The end, when it inevitably came, arrived much sooner than any of us – even David – had expected. After his diagnosis with terminal stomach cancer in July 2015, David MacKay, with characteristic mathematical clarity, sketched out his likely life expectancy on a whiteboard graph. As you can see in the image, he gave himself a 50:50 chance of surviving until January 2017, while “>2 years” was “v. unlikely”.
Probabilistic thinking came easily to David, as the many aficionados of his landmark textbook Information Theory, Inference and Learning Algorithms who are currently deluging Twitter with moving and heartfelt condolence messages will attest. However, sadly for David, for his family, for all of us, and indeed for the world, his remaining lifetime was very much at the left side of the S curve, and he died on the afternoon of 14 April 2016, just a week short of what would have been his 49th birthday.
David’s early death is desperately sad. David was not sentimental, but I simply can’t think of any other word. His children are so young – Torrin is 4 and Eriska only 1 – and now their father will not get to see them grow up, and his wonderful wife Ramesh is left alone. It seems so bitterly unfair. David wrote about all of this in his sparklingly honest and witty blog ‘Everything is Connected‘, and his post ‘What do you tell the children?‘, written on 13 September last year, dealt unsparingly – and humourously – with that uniquely awful dilemma. He knew he was running out of time by early April, and posted a farewell message of sorts: ‘Perhaps my last post – we’ll see‘, with this opening:
I noticed that the posts of a friend who died of cancer trickled away to a non-conclusion, and this seems an inevitable difficulty, that the final post won’t ever get writ. I’d like my posts to have an ending, so I’m going to make this my final one – maybe. While the doctors haven’t expressed an opinion, I think it’s possible I haven’t got long to go…
It wasn’t quite the last one, but subsequent posts – including a poignant but hilarious one about the hospital’s lack of “intelligent thermal environmental control” (which I really hope Addenbrookes and other hospital authorities will read and act on) are labelled Appendices “to my completed cancer story”. In the end, David got it just right, with his final Appendix 3 blogpost being a JustGiving site to raise money – in his memory, it now turns out – for the Arthur Rank Hospice Charity (please do donate – it is incredible how his fundraising target has already been quickly surpassed). I got a final text from him, just a day before he died, saying with characteristic straightforwardness “no visit thanks”.
David MacKay had more personal and professional integrity than anyone I have ever known – and yet somehow he managed to combine it with a warmth that underlay everything he did. (I was privileged to attend his celebratory Symposium in Cambridge just a month ago – I don’t think I have ever been in a scientific meeting with so much love in the room.) He wore his super-intelligence – people use the word ‘genius’ rarely these days, but I’ve heard it used for David a number of times – lightly, and always interacted with humility and an enduring sense of fun.
David had a strong moral compass and sense of justice – his work was fundamentally driven by a desire to make a difference, and to help solve real problems, even intractably huge problems like climate change. His massive contribution was bring numeracy to a debate obscured by mudslinging and ideologically-motivated rhetoric (both of which I’m as guilty of as anyone). It was characteristic of this desire to see real change that he accepted the immense challenge of taking on the role of Chief Scientific Advisor at the UK government’s Department of Energy and Climate Change, rather than staying in the ‘ivory tower’ of Cambridge, after the enormous success of his epochal book Sustainable Energy, without the hot air.
As you’ll see from the above lecture (perhaps my favourite, given in Oxford in 2014) David was most of all a quite simply brilliant teacher. It was not just that he was a gifted communicator – perhaps a rare thing for an expert on information theory and machine learning – but that he understood the need to help other people understand. This could take a lot of him time, which he always gave generously: on one occasion, on a Skype call, he spent an hour sketching graphs on a handy piece of cardboard in a doomed effort to impress on me the basics of Bayesian statistics (we were talking about p values being misused in scientific papers to inappropriately infer statistical significance). I still don’t understand Bayesian inference, but I trusted David to be right on that and a lot else besides more than I’ve ever trusted anyone before or will again.
That’s the problem with wise teachers like David MacKay. Normal people like me just want to be given the answer by someone super-smart that we trust, and thereby be told what to think. But David wouldn’t be drawn into handing out easy answers, whether as a university professor, a writer, a speaker or a government advisor. His whole effort was focused on giving people the tools and understanding so they could figure out the answers for themselves – even government ministers. And that is his legacy. Though the world has lost an incredible brain, and is the poorer for it, David left us everything we need to figure out for ourselves how to proceed with solving climate change – and other problems, however huge and complex they may appear. His formula is disarmingly simple: we have to get the numbers right, and think things through rationally from first principles. To borrow David’s best-known quote, we don’t have to be anti-this or pro-that: we just have to be “pro-arithmetic” and the rest will follow. I hope that we can do this, both to cherish David’s memory and because such an approach is after all the only way to get to the right answer.