Media Lab Conversations Talk Summary: Reid Hoffman

Graduate student Matt Stempeck's transcription of Reid Hoffman's conversation with Joi Ito on April 4, 2012.

Joi introduces Reid Hoffman as a public intellectual, adult supervision at Paypal, founding member of "The Paypal Mafia", Joi's life coach, early investor in Facebook, Zynga, and Flickr, Founder of LinkedIn, partner at Greylock, author of The Startup of You, and stealth member of the Media Lab's unannounced Advisory Board.

Reid Hoffman: It's not that everyone should start a company, it's the fact that a career ladder is no longer a strong model for how you do your work and pursue your career. The Good grades -> Good university -> Good career path model has been broken for years, by globalization and technology's disruption of industry. The model for how to think about your life, career, and work is different. How entrepreneurs think about product market fit, product differentiation, creative risks, all apply to how you, as an individual, live your life.

Be the entrepreneur of your own life:
What are the pragmatic concepts that are both mindsets and tools that enable you?

Joi: It's like having a Media Arts and Sciences degree [from the Media Lab]. It's tricky. Everyone here is really different. It's very different from an engineering school where you know what you're going to do. No company is planning to hire 50 Media Arts and Sciences graduates.

Iteration is in the DNA of the Media Lab. What you're bringing into this is the Silicon Valley entrepreneur perspective. How would you give advice to our students?

Hoffman: Product and market fit requires you to figure out the earliest tells. How do you bring in as much networked intelligence into that process as possible? In Silicon Valley, you bring in early advisors, employees, customers. What your'e trying to figure out is, is the path I'm trying to build the company around accurate? Most people begin with the financing process as a series of hoops to get a certain amount of money in the bank. But the most interesting thing of the financing process is getting network intelligence on the critical question of, "Is this a good plan?" What is the piece of common intelligence about my project? What are the risks in their investment? With LinkedIn, there's zero value for the first few users.

Joi: Someone's described the Media Lab as a bunch of answers looking for problems. There was the sensor for Yo Yo Ma's bow that turned into a sensor for child safety seats. There's a lot of serendipity, and people ask, "What's the difference between your serendipity and just randomness and chaos?"

Hoffman: We're moving from an information age to a network age. Part of that is, how do you increase the possibility of positive outcome from serendipity? There's still luck, but you can increase the probability of the right decisions made. When you have a problem connecting challenges and solutions, that generally involves connections in a human network.

Joi, citing a paper on luck: Serendipity and focus are at odds. An experiment put a dot on the screen and showed peripheral things, and people saw them. But the moment you provide a financial incentive for seeing the dot, people no longer see the peripheral items. In another study, the people who consider themselves lucky were more likely to see pattern-breaking information hidden in a monotonous photo-tagging task.

Hoffman: Entrepreneurs are often given two pieces of contradictory advice: persistence and flexibility. Have a vision and pursue it through years of people telling you you're out of your mind. Or, be flexible: look at data, iterate, and change based on the signals you're getting. There isn't an actual algorithm. You have an investment thesis about why this project is likely to work and have some outside result, and usually that's expressed in a set of statements and hypotheses, that if you're right about, adds up like a logical proof and gives you the output you're looking for. And you can have varying level of confidence in how these pieces are adding up and supporting your theses. Paypal considered helping people split up restaurant bills using their Palm Pilots, but they thought through how few people that would reach, and came upon the far more successful product of email-based payments.

One of the chapters in the book is a flexible planning framework. People picture planning as a train on the tracks, with a few switch gates. No plans work like that. I describe it as Plan A and Plan B, but it's not cascading, where one fails and you move to an entirely different plan. It's having parameters within which you can shift. Plan Z is an intelligent risk you can take to reset when your confidence if Plan A and Plan B are waning. It's a hard reset you can do in a way that allows you to continue to play.

Joi: We talk about celebrating reinvention and failure. Fail fast...

Hoffman: That phrase...I would celebrate learning. Your Plan A should be sufficiently disruptive, a new way of doing something that sufficiently disrupts the ecosystem. You remove $10 from an ecosystem and replace it with $1, introducing a product that works on a different order of magnitude. In the book, we pulled examples from companies. Flickr was started as a game. What they found was that the actual engagement around the photo features was getting far more use than the rest of the game. Pivots based on reading the data can be successful reinvention, and not just following a failure. Plan A should be reinvention to begin with, and Plan B is resetting Plan A if it doesn't work, but along the same parameters.

Joi: Nicholas [Negroponte] always talks about incrementalism being our enemy. Companies pay us money to think of the things they wouldn't come up with in their planning rooms. At a faculty retreat, we came up with three words:
Magic - that's kind of the thing that engages people -- Andy started describing a project with, "Remember that Star Trek project where..." That's how good Media Lab projects start.

A lot of companies are looking for incremental things - sharper pens and an improved bottom line. Companies, like animals, don't intuitively want to find out that they're wrong, but they also know that they need to find out if they are. How do you instill the disruption of DNA to make it comfortable, and how do you convince people to put their money behind it?

Hoffman: There are existing companies and new companies. For existing companies, there's a whole process, and the reinvention process usually requires some person with the mandate to make that happen, and there usually isn't that person. At a start-up, or early-stage project, the only really massive early-stage projects are where you're contrarian and right. The projects where everyone agrees often end up having less overall success than the projects where there was some disagreement amongst the board. Where's the contrarian thinking that, if they turn out to be right, could be really, really big?

Sometimes it's as simple as a new technology or broader adoption of a new technology. You want to start building a company 1-4 years, maybe 5 years, in advance. So that when the technology converges with the change in the world you see coming, you're positioned to capitalize. You have to be right about a set of things, including what your competitors are going to do.

Joi, again channeling Nicholas: Peer review is the death knell of disruption, because it means everyone agrees. We say we're anti-disciplinary here. It maps to your point that consensus can be a bad sign.

Hoffman: Consensus indicates it's probably not a total break-out project. If your thinking isn't truly contrarian, there's a dog pile of competitors thinking the same thing, and that will limit your total success.

Joi: So I can now say that the disagreements at the Media Lab are on purpose. A bunch of Sloan students did a study on the Media Lab, and the first page said, "Everyone thinks it's something different. No one agrees on what it is."

Many internet companies were created, and the product idea tested, without requiring a ton of money to get going. My theory is that Silicon Valley had a regional advantage of critical mass of investors and startups, but for that reason Silicon Valley hasn't been as focused on things like biotech and hardware. The Media Lab and Boston more generally has done more provisional tinkering in these areas, with synthetic biology and fab labs. Now, as prototyping gets cheaper and the cost of creation and eventually distribution go down, you see hardware startups coming out. I think the same lowering cost of innovation is going to happen in biotech and drug discovery and drug testing. You'll see the same dynamics of startups taking over from labs in these areas. If we get our act together on the investing side, can we develop a regional advantage? There's one really low-hanging piece of juicy fruit that Silicon Valley is focused on. But most of the world's companies don't fit this profile. Do we have a shot at capitalizing on this?

Hoffman: There's a huge amount of tension in Silicon Valley, which I have joked should now be called Software Valley. There are additional channels for innovation, some of which may be available for Boston and the Media Lab to jump on and own. One is that as Moore's Law begins to touch other areas, besides just IT, like 3D printing, and moving from bits to atoms, involve huge patterns of companies and products that can be created. And there's network density and an advantage that can be gained in an area. If a [regional] network becomes so good at launching that kind of business that people move there to do it, you know you've reached that stage. The network density creates a feedback loop and can lead to breakout hits.

Joi: We've talked about doing a post-graduate startup program. There's talk of a Minimum Viable Product, and Reid's famous for saying, "If you've released it and you're not embarrassed by it, you've released too late." [Hoffman smiles] So, in that spirit, we're going to officially announce that we're doing it, even though it doesn't have a name yet. The initial members are Google, Greylock Partners, Matrix, and the Media Lab.

Hoffman: So many entrepreneurs are worried about protecting their precious ideas, but the truly valuable thing is that you're in motion, you have momentum, you're gathering all the necessary resources to actually make it happen.

Chris Peterson: Crowdfunding's been in the news lately, with a bill going through Congress. (

Joi: They wrote the bill with a fundamental misunderstanding of what's going on with KickStarter.

Hoffman: The cost of making a Minimum Viable Product is so substantially reduced that a small amount of money can make a try, and a set of people, through crowdfunding, can then allow that innovation to spread out through the networks. The benefit of capitalism is distributed capital and business models, the same thing's true with crowdfunding. I don't know about the geographical advantage it brings, but it does create massive amounts more innovation if [innovation] is not gated by heavy peer review where lots of people need to agree for something to happen. The ability of people to build and even deploy products at low cost allows for more innovation.

I do think there are several parts of the legislation which are critical -- you should actually allow more shareholders before it's public. I don't know if the legislation is right or not.

Joi: Kickstarter's about funding neat things -- it gets really funky when you expect a financial return. I can see Wikipedia being funded on Kickstarter, but not at a university or from a company. Benjamin Mako Hill studied that even those who bet on the Wikipedia model, and there were many, didn't all succeed.

Hoffman: There's a bit of a meme in the Valley now of people running with their ideas rather than go around protecting them. How do you leave room for innovation without fear of voracious copying or being eaten by massive companies?

Question from @hatchsteph: Talk more about how the Media Lab can work on corporate ideas without losing non-disciplinary innovation.

Hoffman: There need to be labs doing early development and base technologies and that's really beneficial to the entire ecosystem, so the question is how you arrange incentives so that happens. This is one of the reasons I was willing to join the Media Lab's advisory board. I think one of the key things is to figure out how to help the folks who want to start companies from here get the best possible intelligence and connections here to reach the highest water mark possible.

Joi: It goes back to impacting people, whether it's 1 billion or 7 billion. If we can knock a few out of the park, we'll be fine, and we won't have companies coming to us for sharper pens. We have 350 projects, 26 groups, but I still get the question a lot: What have you done for us lately? And that's always going to come up, and there's a process we're going to go through to keep the environment open and free -- no pressure anyone -- but we need to hit one out of the park. Just like VC firms like Greylock - you just need one home run every 5 or 10 years.

Hoffman: The top investment is worth the total amount of all the other projects and more. You're looking for the one high water mark, not the average. And then people don't come to you looking for singles.

Joi: We have groups that will see ups and downs, but when you see something tracking, we need to provide unconditional support for those things that are posed to go crazy, and be aware.

Hoffman: Uniqueness doesn't need to be, "Oh, I've never seen that before." It should be construed broadly.
Impact: the scale of impact is a three conditional vector: number of people, depth of impact, and over time. Something that lasts over a long period of time can have incredible impact.
And Magic: This word emphasizes the contrarian-ness. It doesn't mean everyone should look at you like you're insane, but something that is a gambled risk on a thesis, where, if you're right, you can have a discontinuous impact.

Q: A couple of years ago I was at an event with a VC at the Media Lab and he was getting really irritated. Serendipity and focus are opposed to each other and you can't really bridge them.

Joi: Serendipity's hard. It's like meditation. You don't meditate by falling asleep. It takes a certain amount of focus. When you're a good hunter, with a bow and arrow, there's a point where you stop looking, and see everything. Without this hunter's eye, you can never see all of the motion. It's the same thing when you're looking for mushrooms. Part of it's pattern recognition, and part of it's this disciplined way of looking. Serendipity shouldn't be random. It requires hard work. Most companies in the business of elephants can only think of elephants. If you look at agile development, for a week you put your heads down and nail all of the things you need in your sprint. And then they may suddenly decide to flush all the code down the toilet and pivot. And it takes an incredible amount of courage for YouTube to shift from being a dating site, and Flickr from being a game. Some people mistake perseverance for courage.

Hoffman: The vast majority of successful companies in the Valley have made major pivots. At Paypal we abandoned Palm Pilots one year in. Plans fail as soon as they make contact with the market. The possibility that I'm going to make a major pivot is very real, but that doesn't mean it invalidates my thesis and what we're doing, it means you've spent enough time with the problem to see a very real opportunity.

Joi: The sprint is hard, focused work that gets you the prototype. But then when you've got it, you can pivot.

Q: I'm a strong believer in intuition. Pivoting on a project is like driving a car somewhere and thinking about how the engine's working.

Hoffman: I describe it as writing and editing. Writing is the work, and editing is questioning whether it's going to work or not. And rarely is someone good at doing both simultaneously, and that's why a team is good.

Intuition is rarely completely data-driven. You're frequently running across a minefield in a fog with limited information and limited time, and all of that combines to where intuition is a combination of good judgements and intuition.

Joi: Or right brain and left brain. I'd dive into Ed Boyden's work. He's saying, how do we talk to these neurons? Oh, there're these algae that use light to do it. And here's this virus that gets them there. Now we can read and write to the brain. It's funny, you use virality as a metaphor, we use it literally.

Hoffman: It's actually a very close metaphor, because virality online often follows a lot of the same mathematical rules.