Sebastian Thrun, CEO Udacity, talks about the growth of the online education company at the Udacity office in Mountain View, California, Friday, Feb. 5 2016. Thor Swift for the Financial Times
“We are going to break the laws of physics in some funny way”, says Sebastian Thrun, on relentless Silicon Valley innovation © FT

Scrap internal meetings. Do away with budgets. Put an end to the idea of continuous employment.

When it comes to the practice of management, Sebastian Thrun likes to present himself as a radical outsider. These are among the things he tried to ditch while running Google’s X lab, the Petri dish of the future, where the internet company nurtured projects including driverless cars.

But at Udacity, the Silicon Valley education company he co-founded and now heads, Mr Thrun always seems to be running between meetings — a way, he says, to constantly fight the inertia that creeps into corporate life. And, like managers at most companies, he is keen to cite initiatives that have broken through normal decision-making processes to show that speed and innovation are possible.

The big enemy is complacency. “It is a bit like fingernails. It kind of grows back. You cut them off and they grow back.”

Glassdoor, a website where workers anonymously rate employers, provides clues about how this style feels to those lower down the food chain: Mr Thrun’s personal approval rating is only 68 per cent, very close to the average for all chief executives on the site. But he insists that his 200 Udacity staff have welcomed policies like “Project Liberation”, his way of nudging workers out of the door when they are no longer a good fit.

Fighting the average is what seems to motivate Mr Thrun most. Udacity grew out of what he says was a bigger idea even than Google X: remaking education so that it fits what he calls the “geekster economy”, in which job tenure is short and rapid learning of new skills essential.

As one of the instigators of the “Mooc” craze — the idea that free online lectures, the massive online open courses, would spread higher learning to the masses — he has since refined the idea into more structured courses that involve a higher degree of human interaction. The result is a new qualification he has named the nanodegree, something Udacity says takes about six months of part-time study. Udacity has operations in the US and India and expects to launch in China soon along with South America and Europe. Udacity currently reaches 100,000 online although, Mr Thrun adds, with characteristic verve and ambition: “We often talk about [reaching] 100m people to have a real impact.”

Mooc Tracker

See our directory of free online business courses​ ​on entrepreneurship, management and more

Mr Thrun’s most obvious management tool is a powerful sense of mission. Silicon Valley’s brand of change-the-world idealism has become easy to mock, but he says it is what drove the thinking behind the “moonshot” projects at X: “I know that sounds funny, but we believe they will have a large impact on society.”

He also wields his scientist’s distance as a weapon. Most managers, even if they dimly perceive disruptive changes coming, are too invested in the present to follow an idea to its logical conclusion. Or, even then, to do much about it. “What we are often good at is looking at the past and seeing those signs, we are often really bad at extrapolating.”

Over the long term, developments once unthinkable can become reality: if aeroplanes and mobile phones seemed almost magical 150 years ago, he says, why won’t the next advances be just as astonishing? “So we are going to break the laws of physics in some funny way. Maybe we live forever. It’s quite feasible for us to live much, much longer.”

When it comes to considering whether far-reaching and disruptive forces will upend higher education, Mr Thrun becomes uncharacteristically hesitant. It could be because he faced what he calls “direct animosity against me as a person” — news of the nanodegree’s invention was greeted with derision by some educationalists. He still enjoys many of the benefits of being an honorary professor at Stanford, although he gave up tenure after leaving a career as a star computer science professor there.

Whatever the reason, when it comes to discussing the current education system, he ties himself in knots. Like many, he argues that the level of student debt in the US is unsustainable: “It’s a time-bomb that is ticking.” But he says that ideas such as Udacity, despite being designed to be far more cost-­effective and responsive to a much broader group of people, will not replace universities.

Even his fiercest criticism is careful: “In higher education, we’ve been def­ending and defending and defending a model that is good in many ways — and wonderful in many ways — but also by virtue of the fact that it’s 100 years old isn’t necessarily the best.”

At his own company, meanwhile, he says he is bent on stamping out the normal pathology of corporate behaviour that leads to mediocrity.

Mr Thrun warns that meetings instil groupthink: “The more people com­municating, the worse it is, to be honest, contrary to conventional wisdom. By driving this consensus-driven culture, we massively slow down innovation.” Meetings also produce a fear of failure, suppressing radical ideas, he says. And the more people that are added to the process, the slower and more cautious it becomes.

Managers, keen to justify their existence, add their own unnecessary ideas. The result, he says, is that “something that is pristine and nice and simple becomes complicated: it never becomes easier and simpler”.

At this point, with a project slowing down, “people put more of their heart into it, and their ego, and their identity into it”. And that, in turn, makes it hard to apply the cycle of rapid iteration that is a hallmark of Silicon Valley innovation. The “human reaction”, he says, is “to find 1,000 different ways to explain why it’s actually not a failure, it’s just not good enough yet”.

The Udacity boss holds up two initiatives — one top-down, the other bottom-up — that he claims show it is possible to break out of this cycle. One involved his own decision earlier this year to announce that Udacity should refund students’ tuition fees if their studies fail to result in a job.

The other was last year’s launch of a network of self-employed markers to grade student assignments, similar to those used by exam boards. The idea came from the company’s head engineer and Mr Thrun says he turned a blind eye to the project.

“Had I done the classical thing — make a product review and a budget and staffing plan and PowerPoints, you would have found that the staffing was not available, there are more important priorities,” he says.

Lessons from the X lab

As the first head of Google’s X lab, Mr Thrun says he deliberately tried to isolate teams working on “moonshot” projects. Other Googlers were barred from the building.

He also avoided giving project teams budgets, instead setting milestones and assuring them they would have the resources. “I have learned, if you give a team a budget, then the team tries to maximise the budget so that they get the same next year.”

Small teams, he says, stand a much better chance of hitting even big milestones, because they move faster. When Google’s driverless car team reached the 1,000 miles of autonomous driving mark, it had only 12 engineers, compared with the 100-strong groups he says other carmakers used. A team of fewer than 10 on Project Loon came up with the longest balloon to stay airborne.

Another deliberate decision was to build temporary teams, rather than create jobs. “I told everybody, expect to leave Google X and the company when you are done.”

People in research can remain enamoured for far too long with ideas that were cutting edge when they were young.

“That filled me with pain,” he says. “I literally worked at research labs where the staff really tried to steer management away from the modern technology that was actually better.”

Copyright The Financial Times Limited 2024. All rights reserved.
Reuse this content (opens in new window) CommentsJump to comments section

Follow the topics in this article

Comments