JON OBERLANDER (1962 – 2017)

It is hard to believe Jon Oberlander has been gone for more than a year. The news of his death reached me late, in March. I wrote the below and never published it because, frankly, I am not sure it does him justice. I publish it now because Jon should be remembered.

Jon Oberlander was a “yes, and…” kind of man — full of ideas, connections, tips — delivered always with remarkable energy, generosity, and style.

With his fascination at what might be around the corner and his sci-fi haircut, he sometimes struck me as a messenger from the future. At Edinburgh, his willingness to bet on new ideas and his support for innovators helped to foster an atmosphere of experimentation and risk-taking, making him beloved of the university’s entrepreneurs. More importantly, he believed in the potential of technology to improve people’s lives, and blended informatics with psychology and the humanities in his research and projects. In the time that I knew him, he was everywhere, dashing into the classroom with a lecture hot off the press and delivering ideas faster than we could write, in perfectly formed paragraphs, in a melodious baritone; flying off to conferences; introducing creators at InSpace; walking up Clerk St at a brisk clip, raincoat trailing in the wind, smiling. He was optimistic, a tad mischievous, an idealist even.

For all the showmanship, he was also calm, warm, contemplative, understated, stoic, and reserved. Upon entering his office, a visitor would see, taped to his door, some version of Sayre’s law: “The reason academic politics are so vicious is because the stakes are so low.” With this for a north star, he managed to be respected and liked, loved even, by students and colleagues alike, a feat all the more remarkable given the various leadership roles he took on. He spoke ill of no one. But he showered praise at every opportunity. Years on, I remember him vividly, framed by the Salisbury Crags beyond his window, holding up the dissertation of my friend, Sia, asking whether I knew the author of this excellent piece of work. He also rarely spoke of himself. For a long time, I knew nothing of his life beyond the intellectual heritage revealed by his webpage. Once, in a rare moment, with a hint of paternal pride, he told me that his daughter had just graduated university. And he never complained. If all the demands weighed on him, you never knew it.

I met Jon in early 2010, in a course on natural language generation that he was co-teaching with Johanna Moore. I had come to Edinburgh for an MSc in speach and language processing, and was determined to do a self-proposed dissertation around something with commercial potential. Towards the end of 2009, I became convinced that people would begin meeting strangers through GPS-enabled smartphones, and that machines would match them based on interests and personality (one of Jon’s areas of expertise). Fresh from Victor Lavrenko’s information retrieval course, Text Technologies, I suggested modeling humans as walking search queries and documents, and matching them using TF-IDF. To my delight, Jon, together with Johanna, agreed to supervise this fanciful proposal. Jon then found funding for my experiments, advised generously throughout, and made sure I finished on time. And after I graduated, he continued to help me with the startup that came of it. A year and a half later, the company struggling and me depressed, I bumped into him at the Cameo. He asked how things were going. I don’t recall what I said. But I remember his compassion.

As fate might have it, I last saw Jon in early December 2017, two weeks before he died, on my first trip back to Edinburgh in six years. I had come, once more, seeking his advice and involvement. It was a last minute trip, hastily arranged, but he made time. We spoke about my latest venture, and for thirty minutes flat, he was a firehose, jumping out of his seat to sketch on the whiteboard, pulling up papers and slides, suggesting people and ideas. When he met me, he was coming from one engagement, and when we parted, he was rushing to another. As we walked across the Informatics Forum, he pointed out the new center for data science that the university was building and which he was to lead. I left exhilerated, delighted to find him as dynamic as ever, and grateful. I believed we would collaborate again.

The news of his death reached me last week and hit me like a thunderbolt. I, too, will miss him terribly. He was a good and brilliant man, a true teacher, an inspiring human. Everyone who met him was elevated by him. He walked alongside you, dreaming big.

Meknes to Rabat

The train was late, the A/C broken, the window screwed shut, the compartment an oven. A few stops before Rabat two young women got on, one in a tight skirt, hair flowing, made up. Hello, coast: goodbye, interior. We stepped off the train into the ocean breeze.

Meknes to Rabat, Porto after Madrid — by the ocean, one breathes freely.

Or is it…in capitals one breathes freedom…?


[From a comment on Fred Wilson’s blog, lightly edited:]

Sometimes I envy VCs for the chance they get to refine their judgment about people. To invest or not? Hit or miss? What a neat quadrant!

I have always trusted my instincts about individuals — being unafraid to trust where trust felt due, lending a bike to a near stranger, hitching a ride with a complete stranger, etc — and the few times I’ve been burned, it was usually because love was involved.

That said, I’ve been bothered by another thought, namely, that I may be setting the bar too high and missing a lot of good people or opportunities. In other words, to borrow from information retrieval, that my false negative or miss rate is actually quite high, unbeknownst to me. Maybe I am too conservative. Maybe my judgment isn’t so great after all.

Unlike VCs, I don’t get to see this as explicitly, in the form of competitors making a fortune off my passes. But then again, who knows whether the investments they pass on would have been as successful with them? Chemistry matters.

In the context of this post [about keeping or firing people], I suppose the question is more: to hire or not to hire / to fire or not to fire.

Perhaps VCs and employers can refine this process / their judgment by tracking such decisions (if they aren’t doing so already).