Nathaniel Barling: “Objectivity is an impossible pursuit”


Nathaniel Barling, who graduated from Durham in 2016, recently secured a place on Forbes’ 30 under 30 media list, alongside his fellow co-founders of Knowhere – an AI-driven platform designed to strive towards objectivity above all else. It was an idea conjured up by Barling as an 18-year-old and nurtured during his time at Durham; A philosophy text on his first year reading list led to the name ‘Knowhere’, while the Bill Bryson library was the scene of its legal incorporation. Palatinate spoke to Barling about the early days of flying between Durham and Stanford, their exponential success in the coronavirus pandemic, and the inevitable impossibility of their goal.

The origin of an idea

Barling first conceived the idea while on his gap year, and it began to materialise after reading The View from Nowhere, by Professor Thomas Nagel, during his first year studies. “The argument that he presents in the book,” Barling explains, “is around the impossibility of objectivity. But his suggestion that we can leverage subjective lenses on the world to get closer to objectivity struck me as a powerful way to be thinking about the role that AI (artificial intelligence) could play. That’s always stuck with me.”

Nagel’s philosophical treatise ultimately inspired the name of the company. “What we’re trying to do is use technology to find a ‘view’ that is as objective as possible. It was a foundational idea to the technology – I just decided to stick a ‘K’ on the front for a little play on words.”

Throughout first year Barling sat on his idea. “It was during rugby preseason week in second year, at the Billy B, when I originally incorporated the company in the UK.” Barling began to work on the project without any particular sense of urgency. “My plan was always to spend five or ten years in finance and then one day build this company that I’m drooling about building.”

But, following a series of personal tragedies, including the death of fellow Durham student and close friend Euan Coulthard in 2015, Barling decided there was no point in waiting. “It pushed me to think, ‘I’m going to get on with the thing that really feels like I should be doing.’ And that was what kicked me to get going. Between then and the end of second year, I was trying to pull together a business plan. I went and spoke with a whole bunch of my professional mentors and ex-colleagues to see what their take was on what I wanted to build – a sort of AI driven company. I got a positive reception and I managed to get some commitments from people to back the business.”

I was spending about three months of that year on Stanford campus rather than Durham campus

It was that summer, between second and third year, that Barling met the first of his fellow co-founders, Alexandre Elkrief. “We were doing an internship at Deutsche Bank in London. Alex had just finished his undergrad at Imperial and we’d both been working on AI driven projects. He was literally the first guy I spoke to on the first day at the internship and we hit it off. By the end of that week we’d agreed to build the company together; And here we are still, six years later.”

By the end of the summer, the pair had managed to pull together significant backing from their network in Europe. Elkrief, however, had plans to go to Stanford to do a graduate programme in computer science. “I moved out to spend six weeks with him,” Barling recalls. “Before he’d even started at Stanford we were both crashing in his apartment on campus.”

Soon afterwards, they met their third co-founder, Dylan Rhodes, who was also completing a computer science course at Stanford. “But I had a whole year still to go at Durham. I flew back at the end of Freshers Week and proceeded to spend the next year trying to get the company off the ground in California, get myself a Visa, and raise money so we could get kickstarted.”

The AI in our company is basically the foundation to everything we do

Why did the community of Stanford appeal to Barling as a base for the early days of the company? “I always envisioned it as a global company that would emanate from the States. During my research in trying to pull together a business plan, one of the things that had stood out was Stanford and specifically the NLP (Natural Language Processing) and AI research happening there. So, when I met Alex and learned he was going out there for grad school, I showed him all of the plans I’d put together and told him that I really wanted to leverage the research community at Stanford – the best in the world from a university standpoint. It was a completely natural thing.”

After graduation, Barling continued to split his time between the UK and the US. “We raised some money in early 2017 in order to build out beyond the initial prototype. That was the start of the real deal because we could really start hiring employees.”

“That final year was pretty intense,” Barling adds. “I was co-president of the Finance Society and also trying to finish my bloody degree ­– and doing that while spending about three months of that year on Stanford campus rather than Durham campus was a little bit nuts, and did frankly stretch me to the very limit. I wouldn’t have done it any other way but it was a tough year.”

The pursuit of objectivity

Within Knowhere, staff rely on a perpetual groundwork laid by intricately designed AI machinery. Barling describes how their use of technology has informed and enabled their ultimate goal of objectivity. “The best way to think about the AI in our company is that it’s basically the foundation to everything we do. Everything else sits on top of our infrastructure, which is really about discovering and organising information as efficiently and accurately as possible, that might be interesting to a reader or important for them to know about their local community.”

“We pull together all of that information into our systems and do a whole host of analysis around it, where we try to predict how reliable that information might be and how newsworthy it is. We also try and look at the partisan nature of the information – is there a lean to the way in which this is reported? Are we getting this from official republican or democratic sources? Is this coming directly from an affiliated elected representative? Is this coming from a group that has financial backing on one side of the political spectrum?”

No algorithm or human is less biased than the amount of information they consume

Once all the analysis has been poured over by AI technology, hired reporters and editors use the data to produce daily stories.

Barling explains how working from an AI base is actually a very sustainable way for the business to function. “Our algorithms are at the point where the ‘newsworthiness’ analysis and the headline generation is so good that we can deliver newsletters each morning to communities just on that basis alone. The goal is to at some stage self-finance the reporting of each community. At about 5000 subscribers on any one of those newsletters, we’re going to be covering the salary of its reporter. We’re now able to much more aggressively expand into new markets and build from the ground up in order to hire new reporters.”

There is a careful balance between human and machine within Knowhere. As they expand and gain access to more data, they will become progressively more able to eliminate biases in their reporting. “We believe that the more sources we gather, the better we’re able to differentiate what information is likely to be biased. Every additional source we’re able to pull in is giving us an additional lens on the world – I guess one way of framing it would be hearing as many voices as we possibly can. We’re giving ourselves more tools to be able to differentiate between information.”

We’re not perfect now and we never will be perfect

Diversity, as one of their central values, is another aspect that enables the company to combat inherent biases. “We look to bring in a very diverse set of reporters as possible ­– in terms of background, reporting background, and cultural experience – to counterbalance each individual’s particular biases. And also, we don’t think that this just applies to the reporting piece. It is just as important that we have that emphasis on our engineering and product team as well, so that we accurately reflect the communities we’re covering.”

Barling, however, acknowledges the ultimate impossibility of their task. “We’re not perfect now and we never will be perfect. No algorithm or human is less biased than the amount of information that they consume. Our goal is to help our reporters get as far away from their own individual biases as we possibly can. In the company values, we talk about how objectivity is an impossible pursuit. We have to accept that and nonetheless work towards it.”

Knowhere to go but up

Along the timeline of this pursuit sit the unprecedented events of last year. Barling describes the effects of the coronavirus pandemic on the company. “It sent our growth through the roof in March of last year, when shut downs started rising and the gravity of the pandemic started to hit. In the space of three or four months we went from 80,000 subscribers to around 400,000 – on a national level. There’s been a lot of interesting things that we’ve had to figure out over the last nine months. We’re lucky in that, compared to a lot of businesses, we have a very immediate feedback group in our readers.”

For Barling and his co-founders, the Forbes nomination came as an unexpected, but welcome, surprise in the form of an email. “My understanding of it – because it came out of the blue – is that multiple people have to nominate you for you be potentially recognised. I think what happened on our end is that some of our investors, partners, and maybe even readers put us up. We don’t know – they didn’t share that with us! I would have been curious to know for sure.”

I always envisioned it as a global company

Forbes sent them a list of questions about the company, alongside some more personal ones: “What do you think is the long run impact of the pandemic on the US economy? Has it made you more bullish and bearish? What’s been your greatest challenge so far? Things along those lines. There was an interesting one about how student debt played into the process of starting our company, if it did ­– which it did for me, because it actually prompted me to get moving faster rather than wait around. My co-founders and I submitted some answers to those questions and then we found out at the same time as everybody else, when the list came out in early December.”

Barling himself first heard the over an excitable phone call from his dad. He reflects on the achievement. “We’re still very early on in the journey. We’re 20 people who are trying to build the world’s most impactful organisation and do it very differently to how it’s been done before – so we’ve got a hell of a long way to go. But we’re obviously super grateful for the marker to recognise what we’ve done so far.”

Visit their Forbes profile here.

Image: Nathaniel Barling

Leave a Reply

Your email address will not be published.


This site uses Akismet to reduce spam. Learn how your comment data is processed.