The technology-biology interface with Bill Liao
In this interview, Bill Liao general partner at SOSV, Synthace board member and co-founder of social network Xing sits down with the Synthace team to chat coding, biology and starting companies at the technology-biology interface.
(Synthace) To begin, should every biologist learn to code?
DNA is digital. At a fundamental level, DNA is the life equivalent of machine code, and what was difficult in the computer industry when I started was that you had to programme in machine code. This was a laborious, complex, and just a really boring task. To generalise, most biologists would find that kind of digital skill challenging. And so, it isn’t ‘should or shouldn’t biologists code’, it should be that there’s a sufficiently-accessible abstraction layer. This is to ensure that the automation and the coding is not so painful as it has been, and therefore, it can have a higher degree of efficacy. The thing about Antha (Synthace’s cloud operating system for biology) that’s miraculous, is that the abstraction layer is so easy, that you can, without any prior coding experience, dive right in and make something happen, as easily as you might edit a document in Word, or a spreadsheet in Excel. It’s undoubtedly going to cause a seismic shift in the world of biology.
So why did you think there’s been a bit of lag in digitisation within biology?
In my experience, most biologists have traditionally eschewed an engineering degree to go into biology, because they’re into the idea of life. But as biology has progressed, they found themselves having to deal with evermore complex mathematics, and evermore complex computer software. If you got into the field because you hated all that stuff, you would naturally be resistant to its burgeoning growth around you. I think biology has by virtue of an artificial and arbitrary separation of the sciences, fallen foul of that. If we hadn’t specialised sciences so much, we would be in a much better position in biology.
When I started in the computer industry, my computer was very disappointing, because there was no internet, it wasn’t connected to anything. I could run a programme, but it was a little bit like spinning plates, because I could create a game, and then, I could play the game that I created, which was kind of pointless, because I knew exactly all the mechanics of what I just built, and I couldn’t share it with anyone.
Now, if you write a few short lines of code, millions of people can be using it in seconds. What does this mean? It means that everything beyond human scale is now inter-mediated with computers, and effectively done so that we get far better results. If you want to look at the very large, the biggest image in the world was created with the Hubble Space Telescope. Guess what – all driven by software – it doesn’t work without it. You want to look at the smallest things? The Large Hadron Collider, high energy collisions with the smallest particles, and again, every element mediated by code. You want to look at the very slow, it’s only when we did extreme time-lapse photography of forests that we could see trees actually touching and communicating – exhibiting behaviours that we normally associate with mammals. If you want to look at the very fast, look at the advances in slow motion, anything fast, we need to look at that through computer-mediated interactions. And life itself, in every aspect, can be seen through the lens of computers in a way that individual humans can’t even begin to approach.
What role does the wider community have in training that next generation of digitally competent scientists? Is it up to the University system? What about all the highly skilled scientists currently in industry? What can they do?
Coding is a language skill, and languages are best learned young. So, my own initiative (CoderDojo) actually is a club for kids to go and learn coding for free. We take them in the ages 7 to 17. So for the hard-core coders, I’m not sure it’s a school’s job. I think it’s actually better done outside of schools, because frankly, learning languages is easier done in interaction with other humans, rather than being taught. When it comes to the older biologists, the answer for them has to be a high level of abstraction. You have just got to make it super easy, visual, comfortable, and it has to look organic. I think one of the most brilliant things about Antha is that it does procedural language with an organic look and feel, that I know biologists love.
Moving to a different topic. Do you see there being a culture clash between biotechnology and technology, both in the companies themselves and their respective investor bases?
I think that calling it a clash is downing the conflict with faint praise. There is almost a wilful ignorance in the tech sector about biotech, and a wilful ignorance in biology, about technology. I think that investors have over-specialised. The ideal case is where you get automation, and technology enhancing, and speeding up the biology to the point where a biotech start-up is delivering as much value as a tech start-up, as quickly, for the same amount of cost. And not only is that possible, it’s inevitable.
And so, the conflict is in the growing pains of that, and those that identify this trend and get on board are going to be the winners. You look back at the overused example of Steve Jobs. He clearly saw the trend, and he got on board early. Some would say actually too early. Really wasn’t until his second coming that he really nailed it, but nevertheless, if he had never gotten on board the train, the most valuable company in the world wouldn’t exist. The most valuable company in the world is not going to remain Apple, unless Apple starts getting into biology.
What’s your advice to entrepreneurs looking to build a company that works across technology and biology?
So when we look at team formation with companies that are multidisciplinary, I always advise the companies to spend time in each other’s shoes. I always advise founders to actually skill up enough to actually do each other’s jobs. Partly because they’re gonna have to pitch in on things, when things get difficult, and partly so they develop a core understanding of just how hard each piece is. Naturally, there are key specialisations, and again, in the cross-skilling, you can enable parts of the specialisation that previously you hadn’t before, and I do think people get over-specialised. Especially in early stage, you need a well-rounded team, and you need each of the team members to be willing to do any part of the job, at any time.
Do you think that big-pharma is intent on true digital adoption or is it just hype?
Most of them don’t even know what languages machine learning is written in! I think that when you’re beholden to a lot of shareholders, you get to come out with the latest buzz words, and you’d be foolish not to at least recognise that. But there are some companies making inroads, including the ones that Synthace is working with, because people have gone, “You know what? AI or not, automation is definitely the way to go.” And flexible automation that biologists can understand is even more so. Most biologists don’t need an advanced AI solution yet though, they just need a flexible automated pipette.