Interview with Mike Lynch OBE

Mike Lynch was born in Essex in 1965, his father was a fireman and his mother was a nurse. His father regretted not having had the opportunity to go to university, and so he valued education very highly. His parents were prepared to think a little bit differently and were very encouraging. At the weekends, they would go off and visit a place of interest.

His father heard that a very good, local fee-paying school, had a scholarship programme and Mike got in through that.  The school was founded by Francis Bancroft, who became Lord Mayor of London and who, when he died in the 1600s left some money to educate poor boys. Mike feels There were two reasons why it was a good school. First, almost everyone in his year wasn’t paying fees, they were either paid for by Francis Bancroft, or, got in through the Eleven Plus.  So, he grew up with a normal cross-section of society in an ‘amazing school’.

Secondly he says the teachers were incredible and after school they would do things that were way beyond the curriculum. Most relevant was that the maths teacher managed to get an old teletype printer connected to the university at Queen Mary College, London, with an old, acoustic modem and acoustic coupler. So they started writing programs long before anyone had ever heard about this.  Initially they had to do it on punch cards, which he says was very annoying because if they made one mistake, it came back the next day and wouldn’t run. But then they got a teletype and then, they built a Compukit UK101, which was a sort of precursor of the BBC Micro and this was all happening before it was obvious that computers were interesting. He was there until 1983 and studied maths, further maths, physics and chemistry.

Mike then won a place at Christ’s College, Cambridge and studied there for three years from 1983. Mike loved his time there as he was being taught by people who were at the cutting-edge and doing the most advanced research, and was surrounded by amazing students, who were often multi-talented. At Cambridge he met Peter Rayner, who had basically become an apprentice in signal processing and worked his way up to become one of the world experts in that field. Mike was very interested in his class because he was desperate to build synthesisers and the big transition then was to use digital methods. He found Peter to be an absolutely wonderful mentor that let him get on with interesting things. Mike completed a PhD in adaptive techniques in connectionist models, which is what would now be called machine learning.

When Mike was at school he was a musician, and wanted a digital sampler, which was out of his reach as they cost around £100,000.  So, he set out to build one, which, without realising it, made him one of the few people in the world doing things with digital signal processing chips.  After designing his own digital sampler someone heard about it and bought the design.

“Designing synthesisers was incredible training, because it’s all real time, it’s all assembler code, multiple processors,” he says. That was his first foray into commercialisation of technology. In 1991, Mike formed a new company, Neurodynamics which was about fingerprint recognition.

He realised that very similar mathematics to that he was already using could solve the problem of fingerprint matching. In those days, most fingerprint matching was done manually, which meant 30 or 40 highly-trained fingerprint experts had to go through the files for about three weeks to see if they could find a match. Mike helped build a machine, again on digital signal processors, that would do the matching in about five minutes.

They demonstrated it to an interested policeman and he loved it: “And the fact that we were two people and not IBM didn’t really matter, because he tested the machine, it worked and he bought it, and it changed policing for him.”  The company consisted of Mike and a ‘chap’ who he had employed to work with him.

In 1996 , in response to a request from the police, Neurodynamics developed something called the DRE, the Dynamic Reasoning Engine, which then became Autonomy.  They realised that it was applicable to much more than it’s initial purpose because in those days there was very little ability to handle textual information in a computer, and most of the world’s information is textual.

Autonomy became the UK’s largest software business, and was sold to Hewlett-Packard for $11bn, moving from start-up to being the second largest software company in Europe after SAP.

In 2012 Mike started Invoke Capital based on the principle that young entrepreneurs should be concentrating on the clever bit, not running the salesforce or the customer support desk as Invoke does all that. They provide that and the money which given his track record was easier to raise than for the people they back.  Their approach is very simple, they know how to make lots of mistakes and would like their entrepreneurs to make new and exciting mistakes.  It’s still early days but their success rate is good.  Darktrace, which they funded as a start-up at a valuation of a few million, is worth over $500m now, and that’s in just two years.

They have two rules in Invoke: The first thing is it has to be fundamental technology, because they like to have a big advantage.  And then, because of all the things they have already done, they only do things that they think can be very big.

Mike thinks that two companies really turned around the ability of the UK to compete in tech and they are ARM and Autonomy.  They created, in ARM’s case about $24bn out of the UK and Autonomy got to $11bn, when there had been no previous start-up to major players out of the UK.  They created an ecosystem where those things are now happening and there is a very vibrant set of technology businesses and a lot of the people that are doing great things were trained in Autonomy or in ARM at some point.

Mike feels that sometimes it was a mistake to listen to industry analysts. He remembers when Gartner declared that the internet ad model was dead, and they had a technology at the time that was more advanced than Google, and pulled the plug because we didn’t see a business model for it. And of course, he says, Gartner were wrong and the rest is history.  So, he doesn’t tend to listen to them as much.

Mike thinks the challenge initially in the UK was having no start-up scene, and so it was considered a very questionable thing to do, and because of that, there was no finance.  So he feels it’s great that that has changed. Another major challenge is collecting good people and Mike feels the trick there is to realise what you don’t want is ten copies of yourself: you want people that are very different to you.

Businesses are all about the talent, and so, look after the good people: don’t put up with the bad people.  And those are the two sides of the same coin.

Mike thinks learning principles about things is better than learning things.  There’s no point in learning about things that are going to change.  Also, to really understand what is technology and what is the application of technology.  Wanting to be in technology, is different to being in something that’s enabled by technology.  Another thing he thinks is very important is understanding that the technologies which are appearing now mean everything can be rewritten. “Have an open mind about how things could be done differently.”

Mike thinks that we’re in a period of incredible change.  That recent breakthroughs in AI are going to change a lot of things.  He thinks that personalised medicine where the treatments are down to an individual, especially in areas like cancer, are going to be key.

Interviewed by: Richard Sharpe on the 31st January 2017 in London

Transcribed by: Susan Hutton

Abstracted by: Annabel Davies