Interview with Richard Hopkins

Richard Hopkins was born in 1969 in Newcastle upon Tyne. He is an only child, he says of the experience: “During my upbringing it meant the home was a little bit quiet. From the age of about seven onwards, I used to try and build robots just using what was around.  So I had a lot of relatively solitary hobbies, but I shared those solitary hobbies with a whole bunch of friends who had very similar ones.”

Richard’s grandfather and father were engineers and Richard considers himself a software engineer. His son has also followed on with the tradition having just completed a degree in mechanical engineering. Richard says of his own experience: “I’ve been involved in all different kinds of aspects of the IT industry, because IBM is involved in all kinds of different areas. The thing that has kept me most engaged has been the development of large-scale complex IT systems for IBM’s clients – that’s always a fusion of software and hardware – but the creative element tends to be on the software side of things and sometimes in research, and I have done some things that have involved, innovative stuff on the physical side as well, but mostly software.”

Richard says he was the generation that grew up with ZX81s and Spectrums, adding: “When I saw the ZX80, I just went, I’ve got to have a ZX80, Dad!  At the time, my school had one computer which was a Commodore PET.  Computer club lessons consisted entirely of writing code in BASIC in an exercise book, so that you could try and get the program to work in your head, because there was only one computer. You’d then have to take it in turns typing in to see whether you could actually debug it and get it working or not.  It was very frustrating, and not particularly a great experience, so the idea of having your own computer was extraordinary.” Richard did not get the ZX80, but he did get the ZX81 as a Christmas gift.

Richard attended his local state primary school (The Links in Eaglescliffe). He says: “I loved it, I absolutely adored it.  It encouraged students across the board whether you were sporty, arty, science-y, it was equally enthusiastic about all elements, which suited me down to the ground. It encouraged the interplay between arts and science as well at that stage, which I didn’t realise was so unusual until I entered secondary schools and beyond, and woe betide me when I went to university.  I still think that was a better educational environment than anywhere I’ve been to since.”

“I resist very strongly this idea that there are different fields of subject and study that you must follow if you are a scientist or an artist. I believe in following both pretty equally as much as I can, because I genuinely think the thought processes you need to go through to come up with new ideas and new ways of doing things, or even to do simple things like empathise with an end user, requires you to develop both sides of your brain and the creativity is bringing those elements together. I’ve always been frustrated and resisted quite strongly this hard line between the arts and the sciences.”

Richard’s secondary schooling was at a school formed by a group of local parents. Richard says: “If there’s such a thing as a start-up public school, or private school, this was it. There was a very well-established private school in the area for the ICI employees to send their girls to, but when my parents looked for the equivalent for boys, there wasn’t one.  At that stage my father was very adamant that there should be the ability to have a good science education background, and although the schools around us are exceptional schools, and they continue to be, in those days both the major state schools that were in the catchment area were both very arts-oriented and there was no science stronghold.

“So they decided to start their own school. There was a core group of about five parents, and I think my parents formed the next ring out of another core of about ten or 15 parents, who essentially started a school.  They found a headmaster, found a site and literally started a school.  I was still at primary school when I found myself painting the chemistry lab walls of the old grammar school that we were taking over.

“Yarm School is now a massively huge successful school that’s probably got over a thousand pupils in it, it is now probably one of the best schools in the north-east.  It wasn’t when I was there, we were building it bit by bit, or wall by wall. It’s sad that it’s changed its character quite a lot since then, but it was very eccentric.  Partially military, partially adventuring, partially driven by the minds of the most eccentric of the teachers.  So it had a bit of everything really.  It was a very good secondary school from that point of view because it was quite malleable in terms of what you were able to get away with.”

Having gained A grades in English, Maths, Physics and General Studies, along with S level English. Richard applied for and was accepted at Durham University.

 He says: “There was the basic assumption this English thing was just a fad that I’d grow out of and I’d go to Oxford and do Physics – at least that was what my parents and the Headmaster assumed.  As it turned out it was going to be quite different.”

 Richard had hoped to read Physics and English, however the system did not allow for it. He studied English and decided to self-teach the subject he loved. He says: “I virtually abandoned the formal element of the English course and started my own education.  The first year of University’s fairly rigid and structured, you have to do certain things, but after that it all gets a bit easier, so I structured my own learning, did my own things, and did a lot of self-teaching.”

Looking at the way universities operate today where the system is a lot more rigid, he says: “I think we’ve rowed even further back from the idea of it being an institution of learning to being a grades factory in the same way that schools have become grades factories, which I don’t agree with, because fundamentally I think we disenfranchise a lot of bright people through these very rigid, structured ways of doing things.”

Having had a ZX81 as a present, when it came to going to university, Richard turned down the idea of an IBM PC which his father offered and had an Amstrad word processor instead. When he got to university, he started to experiment with the machine in CP/M mode, he explains: “I had stuff published in magazines because I managed to convert some of my Spectrum games across onto the word processor, which was tricky to do, but they all worked.” He also programmed it to write poetry and submitted two poems as part of his course work.   He says: “One was based on Heart of Darkness, I can’t remember what the other one was based on, but it was based on quotes from those two things.  So, every time you typed in a different quote it would give you a different poem.  … I submitted these two poems to one of the creative criticism courses I was doing possibly forgetting to mention they were written by a computer, and they came back with A minus and B plus from the Professor of English at Durham University.  The next day I submitted the program along with an explanation of how it worked to the professor so that he was aware that he’d just given out an A minus and a B plus to a computer written bit of poetry.”

Richard also got into chaos theory at university.  He says: “I then started trying to create my own strange attractors on my Amstrad word processor.  I was let down by the resolution of the screen on the word processor, so I wrote a little program that would generate bits of strange attractors, and I could take the individual files then merge them together in a desktop publishing application and print off these amazing high definition graphs, that led me to IBM.”

Richard entered an internship competition called Elite for IBM while at university. He says: “The idea was that you were a computer science graduate and you would do something with the local firms and then IBM would give a cash prize out to the people that came top.”

Although, Richard wasn’t a computer science graduate, he shared some of his programming with them and was accepted to the scheme. He adds: “I went on this scheme in my holidays, I computerised the accounting of a stainless steel fabricators, I had a great time. Little did I know that by doing that, I was able to produce a list of all the people that owed them money on a daily basis, and I took it on as part of my job to phone those people up every morning to get the money in, I thought while I’m doing this I may as well do it properly. The guy absolutely adored what I was doing. It was the first time they’d ever had real time accounts or knew what their cash position was, and by calling round every morning I turned around their cashflow.  I didn’t know, but they were in a cashflow crisis before I arrived, by the time I left they were fine.” The owner was so happy with Richard’s work, he asked him to return the next holiday to work on their stock control. Richard says: “I went back and computerised the stock control.  I did a little bit of programming in terms of shell scripts and this kind of thing, to make the software automatically boot and give them easy access to the word processors, I created a lot of menu system things.  I did no real programming, whereas everybody else on the Elite course took it as an instruction to go off and program something.  IBM looked at what I’d done and were surprised that I had turned around the cashflow of an organisation around, so they gave me the prize, much to the chagrin of all the computer science graduates,  who thought they’d done much more impressive things, which they had. They’d written more imaginative code than I had, that’s for certain.”

As a result, Richard was offered a job by IBM for when he finished his Degree, but having already got a job lined up with the BBC, he initially turned it down. However when he started to look at what the two organisations would offer in the near future, he found a very different outlook. Richard found that he would not get to do the BBC Director course he was aiming towards until he’d worked for them for eight years, while IBM said they didn’t know what he would be doing in three years. He adds: “They said I could be doing anything, and they painted a very different picture of an organisation.” This changed Richard’s mind and after attending an IBM selection day, and completing an aptitude test, he was offered a role in Newcastle and started in 1990.

He says of the culture of the time: “it was interestingly regimented.  The day I joined was the day they released the ES/9021, the 3090, this was the next generation of mainframes.  Everybody in the branch had to go into a conference room, including the secretaries, they literally left one person answering the phone, and everybody got briefed for a whole day on the intricacies of the hardware, software, the new features of the IBM mainframe.  Everybody, irrespective of what their role was or what, secretaries to systems engineers to customer engineers, to people like me who’d just joined, sitting there going, what the hell’s a mainframe?

“It was an interesting culture, but at the same time there was the recognition that things had to change, and change quickly, because a year after I joined, we registered the largest ever single quarter corporate loss in history.  I don’t think anyone’s ever broken that record since.  Other companies have lost more money over a year than we did since then, but in one quarter we registered the biggest ever single quarter loss and I believe that’s an unbroken record even now.

“So, we were in an existential crisis almost from the point at which I joined IBM. I was part of that new guard, I was part of the new way of doing things. I was sat next to somebody who’d been an IBM-er for 20 years, unbelievably his job was to look after, from a technical viewpoint, the customers of Lotus 1-2-3/m, which is a spreadsheet running on a mainframe, which is the worst possible experience in the world.  They sat me next to him so that some of my enthusiasm would rub off on him.

“They needed to reinvigorate the workforce and the idea of reinvention of the individual had not really taken root in IBM at that point.  That was the point at which it became obvious that people needed to be able to reinvent themselves during their careers on a continual basis, because probably half the people in the branch were not that busy. There was over a hundred people in the Newcastle branch, we had customers all over the place, Nissan and all the shipbuilders and ICI.  We had a huge customer base.  I was part of that new guard, I was PC and UNIX, so no one knew what I was talking about.  It was a time of radical change.”

In 1993, Lou Gerstner took over as Chairman and CEO of IBM replacing John Akers.

Richard says of him: “He was and remains brilliant. He did reverse virtually every decision that his predecessors were making and completely turned the company around.  It was amazing to behold.  He came in, looked around and went what makes you unique is you doing the whole thing, is you being IBM, is being end-to-end, I can’t see that anywhere else in the market, why on earth would you change that? That is what resulted in all the e-business stuff and the growth of our services area and all that things that resulted in a huge regrowth of the company.

“It’s hard to explain just how much of a cultural difference he brought in such a relatively short space of time, every CEO since then has massively benefited from what he did.”

Asked about the training culture of IBM, Richard says: “It was exceptional and occasionally mind-blowing.  A lot of it was driven by the people on the ground, there wasn’t as much centralisation, there wasn’t so much learning by watching videos, there was an expectation that learning was experiential, generally speaking, and was taught by practitioners. I just found that wonderful.  …. To be able to absorb that much knowledge in a relatively short space of time through experiential learning via an expert, just can’t be beaten.  The other element of it, which I found extraordinarily powerful, was that whether you were technical or sales, you all went through this process of doing sales, essentially cold sales calls.  But if you were technical you had to combine those cold calls of going in to see somebody you’d never met before to solve a technical problem.”

Richard says that this forced those who were technical to become very good communicators and listeners.

Richard became an IBM IT architect a role in which he remained for much of the rest his career. One of his early projects was IBM’s Post Place solution for counter automation in the Post Office. The Post Place solution was a competitor to the Horizon computer system which was eventually selected to win the contract and which has since been at the centre of the British Post Office scandal.

 Richard worked on the £2 billion contract bid to build a system for both the Post Office and DWP’s payment systems. He says: “It was a huge procurement. Eventually it got whittled down to just the Post Office bit. It was a very interesting procurement.  I learnt several things in that procurement.  First thing is that is that the fundamentals of engineering are really important.”

He explains that Post Place was ultimately sold to the US Post Office which is approximately the same size as the British version.  He adds: “The USPS system has worked fine for many years, has not got people into trouble, has not dropped transactions because it is transactional. In any system where you’ve got 20,000 locations and 40,000 terminals over those locations and they’re doing hundreds to thousands of transactions a day per terminal, a one in a million chance of something going wrong because you haven’t engineered the thing right becomes a daily certainty, not a blue moon thing. Until you get that into your heads, as an engineer for IT systems, you are completely lost because otherwise you try and apply the learnings you’ve got from building a small system to a big system and that doesn’t work.  You have to fundamentally re-engineer systems to work at that kind of scale so the one in a million chance does not happen.”

He says of the procurement process: “The procurement was a very long procurement.  It was actually technically pretty rigorous, but there was a huge degree of liability being transferred to the suppliers in those days, very unusual, unlimited liabilities, which of course never materialised but that’s what the contract said.  We were doing quite well in the evaluations from a technical perspective, especially on the Post Office side of things, and then at the very last minute our headquarters saw the terms and conditions and said ‘we’re not signing up to that’.”

By this point, Richard had been working 100-hour weeks on the bid.  He says: “So, I learned lots of things on that one, never work that hard again, never assume that you’re in control, and ensure that you’re engineering correctly at scale.”

He continues to explain the importance of ethical decision making as an architect or software engineer, he adds: “I learnt this long before the thing became a scandal, because I did lots of work on welfare systems. You learn that every single decision you make as an architect, or every line of code you write as a developer has an ethical implication and you have to think those through. You have to consider how people cope if they can’t get access to that? How do you prevent people being digitally left behind? All of that has to be part of the landscape of the IT professional.  We’re meant to be chartered engineers, ultimately, we need to start behaving as if we were chartered engineers.”

Asked if Y2K was a scam or a necessity, and although he did not work on any Y2K fixes because he was engineering new systems at the time, Richard says: “I still find it bizarre that because a disaster didn’t really happen it’s therefore regarded as a scam, whereas I regard it as being a really good example of how the computer industry can pull its socks up when it needs to.”

Asked about the perception of public sector IT projects and why they tend to fail, Richard says: “The general problem used to be the distance from the user and  distance from the policymaker. The requirements are coming from the policymaker, and another set of requirements are meant to come from the user, and then of course there’s a third pillar which is the existing complexity of the estate, the brownfield elements.  So those three things combined are almost guaranteed to be a train wreck if, first of all, you don’t engage with the real user and you accept the fact that you’re going to work with a bunch of proxy users who are analysts from the government department who pretend to know what they’re talking about, but often have never even met a real user. This has disappeared over recent years because of Agile, which has been incredibly useful for reducing that gap. We now do Agile and User Centered Design practices in government that make a huge difference on that side. We still haven’t fixed the policy one. So those are the fundamental reasons I think why we’ve had problems.”

Richard compares it to the private sector and says that he still prefers working in the public sector, adding: “When you get that combination right, it is the best place to deliver, because you know you’re working with the people who are trying to do the right thing on one side and the people who you’re going to impact on the other and they’re all within talking distance or walking distance, almost. It makes a huge difference.

“I’d still like to see more radical things going on in public sector in terms of really thinking through whether the structure of departments fit the modern world, are there unnecessary boundaries in place and this kind of thing or do things need to get reshaped to make them flow better.  But, those things are big philosophical things about how government should be organised.”

Fifteen years after joining IBM, Richard became the Chief Technology Officer for the public sector and in 2011, he was appointed as a Distinguished Engineer.

Asked to comment on a recent report published in Nature by academics, that suggests that the application of quantum computing will affect GDP because it will take some time for organisations, public and private, to compensate and invest in it, Richard says: “If that’s what happens with any significant technology, then maybe it will happen.

“However, on this occasion the UK government have said they’re going to invest £2.4 billion over the next decade in quantum and create the conditions in which the UK can generate some money off the back of quantum technologies.  That might mitigate the impact somewhat.

“But I would concur with the academics that the area that the government and all the areas around government are really lackadaisical, is the industry adoption and exploitation of the technology.  I am seeing nowhere near enough focus on communicating into those areas that there’s something they’re going to need to adapt to and absorb.

“The mega-corporations already know about this stuff and are making plans and doing things.  I don’t think the large firms are to much degree and I’m sure the ones below them definitely aren’t.  And it will impact everybody, it will be available to everybody, because it’s going to be cloud-based.  They need to start asking themselves how am I going to bridge that skills gap? Do I need to work with a small software company that’s going to give me a quantum optimisation algorithm? How am I going to use it to optimise my cashflow or my stock control?

“Absorbing a new technology’s always tricky and the smaller you are, the trickier it is. I’m not seeing enough focus on building that ecosystem of enablers around the industries that matter to the country.  So, there could be this backward step on GDP if we don’t get that bit right and I’m sure other countries are pushing harder and faster.  The US certainly is.  The US has already issued edicts about quantum safe cryptography.  We’re still waiting on that from the UK.”

Talking about the progress that is being made in quantum computing, Richard says: “I’ve been involved in quantum computing for five or six years.  It’s quite long, it feels like longer, but fundamentally every time you look, things have moved forward faster than you expect.

“For example, within the last month or so IBM’s announced in Nature that we’ve done some very precise calculations using a noisy intermediate upscale quantum computer, using all 127 qubits to model the spins of atoms. We’ve been able to prove that the quantum computer, for a certain set of conditions that are well known and well computable, like the Clifford conditions states, that we got the answer exactly right.  For the next set out we used a supercomputer and one of the universities verified that they were able to calculate on the supercomputer and their approximation methods were in agreement with the precise answers that we were able to calculate.  Then there’s a whole series of answers that no one can verify because there’s no computer big enough on the planet to be able to verify them.

“So we’ve issued a challenge to the academic community to come up with an approximate way or a supercomputer way that’s going to be able to verify whether we got the answers right or not, but we think we did because we’ve proven that the other ones are right.  So if that’s the case, then we’ve already shown that for a scientific problem we can use noisy intermediate scale quantum computers using error mitigation techniques – not error correction – error mitigation techniques to come up with a precise answer.”

Richard says that at the moment it is not cost effective to do quantum computing for normal day to day problems but “a day’s worth of quantum computing time might be affordable for certain use cases around optimisations or problems that simply can’t be computed any other way”.

He continues: “Although we’re not claiming we’ve got to that quantum advantage point where it is cost effective to do a typical optimisation for a commercial use case on a quantum computer, because it currently takes a day and that’s probably going to cost you a lot of money, an entire day on a quantum computer, especially one with 127 qubits.  So there’s a certain element of engineering efficiency that we have to solve.  But the fundamental problem that everybody thought was going to require logical qubits, which would require anywhere between 250 and a thousand qubits, noisy qubits to create one logical qubit, that is no longer a blocker.

Looking at the next steps, he says: “If you assume you need logical qubits, that’s a long way off, but if you actually are happy with a long run time, some cost and using error mitigation techniques, then I think we’re kind of in the ballpark already.  We’ve got a roadmap that extends out to low numbers of tens of thousands of qubits already, and that’s within the next few years.”

The IBM roadmap is available on the IBM website under IBM Quantum and by searching  ‘Quantum roadmap’.”

Richard highlights that Google has also shown that using similar technology to IBM’s that “they can create logical qubits which are where the actual error correction is allowing for improved longevity of the logical qubit itself.  So the bits of the puzzle are all slotting into place and rather than it requiring a huge leap or a huge step change in capability to get to the point where it’s commercially viable, what there now appears to be is a relatively smooth path of increasing levels of noise mitigation, increasingly large QPUs, increasingly large bridged together QPUs, increasingly large fridges, and finally the miniaturisation of the control software and control hardware to actually fit within the fridge itself.”

Richard goes on to say there are still engineering problems to overcome. He also highlights new research that is showing the potential of technology such as neutral atoms which he says: “or other ones that as they develop may turn out to be better technologies.  So what we can say is that it looks like there’s a commercial way forward for the next decade or so, and after that who knows?  I mean if enough money goes into some of these new, very promising looking qubit technologies, including room temperature quantum computers, who knows where that’ll take us.  But the point is that by the time we get there we’ll have some really good algorithms, some really good idea how to mitigate noise, and possibly even some logical qubits. It’s not a black and white path,  I think it’s now a series of graduated shades of grey.”

IBM’s research on the subject is being done across locations worldwide. As well as IBM and Google, Richard also mentions Rigetti who are researching the area, he adds: “It’s not just us, Google are doing similar work, Rigetti’s doing similar work and there are other qubit technologies all going on in parallel.  There’s obviously an increasing amount of trade secrets and everything else, but, there’s still an awful lot of publishing going on which is actually bringing the entire field forward together.  Especially all the work around algorithms and applications and this kind of thing, pretty much equally applicable to all the technologies.”

Richard says that AI will probably be the biggest applications of quantum computing, adding: “Initially, I think it will be machine learning in terms of classification algorithms and this kind of thing.  So we’ve already got some good results from real data on a quantum computer around things like fraud prediction and prevention, because it looks like it’ll change the scope and scale of those algorithms significantly.”

Asked if he would you have wished to sign the letter asking to pause GPT development, Richard says: “No, I think it’s the wrong thing to do, because I don’t think we learn our lessons correctly if we do it that way.” He highlights that he talks to people about quantum ethics, adding: “That whole debate and that whole structuring needs to start now, and should have been happening sooner. IBM’s been talking about AI ethics forever. We’ve been pushing this for over a decade since Jeopardy!”

He highlights that the idea of industry adopting GPT in an irresponsible way is not what he’s seen customers doing.   He says: “What I’ve actually seen is most of the customers saying ‘it’s not very reliable is it, it hallucinates, it does all these strange things, I can’t use that, it doesn’t explain its decisions, I can’t use that’.  What I’ve actually seen to that pushback, is that technology grows again.”

 He adds: “People are clued up enough to go, no, no, no, this only works for a series of key use cases. It will be used in an awful lot of places, but I think what we’re going to see is much more focussed use and much more controlled use and safeguards and guardrails being put in place to ensure that it does the right thing in the right context, and therefore not quite job done, because I think that all needed to be thought about and said before the technology was released to the public, because that’s what caused the problem, it suddenly became publicly available.  Those of us who were on the Beta had been using it for a year already, had already worked out its shortcomings and had engineered how to get round them, or where not to use it and when to use it and this kind of thing.”

He concludes: “Once the technology’s out the bag it’s very hard to put it back in the bag.  Do the thinking before it gets out the bag in the first place.”

Interviewed by Richard Sharpe

Transcribed by Susan Nicholls

Abstracted by Lynda Feeley