 I don't want to have favourite speakers, but Sean's one of my favourites too. We met on the TEDx stage actually a few years ago and have been sort of hanging out ever since. He's originally from New Zealand but now based in San Francisco, where he's co-founder and CTO of Quid, an augmented intelligence company. I'm going to read this list then you just see basically he's an overachiever. He's a physicist, a Rhodes scholar, obtaining his PhD from Oxford on the mathematical platforms that underline that modern war. And this research has literally taken him all over the world from the Pentagon to the United Nations and Iraq. And he's often on CNN. I always see on the American news talking about war and cool stuff. So he has had a past life, however. There's a lot of rivalry and competition between Sean and I. Sean was a fellow nanotechnologist. He used to be one of us and then he switched. Who used to work at NASA keeping the rocket theme on self-repairing nanocircuits. And therefore we have three rocket themes today. That was number two. He has a huge list of massive successes and not just in business, but also in sports. He's a two-time former national decathlon champion. Now I was like, okay, whatever. I bet you're not that good. So we went on, you know, those nice cycle renter bike go around why he can drink some wine tours. Two A-type personalities who are sporty nuts did that wine tour on bikes. I don't think we drank any wine. I think we literally raced each other around the whole island and we missed the winery. And we actually got called back by the event organizers being like, we saw you guys. And you cycle past the winery and you kept going. And I'm like, because I'm winning. So anyway, as well as his athletic prowess and quid, he's a super fun, exciting guy to be working with. And I think the thing that's made me most jealous this year from Sean was the fact that he got to work with Morgan Freeman. I know. I'm sharing his research off on the TV show through the wormhole. So if you haven't seen that Google, it's pretty awesome. And Morgan Freeman is one of those amazing people, as is Sean Golly, who is our next speaker. Sean. That'll work. All right. Let's get that there. Thanks, Michelle. I don't really know how to respond to that. But if you ever get the chance to have Morgan Freeman for a few minutes, you should do one thing. Absolutely. And that's going to leave a voice message on your phone. You really can't go past being introduced by Morgan Freeman on the thing. I would also kind of put the caveat on that bike racing. You know, you go to Waikiki Island, you do the kind of the tour around, right? But it's supposed to be leisurely, right? She showed up with a racing bike. So, you know, I was like, at that point it was on. I want to run through Vaughan Sean. Here we go. Brilliant. I'm going to start with this. Because this thing for me really sums up the year that we've had at Quid. It's been a year of selling. It's been a year of getting up and pitching people time and time and time again into the big companies, the big governments, the big hedge funds around the world. It's something that I think when I started Quid, I never really had the sense that that's what the company would actually be doing, would be selling. You always think you're going to build a team, you're going to build technology, you're going to build the product. You never realize that at some point you're just selling that thing the whole time. But it's been a wonderful thing because for me coming into it, I was like, I'm not really a salesperson, right? You know, it's like selling for other people. I don't like getting told no. So I'm probably not a very good person and I really don't like asking for money because it's kind of awkward. So for me going through this I think has kind of been a couple of things. One is you realize when you build a company, you ultimately sell stuff, which should have probably occurred at the start of it. But secondly, it realizes too, is I think we have stories that tell ourselves about what we're good at and what we're not good at, where our strengths are and where they're not. And for me it was always, well, I do technical but I don't do selling. And I think coming through this year, I think I've earned my selling badge at the very least. And so we close the year out and I get a little bit of time now in New Zealand to take a bit of a break and just absolutely thrilled we did 8.5 million of subscription sales this year with the first full year of the software on the market. And so for me that just represents kind of a huge milestone of creating software and having it out and deployed with the sales team and kind of hitting our number. Our number for the year was 8 million, which to kind of go past that was really, really fantastic. So thank you. But it was also important to come back to New Zealand. So when I was chatting with Scott about coming down here, it was the single most important thing. I said, I've got to get down to New Zealand. He said, you know, why have we got to get down here? I said, for me it just gives me a chance to take a break from that whole circus in Silicon Valley and just get out on the surfboard back in Christchurch and just think about where we've come from, think about the next year where we're going but also think some different ideas. And I think it's so nice to be able to come home to New Zealand and take that time and that space. So I want to kind of share some of that stuff that's on my mind is I think forward towards the next year and kind of share some of that with you. I thought about maybe like saying, well, I could tell you about how to build companies but I think honestly if you were looking at advice from me, it's probably the wrong person. I think I've probably made most of the mistakes and I always think it's funny when someone says, how do you do the stuff? I have no idea. I really don't but I do have some interesting thoughts I think you're like on this. So this for me is fascinating, the human brain and particularly the limits of the human brain. What it can do but also what it can't do. And I think we tend to think that it's infinite, right? We've got infinite capacity in the way we can think but there are very, very real limits in what we can do. And one of these is the speed of thought, right? So we can think only so fast and scientists got together and they measured this and they said, well, how fast can you think? And they took a bunch of chess players and they stuck them into an fMRI machine and it started to measure the time that they would react to whether the board was in check or not. And they found that the experts could do it in about 650 milliseconds and the amateurs about 900 milliseconds, right? So no matter how good you were at chess you couldn't think faster than 650 milliseconds as to whether the king was in check or not and this sort of represented the smallest atomic unit of strategic thought. So the human brain can't think faster than .65 which is about this long, right? That we don't exist below that and it seems kind of trivial, right? It seems like that shouldn't really make a big difference to anything except this is what happens because of that little limit in our cognitive processing power. Today we're looking about 65% of all US equities are traded by algorithms and they're traded by algorithms because algorithms can think a whole lot faster than that. And now instead of looking at people trading stocks backwards and forwards it's a bunch of servers co-located to be as near as possible to the source of information because the speed of light makes a difference in these equations and it's developed a whole industry around this. It's an industry now where they're laying cables between London and New York to shave 5 milliseconds across the round-trip time between the two exchanges. So for me this is $300 million spent to take 5 milliseconds off a trade that dominates the market because the human mind doesn't think any quicker than .65. So for me when I look at where the really interesting stuff is I think, well, where does the human brain stop? And where the human brain stops the algorithms will eventually take over and that has pretty profound implications not just in technology but also the physical environment. But in addition to that I think our relationship with these algorithms is going to be one of the defining characteristics of the economy in the next decade and I want to run you through some of the thinking that I've been doing about that. So there's definitely benefits, right? Like if you run this you'll get the bid ask spread it'll go down to almost zero. Really, really efficient markets, massively efficient and it's almost always efficient almost all the time except when that happens, right? And those that don't read those stock tickers for a living you'll probably know it more like that. So Flash crashed at 245 where collectively the market decided to take a trillion dollars in market capitalization for no real apparent reason the algorithms got together and decided to remove it and then give it back 15 minutes later a little bit unhurt and slightly worse for wear and they still don't quite know what went on in that period of time, right? The algorithms collectively decided that they wanted to run the world's markets and they decided they wanted to impose their own will on things and they traded stock in Sotheby's for $999,000 and stock in Accenture for a cent maybe they knew something we didn't know but that's where they got to and it's because if you look at the thinking time, right? The humans, we think if you read a scientific journal three hours watching a post article, you know maybe seven or eight minutes you get down to about three seconds, a 140 character tweet, right? But beneath that you're moving into the sub thousand millisecond the limits of human cognition and at that point it's the algorithmic ecosystem, right? And this is the, where you're looking at the time for light to travel from New York to London, 65 milliseconds you can execute a trade in about 100 microseconds on the NASDAQ and now they've got hardware executable chips so they'll do 740 nanoseconds to do a trade now nanosecond for those that aren't physicists is about the amount of time it takes for light to travel that far that's why, you know, the position of your servers inside of the co-located infrastructure is really important they actually measure out the cable so everyone has the same amount of cable but that's the kind of the world they're in, right? And that's what it looks like we can't see that, again that's 900 milliseconds up the top but that's, you know, a stock being traded by a bunch of algorithms you know, here's what happens Knight Foundation, not Knight Foundation, Knight Capital is a massive trader in this, right? Knight Foundation moving that would be strange but these guys run, you know, some of the biggest high-frequency trading and they showed up one morning they were releasing a new algorithm into the system they got this kind of set of servers that they run the algorithms on they test it, they do all that stuff and then when they're happy they release them into the market, right? They go and trade well, they did that this morning one of the mornings back in 2012 and the markets started going a little bit strange and they knew the market was strange they didn't know that they were responsible for it going strange but it was all of a sudden they're like, the market's weird and 45 minutes it took them to figure out that they'd release an algorithm into the market that was never supposed to be in the market it hadn't been fully tested in fact it was the opposite of what should have happened it was an algorithm that was used to stress test the algorithms that were supposed to go into the market so it was a dummy algorithm that would sell high and sell low and buy high, right? do the opposite and it got released and they lost $440 million in 40 minutes and then they pulled the plug and then they got effectively got bought up for distressed assets at $0.15 on the dollar right? and it just sort of shows you these things they're sort of like crazy, right? like this shouldn't happen but again, this is stuff, this is reality this is going on and it goes on because we don't think at the speed but the algorithms dominate at the speed and again, it's start thinking about well, what does this mean for our economy? what does it mean for us? so algorithms like the night stuff they read zeros and ones, right? zeros and ones, basic time series analysis really, you know driving the quant driven revolution of financial markets but they also now are starting to read news, right? they're starting to read not just ones and zeros but, you know, letters and digits and this is really fascinating because this sort of stuff happens this is a bunch of E-mini futures there are a type of thing that you can trade and you'll look at the volume there, right? that's the total volume and then you see it sort of stop and get very, very small, right? so the algorithms all leave the market for some reason, right? remember, the algorithms dominate the market and they all leave and you think, well, why are they leaving? and they're leaving because there's a schedule news announcement about unemployment data coming out and the unemployment data is coming out at 12.30 so all the algorithms get out of the market at 12.25 and they say, I don't know how to read that stuff so I'm not going to trade and what you're left with is humans and the humans read the stuff and they go, yeah, nothing, nothing really significant here and all the algorithms they say, push the button, come back in but you look at it, you think, jeez, like the algorithms decide to leave the market, the market's kind of screwed as far as liquidity goes and that's because, you know, they can't read but if they could read, what would they do? and that's exactly what they're starting to do because in a world where you're competing for the speed of light and the distance that it can travel the 10 minutes that they're out of the market is an eternity huge amounts of time remember, we spent $300 million to get 5 milliseconds on a cable, right? so trading algorithms to read news is going to open you up 10 minutes what's that worth? so this I think is going to be very, very significant over the next couple of years the algorithm's getting very good at reading news now that seems all well and good algorithms reading news, trading markets get a machine readable news system read through an article a bomb goes off and bagged out I'm going to trade the VIX index or I'm going to trade the currency, blah, blah, blah it's going to be more volatile, I'll make some money until stuff like this happens, right? breaking news to explosions in the White House and Barack Obama is injured never happened, this was a hacked story done by humans this time from the AP Newswire and when that happened, this happened the algorithms left the market they said there's a bomb, I don't know what to do and $200 billion was removed from the market and it came back and again the volatility changed and for me again is like well we've got algorithms now being able to read news, not sure what to do there's opportunities if they do start understanding it they'll start trading it, but of course if they start reading the news, won't they start writing the news? and they start writing the news how do we believe and what about us isn't the news meant for us, right? so we log on, right? we go on and we think well this is again a kind of step back, okay so financial markets are gone we don't run them anymore we can't think fast enough that the new programs are going to start reading our news and potentially writing it well when you open the New York Times you open the Herald it takes a little time to load, right? you'll notice it takes a few seconds you blame New Zealand's shitty internet for that it takes about 800 milliseconds to load and you'd be right for the most part to blame New Zealand's shitty internet for that, but you should also realize 150 milliseconds of that is built in for an online ad auction that happens so when you go to the website it takes a few milliseconds and says we're going to auction off your presence we're going to auction off your presence to the highest bidder to show you an ad, yes or no, right? so there's like a little market that evolves for selling ads backwards and forwards and I think that's really fascinating again because 61% now of the internet traffic is non-human so the internet's kind of switched the same way as the markets it used to be the internet was for us, right? we had the traffic now we don't have the traffic anymore and it's up from 51% last year or 49% last year so again the internet is slowly moving that way because there are advantages to being able to think 150 milliseconds that we simply cannot do it's not within our biology so at this point you kind of want to push the button and say stop like, you know, this all seems a little bit complicated and a bit too much, but the reality is we can't really turn back from this so how do we deal with it? and also let's think a little bit about what we're actually dealing with so, you know, Google and Facebook what they kind of think of is having massive information filtering algorithms ways to kind of separate information back out to you and we think of that and we experience them as products and we think about how they do things but we don't often think about the economics and a large part of the economics of these two companies is electricity because you get electricity to store information and retrieve it and electricity is one of the single largest costs of running these two operations if you flip it back out what you see is it costs Google about $12.90 per user per year and it costs Facebook about $1.20 to give you the product I think these numbers are just fascinating that you can have Facebook for $1.20 obviously if you've got a billion people that can join you on that you can have that but it's not a lot of money what they're spending on the electricity to give you the product at the end of the day and that's even kind of more surprising because you look at things like this Wikipedia shows up 85% of the time for the 1,000 most common searches on Google in many ways Google is sort of crawling Wikipedia and you probably wouldn't notice too much of a difference if all it did was give you back results from that so it's kind of interesting that this kind of search and filtering mechanism that exists and I think of this because you think about well we've got Facebook and we sort of accept Facebook as a social network but what else could it look like and we say well why does it look like what it does and we've got a couple of things here the identity is one axis and you take the history so am I real and do I have to be me online and if I post something are you always going to remember it and Facebook is yes you have to be real and I'm going to remember everything and they're up there because that's really really profitable that's a hugely profitable space to advertise against but we could also think of one where I don't have to be real and you'd be reddit right you can have your pseudonyms that's a really different kind of community we can think about what no history and no identity looks like does anyone think about that well that's 4chan if you haven't been there don't go there you'll ruin your day but it's a really different kind of world and then of course on the other side real identity, no history is Snapchat so you start to explore the space of different kinds of ways of being now these are algorithms the algorithm is doing something very simple do you have to be you and am I going to remember you and just by varying those two variables you end up in a difference between Facebook and 4chan and the online behavior that we experience through them is night and day and I think you should just kind of reflect on that and think what power are the algorithms having on the kinds of interactions that I'm having with the world and did I want you know the all history all identity is that what I want because that's kind of what we've got and I think it's coming back and saying well how do you design for these worlds so coming from that I think kind of splitting this out we've been treated I think in many ways as products by algorithms most of us interact with algorithms as products our identities are bought and sold online to give us advertising that we'll hopefully buy things from so we've had that but that's pretty narrow right there's a whole lot of other ways that we can interact we could actually be the owners of these algorithms so instead of the algorithms serving information and the hope that I might buy something I could actually own the algorithm and say well give me things that'll make me a better person make me a better friend, make me smarter make me learn about stuff that I probably don't really want to but I probably should and don't try and sell me that stuff so we could own them we could be a really different kind of relationship we could also be on the other side a sort of technosurf well you can't afford I can't afford to be monetized as a product but there's some stuff the algorithm is not very good at doing and it's going to offset them to me so you can kind of think of being a servant of the algorithms and that's actually happening but you can also think of augmented intelligence actually driving algorithms being in a relationship with algorithms as sort of a conversation you own them of course but you're not just accepting what they're giving you you're actually feeding back and saying yeah I want to drive it like I'll drive a high performance race car so there's a whole different range of ways of interacting with algorithms but what we've been given is very very narrow and I think to our detriment this is a graph that kind of struck me as all the it's my favorite graph for the year and kind of what it says a graph of inequality I don't know if anyone saw this when it came out but it should worry you I think because what it's looking at is the amount of income owned by the bottom 90% versus the top 0.1% and you see that just at the point where they're crossing and you'll also see last time they were crossing was this lovely time period in the 1920s through to the 1930 and so for me this I think this graph and this rise of inequality it also jumped out in 1984 New Zealand was one of the most equal places in the world over the last 25 odd years we have had the single biggest jump in inequality in the world and so the stories that we had of growing up all equal I think are kind of different but the whole world's moving that way and I think what I put to you is the large reason of this is because of algorithms algorithms are fundamentally making our world more unequal because the information and the algorithms that are operating don't affect everyone equally and I think this is something to really consider because if you can afford the high end subscription for an augmented intelligence product you'll drive and find insight in the market whereas everyone else who can't won't the difference between being treated as a product and being driving software to make trades are really really different and really distinct so something again to think about we talked a little bit about technosurfs and for those that know Amazon's Mechanical Turk this crowdflower as well as a platform that lets that run so these are systems that you log in and they're called human implementable tasks and you get paid maybe between five and ten cents to do things like determining whether that picture is obscene or not or determining what color the sum object is things that the algorithm could do if it had money for the electricity but actually you're cheaper so I can run with you instead and maybe I'm not quite smart enough and I can leverage you and so we've got crowdflower boats that have a million human hours available at any moment you can access a million human hours to do things that a few cents a pop things that algorithms can't do because of cost or can't do because of complexity this leads to some really interesting things this is a performance art piece by Aaron Koblin he paid people two cents to draw a sheep in New Zealand you have to have sheep they got seven and a half thousand people to draw a sheep they got ten thousand sheep cost them twenty bucks and that was great because then he took all those sheep and created an art exhibition and sold each of the paintings off at twenty dollars and then people complained they said you can't sell my thing at twenty dollars they said yeah you signed the user acceptance thing so he took what people did at two cents and sold it at twenty bucks it was sheep and people he said draw sheep for two cents he was like okay which you know that's why one of my favorite pieces of art here and it's like is that our relationship with these systems are we the sheep, are we creating sheep who's getting the value, where is the value going and I think you should sort of pause and think about that for a moment but there's actually some really good sheep in there for two cents augmented intelligence this is the space that I've been working in the kind of the human coupled with the machine, the center and you can think of a bunch of companies like Palantir, Record of Future and I asked they all working in that space creating these intelligence platforms that let people interact with algorithms it looks a little bit you can kind of think of that there but it looks a little bit like this this is the software that we sell but you can kind of dive in and see it in time and again go back into all the space here what we're looking at is all the all the active participants in the world of the global space industry you've got SpaceX up in top and red all its first-order partnerships that it's done business with in yellow and its second-order ones there in green you've got thousands of objects tens of thousands of things it's all been extracted out from open source unstructured information projected down to a three-dimensional interactive space that you can go and explore and figure out and understand the structure of the world that's around you so you know at this point it's not like the computer's giving you an answer of like you know I predict with probability X what's going to happen with the global space industry it's actually said here's the information go and explore it and figure out how you want it to be figure out where you want to position yourself and so for people like NASA who use the software a lot of it is about trying to understand the structure of how the space evolves and then figuring out how to position themselves within that space there another one here's is Occupy Wall Street you can kind of see this time the stories unfolding in real time as the political protest moves you can see the stories coming out and this becomes very very powerful because you start to think about it as you know if I'm putting out stories what I want to position them you know where do I want my story to end up is there a white space in the middle but I can start to connect ideas together to get resonance that no one else is really seeing and start to monitor that as that goes through you've got a topology of information that you can effectively interact with and understand you know how do I want it to change so we did this as sort of a I guess a demo about 18 months ago this midterm election we had a bunch of senators and very close races starting to use the software now from the US political campaigns and everyone that used the software actually ended up winning their races most of them are on the Republican side so I don't know what that tells you but for the first time this has actually been deployed in elections and again these platforms and if you have access to this information you can see more and see further than anyone else that doesn't and again you start to think about how that impacts and affects inequality as we drive through this different system I want to finish on the final thing which I think has been for me the most fascinating piece to think about is this idea of subscription bots things that you own things that work for your benefit instead of being treated as a product and there aren't really a lot of companies around on that so you sort of have to turn to the next best thing about science fiction and those if anyone has read Neil Stevenson's work there's a book called The Diamond Age and there's a course inside of The Diamond Age there's a book made with nanotechnology called The Young Lady's Illustrated Primer and it's a book that writes itself and it writes itself for the young woman of the story to educate them to excel and learn and be very very successful inside of the world but it's a book that conforms to the reader and kind of makes the story up as it goes along in some ways manipulating them but in other ways educating them and I think this kind of idea of a primer for people a software subscription that you own that has your benefits in mind is going to be something I think we're going to see a lot of as we go forward we're getting there again if we look forward into actual patents this is an auto generation of social status updates so it's like Google's patterning it's like I'm too lazy to update my status but I want people to know that I'm active so can an algorithm do that for me and you start thinking about it again it's machines learning to kind of write machines learning to impersonate humans and we're seeing there's sort of a crude version of this online it's a Facebook what would I say it came up with your my valet ticket which I don't know if that is either a reflection of the badness of the algorithm or just like I don't really post things that are very sensible but still it posted that for me and I was very grateful everyone looked kind of confused I got a few likes it gets a little more complex and this is a bunch of tweets so those that can't read Spanish it's coming out of Mexico City it's a bunch of tweets here it's Mexican Twitter bots alleging that a reporter was not killed by a drug cartel just think about that for a second we've got organized crime syndicate pushing out propaganda bots to convince users online that an alleged murder of a reporter didn't happen right this is again it's not this is happening today so we think about this it's around it's going to have a big big impact on what we do here's what those bots look like and actually if you run the algorithms you can kind of see it in time they all come out every 10 minutes and it's actually pretty easy to see them but they're evolving right this is what happens in China China's got a bunch of these bots this time protests that were happening in Tiananmen Square it's really hard to figure out from the temporal signatures they figured out that the algorithm the people designing the algorithms are like we need to get rid of the temporal signatures of what's going on so we're going to move into a world where algorithms are pretty good at reading and writing and they're pretty good at understanding who you are and they can think fast in you and they can read more than you and they're going to treat you as products to be manipulated bought and sold or you can own them and have them work for you and I think there's going to be a distinction that we're going to start to move in and that decision is going to come with money and some people are going to be able to afford them and they'll do better most people are not or can't afford very good ones you're my valet ticket right that could be mine people would look at me and say sure you should get a better algorithm and they're not going to do so well and so there's an inequality thing I don't see it closing because of the stuff and it's something we really need to think about and I sort of leave you with this final piece of that and this is maybe a world that might exist I hope it doesn't but it's something that we should all really kind of consider and maybe over the Christmas break as we're having a little bit of time relaxing think about that and how we want to build the companies that we're all going to be building in a world where this is now a possibility so thank you