Research Association founded in 1918.
Of particular interest was why wool stretched and shrunk so differently in the heat and wet to cotton. From these investigations Astbury discovered the alpha and beta secondary structures of proteins, what his close correspondent and indirect collaborator Pauling would call alpha-helices and beta-sheets in his later Nobel Prize winning work.
Astbury’s work was varied and unorthodox. He later added DNA to his list of molecules of interest. He and his graduate student Florence Bell were the first, in 1938, to publish X-ray diffraction patterns of the crystalised molecule, and the first to accurately measure the repeating patterns and distances within its structure.
In 1945 Astbury popularised the term Molecular Biology and saw that his ambitions were greater than the local wool industry via the Wool Industries Research Association could justify or afford to support. So he went to the Medical Research Council in London to try and make his lab, the lab described by Perutz who would later found the LMB as “the X-ray Vatican”, the national centre of molecular biology.
The MRC had four senior decision makers. They all lived and worked in London. Three had gone to University in Cambridge, and one in Oxford. The board as a whole was ten people, seven living in London, one each in Cambridge, Oxford, and Newcastle. Five were graduates of Cambridge, two of Oxford, two of London, and one of Durham.
The MRC board did not choose to fund the lab of “The Father of Molecular Biology” William Astbury. It was too tainted by association with industry, he was too Northern, his work too unorthodox, and his lab too far away.
The MRC funded the places and the people they knew.
Perutz was funded in Cambridge to build what would become the LMB.
Randall was funded at Kings College in London to invest in the lab where in 1952 Rosalind Franklin would use newer equipment to take much better pictures of DNA than Bell had managed in Leeds in 1938.
Leeds got nothing.
Astbury was said, and the best reporting on this whole topic is in Kersten Hall’s fantastic biography of the man, to be dejected. He could not hire the mathematicians he needed to build the better models of protein and DNA structure that would win Pauling, Crick, and Watson their later Nobel Prizes. He could not buy the new equipment that Franklin would use to take better pictures of DNA and other molecules. He could not afford to move far enough beyond the constraints of his foundational industrial funding to win the Nobel Prize he had in him.
This attitude has always existed in England. I have written at length about the efforts and success of Oxford and Cambridge in London to block places like Leeds from having universities at all. When Northern cities later established universities, substantially on the back of industrial funding, people with similar attitudes used that necessary association with industry to deny them funding or influence.
It remains amusing to me how convenient it is that so many of the innovation experts in South England continue to wonder why we as a country struggle to translate our world-leading pure science research into industrial application given the clear history of many of their institutions in discouraging the industrial applications of research.
The Wool Industries Research Association continued its work in Leeds, both with the university and at their own laboratories. Martyn and Synge won the Nobel Prize in Chemistry in 1952 for their invention in 1941 of partition chromatography which they used on the amino acids that make up the keratin in wool. Dr. Martindale invented what is now the ISO standard for measuring the resistance of a textile to abrasion. Others at WIRA created methods and set standards for rapidly measuring fibre thickness and strength and thus helped in the development of the Huddersfield Super Series cloths that remain the finest available in English suit making.
But that output, though world-leading, was a fraction of the LMB’s.
WIRA in Leeds was closed under Margaret Thatcher. Her reforms removed its sources of funding during a push to reduce the UK’s public support for research close to industry in places like Leeds and focus instead on purer research such as at the LMB in Cambridge. The site of WIRA is now a block of flats next to a big and beautiful Wetherspoons pub just up the hill from my house.
The importance of geography.
It was fortunate for me that the Lovelace Institute paper chose to focus on the LMB. The story I have just recounted on its origins are either not known or are retold very differently in Cambridge and London. In the Astbury Centre at the University of Leeds, where I earned my PhD, it is quietly instilled among those of us who will listen, as are the lessons from it. We are told the stories, but then advised not to bring them up — and certainly not the equivalent contemporary stories — if we want a decent chance at shaping the UK’s research agenda in directions that would benefit our careers. Compliance in university academia is a deep weakness, as the paper argues well.
The attitudes that saw Astbury’s application to lead the UK’s national efforts in molecular biology passed over despite being almost certainly the strongest proposal persist to this day. Panels in and around London are excellent at giving the UK’s public research money to researchers in and around London and finding ways to convince themselves they were funding the best science and not just the science that looked most like what the people they know do.
Paul Nurse, the President of the Royal Society, infamously said that only London could host national labs like his Crick Institute.
He was just saying what I find a worryingly high proportion of scientific thinkers in the South East think.
The MRC’s current board are 6 based in London, 3 Cambridge, 1 Edinburgh, and 1 Liverpool. Barely changed in geographical composition from 1945.
ARIA’s board and advisors show a more modern pattern of including Americans while seemingly doing little to bring in opinions and ideas from outside of South East England.
And I know that the underlying attitudes and biases are as present as ever. Which is why Britain continues to fund too much public R&D in places like London and Edinburgh. And too little in places like Birmingham, Liverpool, Manchester, Belfast, and Glasgow.
For nearly a decade I helped Paul and Kathryn to run and make The Open Data Institute Leeds a success. With no national government funding we achieved more in many ways than The Open Data Institute in London — who we paid for a franchise licence — did with £2m per year. In the absence of central government support we succeeded based on support from industry, by creating spin outs that made money, and by winning competitive tenders for work. And for that we were judged negatively as less pure and less worthy than our competition and forced to give up our name. We are now Open Innovations. UK law forbids the use of the word institute without permission from the central government in London.
Recently, I tried to start a conversation to capitalise on Leeds’ commercial success in data industries and encourage the UK government set up its National Data Library in our city, rather than in the South East by default. It has been made clear to me that Leeds just isn’t seen by the government in London as the right place, not least by the fact that the only time anyone from the UK government funding the project have made a meeting likely was in London, and even then at the invite of a third party.
The UK government recently appointed a Director of the National Data Library. Their LinkedIn location is London. I have never met them.
How to make Lovelace Institutes a success.
None of this reduces my belief in the need for something like Lovelace Institutes, it merely focuses my view on what they must be if they are to succeed.
There is a gap for publicly-funded research in Britain outside of the career structures and incentives of universities. Catapult Centres and ARIA have plugged some of this gap, but not all of it.
The paper argues convincingly against a hub and spoke model for such an organisation. I strongly agree. Physical proximity is everything for the type of work that a Lovelace Institute would do. Lovelace Institutes must physically exist, with an outstanding online presence focused on communications, not doing work.
Resources must not be spread too thin, with just a few Lovelace Institutes at most. There might only need to be one or two.
The location and form of Lovelace Institutes must be chosen to maximise agglomeration benefits and to be resistant to capture by the status quo of the science and innovation community. This means they must be large and visibly standalone laboratories, not part of existing universities. They must be in a large city, at a distance from the scientific status quo, at the centre of an agglomeration of economic strength, scientific excellence, and talent.
At the moment in Britain, only three places meet this brief, and only one without reservation.
Oxford and Cambridge are far too small, too dominated by their existing universities, and too unwilling and unable to grow to meet the huge associated business growth we should expect as a spillover from 3% to 4% of the UK’s public R&D budget.
London obviously has sufficient scale, but a Lovelace Institute there is far too likely to be captured by the status quo and revert to the existing academic culture of Britain which is overwhelmingly defined and concentrated in the capital.
Going down the list of remaining candidates,
Glasgow achieves the necessary scale but only through proximity to Edinburgh which brings with it a risk of reversion to existing academic culture.
Birmingham has scale but only just makes the cut for proximity to scientific excellence. A big risk is that with the opening of HS2 Birmingham will be just as close to London as Swindon is. We have seen that Swindon’s proximity to London has almost completely removed any independence of thought that the research councils moving there might have brought. A Lovelace Institute in Birmingham may well be swallowed by the scientific status quo.
Manchester is the clear leader.
The UK’s home of transformative innovation and industrial application. The world’s original modern city. Connected to the world by the country’s only undisputed hub airport outside of London. The centre of an agglomeration including the UK’s highest intensity private sector R&D hub in Cheshire, the UK’s defence and nuclear industries, and Britain’s most successful tech business. A city with a global brand that when combined with its neighbour in Liverpool is the equal of London’s and the first place in Britain to win a Nobel Prize. And with nearly 13 million people living within 50 miles of Manchester, a similar sized labour pool to Paris.
Due to the accumulated biases of the UK’s honours system, well under 10% of the top awards go to Northerners, Manchester is an excellent place to avoid the prestige-based recruitment the Lovelace Institute paper warns against too.
The paper proposes that the next step to setting up Lovelace Institutes would be that “The UK government should allocate a £3 million initial grant for a three-year founding unit”. I agree, but only if that founding unit is based in Manchester. Anywhere else is a recipe for the same failures that disconnect our innovation system from our economy and leave us as the weakest economy in Northern Europe.
If the goal is to build a sense of national purpose using innovation funding, then that funding must be purposefully national. It must escape from the parochial innovation funding system that the UK has today. The nation must be sure that it will pick the best research to fund, not just the people it knows in the places it lives.
I suspect my condition would doom the idea to failure, but I would prefer that failure to another “national” institute in South East England run by the usual suspects spending my money drawing out success from my region in return for handouts.