1 00:00:09,679 --> 00:00:14,680 The specialists whose brains are picked to program expert systems, are delighted to produce 2 00:00:14,680 --> 00:00:18,560 machines that are cleverer than they are. They feel a pride of authorship. 3 00:00:18,560 --> 00:00:26,070 But what about the rest of us who have not been consulted, how will we feel? 4 00:00:26,070 --> 00:00:30,509 Computers seem set to learn at a rate that far outstrips our speed of learning. They 5 00:00:30,509 --> 00:00:35,700 may even understand what they learn, but the hallmark of human intelligence is the: 6 00:00:35,700 --> 00:00:40,600 "Eureka Effect." Will computers be able to come up with anything new? 7 00:01:23,600 --> 00:01:30,510 My name is Karen Sandler, I am a lawyer, I am a cyborg-lawyer! 8 00:01:30,510 --> 00:01:35,869 A few years ago while I was already a lawyer with the Software Freedom Law Center, I found 9 00:01:35,869 --> 00:01:40,240 out that I had a heart condition. I have an enlarged and obstructed heart, it's called 10 00:01:40,240 --> 00:01:43,900 Hypertrophic Cardiomyopathy (HCM). I don't have any symptoms, it's all fine, 11 00:01:43,900 --> 00:01:48,439 but I'm at a very high risk of suddenly dying. If you hear about kids that are running to 12 00:01:48,439 --> 00:01:53,080 second base and drop dead all of a sudden, they often have what I have. 13 00:01:53,080 --> 00:01:57,070 It's all well and good, there are some preventative medications I can take, but the main thing 14 00:01:57,070 --> 00:02:01,799 is that to prevent suddenly dying, I was prescribed a pacemaker-defibrillator. 15 00:02:01,800 --> 00:02:04,900 It was like people following me around with paddles. If I passed out, I would get a shock, 16 00:02:04,900 --> 00:02:07,700 it would be implanted, and it would be great. After dealing with the idea that I needed 17 00:02:07,710 --> 00:02:11,190 to get this device which was quite, you know, stressful for anybody to hear that they need 18 00:02:11,190 --> 00:02:14,849 something implanted in their bodies. I started asking more about it, and the doctor 19 00:02:14,849 --> 00:02:19,659 showed me the device and he said: "This is what you need, and you should get it implanted 20 00:02:19,659 --> 00:02:24,019 right away." And so I asked, because I was a lawyer at the Software Freedom Law Center, 21 00:02:24,019 --> 00:02:29,159 "Well, what does it run?" And he couldn't tell me. That entire doctor's office had never 22 00:02:29,159 --> 00:02:32,079 thought about the issues of software on these medical devices. 23 00:02:32,079 --> 00:02:38,230 It was a little bit strange, I literally had to get proprietary software hooked up to my 24 00:02:38,230 --> 00:02:43,010 heart and implanted in my body. I did it because I couldn't walk around at risk for suddenly 25 00:02:43,010 --> 00:02:47,999 dying every day. My parents, everyone I loved was really angry at me for holding up on 26 00:02:47,999 --> 00:02:51,629 this issue. For them it was this abstract idea: "Who cares what software is running 27 00:02:51,629 --> 00:02:56,200 on your device, it's going going to save your life, you need it!" 28 00:03:00,249 --> 00:03:05,139 This decision about technology and what kind of technology stack we have will define the 29 00:03:05,139 --> 00:03:10,089 rules that are the new laws of how society works. And people will structure society based 30 00:03:10,089 --> 00:03:14,540 on what the software allows them to do and, not based on what the laws are written, because 31 00:03:14,540 --> 00:03:18,139 the software becomes this new set of laws that constrain society. 32 00:03:18,139 --> 00:03:23,949 They're working towards constructing their own immortality, and their own infinite pleasure, 33 00:03:23,949 --> 00:03:28,540 and not just for themselves, but for their whole community, and they perceive even for 34 00:03:28,540 --> 00:03:32,329 all of humanity itself. They dystopian consequences of their work 35 00:03:32,329 --> 00:03:38,169 is not what is most present in their mind. Software is so important to so many things, 36 00:03:38,169 --> 00:03:43,290 and if you can control the software, it's like controlling the flow of oil. You can 37 00:03:43,290 --> 00:03:48,889 turn off access to certain features to other companies. I don't think it's a surprise or a 38 00:03:48,889 --> 00:03:53,599 coincidence that there's a lot of talk about monopolies in software. 39 00:03:53,599 --> 00:03:57,599 The freedom that the Internet brings to people, if you look at all the revolutions and things 40 00:03:57,599 --> 00:04:00,059 that have been happening in the Middle East, it's amazing, and those things couldn't 41 00:04:00,059 --> 00:04:04,430 have happened in that way before the Internet existed. It really makes a difference to people's 42 00:04:04,430 --> 00:04:06,729 life and to democracy. No one contributes to a community that only 43 00:04:06,729 --> 00:04:10,960 benefits one person, that's just not how it works. The core element of community is, when 44 00:04:10,960 --> 00:04:16,799 you make a contribution, the community benefits from it. The altruistic moral perspective 45 00:04:16,799 --> 00:04:19,000 is what drives all of this technology forward. 46 00:05:03,000 --> 00:05:05,400 Quiet Please, I am analyzing. 47 00:05:12,740 --> 00:05:16,960 Shiny hardware gets the attention, but it's the software code behind the screen that's 48 00:05:16,960 --> 00:05:20,449 making the magic happen. A computer is willing to divide by zero all 49 00:05:20,449 --> 00:05:24,930 day long. In reality, the possibilities of technology are limited only by the imagination 50 00:05:24,930 --> 00:05:30,050 of the users, and the license agreements. With the Internet, people can learn and create 51 00:05:30,050 --> 00:05:34,789 in their pajamas. The current copyright system was created by 52 00:05:34,789 --> 00:05:41,229 people who didn't imagine millions of digital projects built remotely. 53 00:05:41,229 --> 00:05:46,520 Or, that it would cost basically nothing to make more copies. 54 00:05:46,520 --> 00:05:52,060 The Baby Boomer generation of programmers made a mistake: they forgot the scientific 55 00:05:52,060 --> 00:05:54,620 tradition of sharing results. 56 00:05:58,340 --> 00:06:00,780 Starting in the late 1980's, a new kind of 57 00:06:00,780 --> 00:06:05,680 license agreement was created, which increased freedom and the pace of scientific progress. 58 00:06:05,689 --> 00:06:09,360 Not everyone will exercise their right to change their software. In the past, not many 59 00:06:09,360 --> 00:06:13,319 people owned a printing press, but it was widely understood that a free press was important 60 00:06:13,319 --> 00:06:17,860 to the civil society. Those who don't write software can still help protect this freedom 61 00:06:17,860 --> 00:06:19,400 for others. 62 00:06:23,900 --> 00:06:26,300 The first computers in the 1960s that people 63 00:06:26,400 --> 00:06:32,830 bought were giant mainframes. They were incredibly expensive to deliver, to cool, to maintain, 64 00:06:32,830 --> 00:06:40,190 and companies like IBM developed these computers, and in end just sort of gave software away 65 00:06:40,190 --> 00:06:44,159 as an afterthought. They realized that no one was going to buy 66 00:06:44,159 --> 00:06:48,160 one of these giant computers that needed a special air conditioner and had to be installed 67 00:06:48,160 --> 00:06:53,280 in a special way, in some special room, unless they could do something with it. So the software 68 00:06:53,280 --> 00:06:59,401 was added as an afterthought and the cost of doing software was nominal compared to 69 00:06:59,401 --> 00:07:04,520 the cost of making the computer. And over time, computers obviously got smaller, 70 00:07:04,520 --> 00:07:10,669 and they still are getting smaller today, and the idea of having a computer that individuals 71 00:07:10,669 --> 00:07:17,830 could buy became initially a hobbyist activity. There were many individuals in the 1970's who 72 00:07:17,830 --> 00:07:22,949 got very excited about buying and sharing computers. And I remember as young child, 73 00:07:22,949 --> 00:07:26,680 the first computers that my father brought home were these kinds of hobbyist machines 74 00:07:26,680 --> 00:07:31,270 that you could look at and modify. And in fact, they were very similar to the 75 00:07:31,270 --> 00:07:35,470 software freedom communities we see today, because they swapped the software once they 76 00:07:35,470 --> 00:07:38,330 got their machine. And of course, there was also the hardware component of putting it 77 00:07:38,330 --> 00:07:43,080 together and assembling it. But then you started adding some software to it, and writing software 78 00:07:43,080 --> 00:07:47,090 yourselves, sharing it with your friends as these things called Computer Clubs. 79 00:07:47,090 --> 00:07:53,169 Well, at the same time that all that was happening, there were various different software companies 80 00:07:53,169 --> 00:07:58,580 there were focusing on making lots of money as opposed to necessarily joining a community 81 00:07:58,580 --> 00:08:02,580 and sharing software. How much do you think this advanced operating 82 00:08:02,580 --> 00:08:07,130 environment is worth? Wait just one minute before you answer. Watch as Windows integrates 83 00:08:07,130 --> 00:08:10,842 Lotus 1-2-3 with Miami Vice. Wait until you see Windows Write, and Windows 84 00:08:10,842 --> 00:08:15,350 Paint, and to listen to what else you get at no extra charge: the MS-DOS executive, 85 00:08:15,350 --> 00:08:23,300 an appointment calendar, a cardfile, a notepad, a clock, a control panel, a terminal, print 86 00:08:23,300 --> 00:08:28,050 spooler, RAM driver, and can you believe it, Reversi. That's right, all these features 87 00:08:28,050 --> 00:08:35,700 and Reversi, all for just, how much did you guess? $500, $1000, even more? No, it's 88 00:08:35,700 --> 00:08:39,200 just $99, that's right, it's $99. 89 00:08:40,200 --> 00:08:43,200 There's a very famous letter to hobbyists 90 00:08:43,229 --> 00:08:50,040 that Bill Gates wrote in the 1970's saying that the idea of sharing software is stealing. 91 00:08:50,040 --> 00:08:55,640 And that's where the division really begins. Ironically, right around the 1980's, as Microsoft 92 00:08:55,640 --> 00:09:00,800 was getting started, is exactly when my colleague Richard Stallman decided to start the software 93 00:09:00,800 --> 00:09:06,100 freedom movement, and wrote a document called "The GNU Manifesto" that in some ways I tend 94 00:09:06,100 --> 00:09:12,260 to think as an answer to the "Letter to Hobbyists." Now RMS didn't write it as an answer, he wrote 95 00:09:12,260 --> 00:09:17,500 it because it was what he believed it was the right thing to do for software users, 96 00:09:17,500 --> 00:09:22,620 but when you put these two documents up next to each other, you sort of see two philosophies 97 00:09:22,620 --> 00:09:28,000 of how software ought to be done. Paid-for software with restrictive licenses 98 00:09:28,000 --> 00:09:34,029 are still the more common type today. Linux runs on about 1% of desktops and laptops. 99 00:09:34,029 --> 00:09:39,300 Many companies depend on free software, but prefer to keep their innovation secret. Google 100 00:09:39,300 --> 00:09:43,860 is working on a lot of artificial intelligence code in private, and many other companies 101 00:09:43,860 --> 00:09:48,000 are also. Google released a neural network system called 102 00:09:48,000 --> 00:09:54,690 Tensorflow in 2015, but most of their software and data to make it intelligent, remain hidden. 103 00:09:54,690 --> 00:10:01,040 If you use Google Photos, and you upload a photo, it knows how to categorize what's 104 00:10:01,040 --> 00:10:07,220 in your photo. Now how does it possibly know that's a picture of a giraffe or a car or 105 00:10:07,220 --> 00:10:10,820 what have you? Well the answer is it's been trained, right, we've showed it picture 106 00:10:10,820 --> 00:10:15,360 after picture. And in the 1970's people tried this and 107 00:10:15,360 --> 00:10:19,160 it didn't work, and then there was what is called the "AI Winter" which is sort of 108 00:10:19,160 --> 00:10:22,970 about 20 years when nothing happened. And then a set of mathematicians figured out 109 00:10:22,970 --> 00:10:28,120 a way to build what are called deep neural networks, and they are layers of these computer 110 00:10:28,120 --> 00:10:33,209 models that are very complicated. And the first layer says: "Ha, this is a picture, 111 00:10:33,209 --> 00:10:37,340 it has boundaries, here are pixels." The second layer says: "Here are where the shapes are." 112 00:10:37,340 --> 00:10:41,149 And the third says: "Ah, here's a shape here, and here's a relationship." And there 113 00:10:41,149 --> 00:10:46,320 are typically 9 or 10 levels, and the 10th level says: "It's a giraffe, it's a car, 114 00:10:46,320 --> 00:10:50,710 it's a person." When I was heading up Google X a few years 115 00:10:50,710 --> 00:10:55,621 back, and one little project we had in there, which is now called Google Brain, which was 116 00:10:55,621 --> 00:11:01,560 this AI effort. But, I didn't pay attention to it at all, to be perfectly honest. 117 00:11:01,560 --> 00:11:09,420 And, myself having been trained as a computer scientist in the 1990s, everyone knew AI didn't 118 00:11:09,420 --> 00:11:15,690 work, people tried it, they tried neural nets, none of them worked out. And this fellow, 119 00:11:15,690 --> 00:11:20,800 who is one of our top computer scientists, Jeff Dean, would periodically come up to me: 120 00:11:20,800 --> 00:11:24,006 "Look, the computer made a picture of a cat." I'd say, "Okay, that's very nice Jeff, 121 00:11:24,820 --> 00:11:32,930 go, do your thing, whatever." And fast forward a few years and now, Brain probably touches 122 00:11:32,930 --> 00:11:37,730 every single one of our main projects. If you look at something like Google Translate, 123 00:11:37,730 --> 00:11:45,199 it's amazing. It's pretty hard to learn 8 languages, let alone 65 languages, that's 124 00:11:45,199 --> 00:11:48,520 really remarkable. The field of artificial intelligence use to 125 00:11:48,520 --> 00:11:54,329 be called AI/MT, Artificial Intelligence / Machine Translation. This machine translation was 126 00:11:54,329 --> 00:11:59,040 basically to tap into funding streams from the National Security Agency, because they 127 00:11:59,040 --> 00:12:04,699 had to translate all their intercepts. But the machine translation part is basically 128 00:12:04,699 --> 00:12:12,840 done. How did that happen? Was it 10s of thousands of programmers making lots of grammar rules 129 00:12:12,840 --> 00:12:16,970 for all sorts of different languages? No, the breakthrough is, it's looking at translations 130 00:12:16,970 --> 00:12:24,130 of UN speeches and EU regulations. And there's many, many language pairs of the same document. 131 00:12:24,130 --> 00:12:29,279 And with millions of Rosetta stones, you are able to construct a computer which can learn 132 00:12:29,279 --> 00:12:34,640 the principles of language and therefore translate. When you get a program from say, a proprietary 133 00:12:34,640 --> 00:12:40,920 software company like Microsoft, you are left effectively helpless. You don't have the right 134 00:12:40,920 --> 00:12:46,560 to study the software and even learn how it works. You just get the compiled version of 135 00:12:46,560 --> 00:12:50,389 the program that's in binary form that only the computer can understand. 136 00:12:50,389 --> 00:12:57,889 I've been to one meeting with Steve Jobs, that's like almost 10 years ago. He was 137 00:12:57,889 --> 00:13:03,630 trying to convince to start working on OS X instead of Linux, and I said: "Sure if you 138 00:13:03,630 --> 00:13:06,720 switch it to Linux." But it wasn't going to happen. 139 00:13:06,720 --> 00:13:13,160 Now I'd like to go through the architecture of Mac OS 10 with you. To start off with, 140 00:13:13,300 --> 00:13:18,040 we need a kernel. And that kernel we call Darwin. 141 00:13:19,080 --> 00:13:21,690 Darwin's amazing, it's a super 142 00:13:21,690 --> 00:13:27,389 modern kernel, and it's got protected memory, preemptive multi-tasking, modern networking, 143 00:13:27,389 --> 00:13:33,560 all the things you'd expect, done in a very elegant way. And it's very Linux-like, 144 00:13:33,560 --> 00:13:37,400 very much so. It's got free BSD Unix which is the same as Linux. 145 00:13:37,940 --> 00:13:44,300 Apple picked the BSD software on purpose. They knew about copyleft, by the time they 146 00:13:44,300 --> 00:13:50,470 picked it, they avoided picking copylefted programs because they specifically want to 147 00:13:50,470 --> 00:13:56,130 lock up users and not let them share and improve their software, devices that they themselves 148 00:13:56,130 --> 00:14:01,550 own. They can't do anything with them other than, well if it breaks, call Apple. 149 00:14:01,550 --> 00:14:06,329 Copyleft uses copyright, but to keep it free. It's like a completely spinning around of 150 00:14:06,329 --> 00:14:10,500 the copyright system, and instead using it to control, it uses it to share. It has a 151 00:14:10,500 --> 00:14:14,889 kind of a snowball effect, if you improve it, those changes, if you're publishing them, 152 00:14:14,889 --> 00:14:18,690 also have to stay under the same license. So the body of free software just grows and 153 00:14:18,690 --> 00:14:23,130 grows and grows. People call it "forever free" and I think it's such an amazing concept. 154 00:14:23,130 --> 00:14:28,350 In some cases, the free software equivalents to proprietary applications are not as well 155 00:14:28,350 --> 00:14:33,190 known, or quite good enough yet, so people continue to pay billions for closed software, 156 00:14:33,190 --> 00:14:34,800 and give little to the free software movement. 157 00:14:35,000 --> 00:14:39,200 Change is very emotional. People don't like change. 158 00:14:39,400 --> 00:14:44,100 No one likes to move house, no one likes to buy a new car, and no one wants to 159 00:14:44,180 --> 00:14:48,700 change software. The relationship that a lot of people have, 160 00:14:48,709 --> 00:14:53,910 like regular people, not geeks, the relationship they have with their computers, is one of 161 00:14:53,910 --> 00:14:58,720 is basically hanging by a thread, and "I am terrified it's all going to go down at some 162 00:14:58,820 --> 00:15:05,399 point." So the idea of changing to this whole new world is terrifying! 163 00:15:05,399 --> 00:15:11,100 You think about most people in front of a computer, they don't actually like their computer! 164 00:15:11,100 --> 00:15:16,829 They're using a computer because it's part of their job. They like using it in the sense 165 00:15:16,829 --> 00:15:22,160 that it does a lot for them but usually they just curse when things go wrong, they don't 166 00:15:22,160 --> 00:15:28,600 want to tweak things, they want to do stuff the way they are used to doing things. 167 00:15:35,480 --> 00:15:40,700 Compuware was formed in 1973 and is one of the largest proprietary software companies, 168 00:15:40,700 --> 00:15:42,180 specializing in the mainframe business. 169 00:15:43,380 --> 00:15:45,940 So why the use of a proprietary software license? 170 00:15:45,949 --> 00:15:49,100 It really comes down to, that's our business model, and that's where we're going to make 171 00:15:49,100 --> 00:15:53,639 our money and our profit, and pay for our employees. So, it really comes down to that 172 00:15:53,639 --> 00:15:58,470 we have something that is novel, patentable, that we can actually turn a profit on that's 173 00:15:58,470 --> 00:16:02,949 in the industry. We are a commercial software company, and 174 00:16:02,949 --> 00:16:09,730 our customers expect a certain level of innovation from us that provides them some competitive 175 00:16:09,730 --> 00:16:12,319 advantage as well. Compuware has always been in the business 176 00:16:12,319 --> 00:16:17,029 of selling software. And we make our money out of the innovation that we put in our software, 177 00:16:17,029 --> 00:16:20,860 and the creativity that we put in our software, as well as the maintenance and support that 178 00:16:20,860 --> 00:16:23,900 we provide behind that software. 179 00:16:23,900 --> 00:16:25,970 In much of the developing world there is very 180 00:16:25,970 --> 00:16:32,240 strict enforcement of companies using proprietary software. And so you can get fined, and you 181 00:16:32,240 --> 00:16:35,360 can have your office raided and the computers seized. 182 00:16:35,360 --> 00:16:39,589 The Software Publishers' Associations were able to push through much more draconian laws 183 00:16:39,589 --> 00:16:43,440 in the developing world. NGOs that were opposing the government in 184 00:16:43,440 --> 00:16:48,150 Russia, were also having their whole NGOs shut down, and their computers seized on the 185 00:16:48,150 --> 00:16:51,639 premise that they were using un-licensed software. It's a rule that you know you're breaking 186 00:16:51,639 --> 00:16:53,820 so they can come in and arrest you for it. 187 00:16:53,820 --> 00:16:57,620 The reason software monopolies happen so easily, 188 00:16:57,620 --> 00:17:05,730 without open source, is that software is fundamentally something you build on top of. So if you own 189 00:17:05,730 --> 00:17:12,670 the thing that you built on the top of, nobody else can really even begin to compete because 190 00:17:12,670 --> 00:17:18,810 they have to build the whole base, and it can take them five years to build what you 191 00:17:18,810 --> 00:17:24,020 have today. In the mean time during those five years, you built on top of yours and 192 00:17:24,020 --> 00:17:27,550 improved it, so it's very hard to really catch up. 193 00:17:27,550 --> 00:17:32,970 When your mail client is hosted at Google, Google could decide that they're going to 194 00:17:32,970 --> 00:17:37,021 suspend your account and then you disappear, and all of your information disappears. If 195 00:17:37,021 --> 00:17:41,940 we can't have free software to run all these services and if we are relying on sharecropping 196 00:17:41,940 --> 00:17:46,390 other people's software, and other people's technology, then they could always remove us 197 00:17:46,390 --> 00:17:50,550 and they could always make us disappear. Your digital life could be destroyed in a 198 00:17:50,550 --> 00:17:53,900 minute, and it could be a mistake. 199 00:17:56,700 --> 00:18:02,180 Free or open source software means that the code, documentation and discussions are available 200 00:18:02,190 --> 00:18:08,210 to the public. This allows anyone to better understand or help improve it. Wikipedia was 201 00:18:08,210 --> 00:18:14,060 created in 2001 and is a great example of free content, and a powerful software system 202 00:18:14,060 --> 00:18:18,520 to maintain it. Many people spend free time in online communities, 203 00:18:18,520 --> 00:18:22,020 whether to play games or improve public resources like Wikipedia. 204 00:18:22,020 --> 00:18:27,500 For a while, at least in the U.S., there was an encyclopedia that was printed, called 205 00:18:27,500 --> 00:18:32,690 the Britannica. I'm not sure why, it was an American encyclopedia, it wasn't called 206 00:18:32,690 --> 00:18:37,270 Americana, it was called Britannica, but anyway, Britannica was the biggest encyclopedia around. 207 00:18:37,270 --> 00:18:45,120 Then we created a little DVD / CD with something called Encarta, and we became the biggest 208 00:18:45,120 --> 00:18:51,360 encyclopedia company in the world! Now there's Wikipedia. 209 00:18:51,360 --> 00:18:56,320 Having lots of eyes, and lots and lots of minds being applied to something really allows 210 00:18:56,320 --> 00:19:00,520 it to evolve and become greater than the sum of its parts. I think this model and this 211 00:19:00,520 --> 00:19:05,330 idea can really be applied to just about anything. The Wikimedia Foundation is all about providing 212 00:19:05,330 --> 00:19:11,380 free content, and keeping the lights on for a whole host of different wiki projects including 213 00:19:11,380 --> 00:19:16,210 Wikipedia, Wikiquotes, Wikispecies, Wiktionary, and a bunch of other ones. 214 00:19:16,210 --> 00:19:21,680 The whole thing behind a wiki is just that it's a way that anyone can create content 215 00:19:21,680 --> 00:19:27,380 and share content. And it's a little bit different than say, just slapping up a web page, because 216 00:19:27,380 --> 00:19:32,290 anybody can come and edit it. And so there's this constant evolution of content that is 217 00:19:32,290 --> 00:19:35,020 going on. There have definitely been a number of studies 218 00:19:35,020 --> 00:19:39,790 done that have actually shown that Wikipedia is as accurate, if not more accurate in some 219 00:19:39,790 --> 00:19:43,430 cases, than encyclopedias like the Encyclopedia Britannica. 220 00:19:43,430 --> 00:19:47,831 Wikipedia itself is in fact an extremely important resource for mathematicians. A lot of the 221 00:19:47,831 --> 00:19:54,440 articles in Wikipedia about serious mathematical research topics, such as the Riemann Hypothesis, 222 00:19:54,440 --> 00:19:58,430 are in fact extremely good articles. It's really, really surprising how good the quality 223 00:19:58,430 --> 00:20:02,780 of Wikipedia articles has evolved. If you flip open a traditional encyclopedia and try 224 00:20:02,780 --> 00:20:06,710 to look for research mathematics you just kind of chuckle at how incomplete it is, but 225 00:20:06,710 --> 00:20:10,310 Wikipedia is really much, much better. Well, Wikipedia and all of these other wiki 226 00:20:10,310 --> 00:20:16,640 projects that the Wikimedia Foundation helps run, are all built on free software. It's 227 00:20:16,640 --> 00:20:22,220 a piece of free software called Mediawiki, which is the most popular and widely-used 228 00:20:22,220 --> 00:20:27,420 wiki platform that's out there. And even beyond just the software that powers 229 00:20:27,420 --> 00:20:33,000 all of this, there's this free and open source ethic that pervades the sort of culture 230 00:20:33,140 --> 00:20:39,500 behind Wikipedia. And that's this idea that we can collaborate together to build something better. 231 00:20:39,500 --> 00:20:44,490 And in Wikipedia's case that's building a body of knowledge, collecting the sum of all 232 00:20:44,490 --> 00:20:45,698 human knowledge. 233 00:20:45,698 --> 00:20:48,870 Wikis are powerful tools of collaboration, 234 00:20:48,870 --> 00:20:53,520 that require an operating system to run on. Wikipedia, along with billions of other servers 235 00:20:53,520 --> 00:20:55,600 and devices, depend on Linux. 236 00:21:02,460 --> 00:21:10,480 Another bold of lightning is that Apple + Microsoft equals 100% of the desktop computer 237 00:21:10,680 --> 00:21:12,820 market. 238 00:21:16,960 --> 00:21:20,420 The Linux kernel controls access to the processor, 239 00:21:20,420 --> 00:21:26,070 memory, and all the other devices. Surrounding the kernel is a lot of user mode software, 240 00:21:26,070 --> 00:21:33,000 to build a complete operating system which is also known as Linux, or sometimes GNU/Linux. 241 00:21:33,000 --> 00:21:40,040 It's kind of hard to explain a kernel, because the whole point of a kernel is to the hide 242 00:21:40,040 --> 00:21:48,050 the hardware details to the programs you actually use. So you never use a kernel. If a kernel 243 00:21:48,050 --> 00:21:53,530 is visible, there is something wrong. The kernel is handling the disk layout for you, 244 00:21:53,530 --> 00:22:00,380 the kernel is handling memory allocations, the kernel is handling CPU and networking, 245 00:22:00,380 --> 00:22:04,050 but it's not doing something that you actually interact with. 246 00:22:04,050 --> 00:22:11,790 I mean, it actually does everything you interact with, in the sense that it, for example, reads 247 00:22:11,790 --> 00:22:17,890 the keys you type in on the keyboard. But at the same time, it's not the thing that 248 00:22:17,890 --> 00:22:24,170 reacts to what you are actually typing. It's kind of this hidden infrastructure, the 249 00:22:24,170 --> 00:22:29,890 thing that you're supposed to ignore. If I'm on a bus or a subway, I wouldn't try 250 00:22:29,890 --> 00:22:37,460 to explain it, it's just too hard. The things I do every day are about looking up filenames 251 00:22:37,460 --> 00:22:44,500 and translating the filenames that normal applications use into the blocks on disk. 252 00:22:44,500 --> 00:22:48,440 Most normal people say: "What blocks on disk?" 253 00:23:06,260 --> 00:23:08,420 I've been a programmer since I was a kid, 254 00:23:08,420 --> 00:23:12,240 doing programming on different machines over the years. In college, my girlfriend came 255 00:23:12,240 --> 00:23:16,440 back from a class and said, "I heard this guy named Richard Stallman gave a talk, have 256 00:23:16,440 --> 00:23:19,620 you ever heard of him?" I was like: "No, what's that?" So I looked it up, and I learned all 257 00:23:19,740 --> 00:23:23,620 about the free software movement. After school, I went off and did embedded 258 00:23:23,620 --> 00:23:27,900 programming, and then one time at one job we were using a Unix called SCO Unix, and 259 00:23:27,900 --> 00:23:31,730 it wasn't really working as well, we had some problems with it, and someone said: "Hey, 260 00:23:31,730 --> 00:23:35,111 let's try something called Linux, it should work on there." And it did, so we used it 261 00:23:35,111 --> 00:23:39,801 in one of our products we made. And then I kind of just used it for a while, 262 00:23:39,801 --> 00:23:43,570 another job I was at, and I was like: "Hey, I can change this stuff." And I did, and Linus 263 00:23:43,570 --> 00:23:44,570 took my patches and I was like: "Hey, this is fun." 264 00:23:44,570 --> 00:23:47,400 And I started writing code and I realized that more people were using the code I wrote 265 00:23:47,400 --> 00:23:52,400 for free as a hobby, than anything I had ever been paid for. I wasn't working for very profitable 266 00:23:52,400 --> 00:23:55,930 companies. So then, over time, I started doing more and more, and then I got offered a job 267 00:23:55,930 --> 00:23:59,930 doing it, and I've doing it, getting paid for it ever since. 268 00:23:59,930 --> 00:24:06,200 Linux as an operating system works on so many platforms and has so many fewer bugs than 269 00:24:06,200 --> 00:24:11,240 other stuff because it has 10 million people who can view it. It has all these people who 270 00:24:11,240 --> 00:24:14,160 can see the problems, report bugs, and fix bugs. 271 00:24:14,160 --> 00:24:19,460 Making hardware work well is hard. Other operating systems like Apple has control over their 272 00:24:19,460 --> 00:24:22,570 hardware. So they can make their operating system work really, really well with their 273 00:24:22,570 --> 00:24:25,820 hardware. And that's great. Windows runs on every hardware, so they have the same problem 274 00:24:25,820 --> 00:24:30,330 we do with Linux in that it has to run everywhere. All the drivers for Linux are in the kernel. 275 00:24:30,330 --> 00:24:33,060 This is unlike other operating systems where all the drivers are outside, they come with 276 00:24:33,060 --> 00:24:37,370 the device. Because of that, we get to see the patterns that are met. When one device 277 00:24:37,370 --> 00:24:39,920 does the same things as another device, we can look at that and say: "Hey, wouldn't it 278 00:24:39,920 --> 00:24:43,350 be great if we did this in the core, in a common way, so that everybody gets it, and 279 00:24:43,350 --> 00:24:47,250 it works." So you don't make the same mistakes over and over and over. And then we say: "Oh 280 00:24:47,250 --> 00:24:50,070 look, these 4 drivers from these different companies can be merged into one." 281 00:24:50,070 --> 00:24:55,550 Microsoft and their way they did their OS is they have to keep the old backward compatibility 282 00:24:55,550 --> 00:24:59,930 to the driver interface. Because our drivers are in the kernel, we can change that interface 283 00:24:59,930 --> 00:25:03,580 and modify it based on what happens. They have to keep multiple versions of their 284 00:25:03,580 --> 00:25:07,050 USB stack in the kernel and maintain them over time. And that's their business decision, 285 00:25:07,050 --> 00:25:11,990 that's great, I understand their niche. Ours is re-write it, we make it smaller. So Linux 286 00:25:11,990 --> 00:25:16,700 driver are at least 1/3 smaller than other drivers for other operating systems. So just 287 00:25:16,700 --> 00:25:21,140 by pure size of the code, you have enhanced reliability. 288 00:25:21,140 --> 00:25:26,130 Free software ends up being able to much more quickly adapt to environments and fix bugs, 289 00:25:26,130 --> 00:25:30,620 because you don't have to wait for a single authority or a single person to get around 290 00:25:30,620 --> 00:25:35,240 to fixing those bugs. Overall Linux allows you to create an operating 291 00:25:35,240 --> 00:25:39,690 systems to build a product on. So like build cellphones, the Android guys did it. Palm, 292 00:25:39,690 --> 00:25:44,240 all those people did phones for a long time with Linux in it before people realized it. 293 00:25:44,240 --> 00:25:47,620 Supercomputer guys built giant supercomputers in ways they wanted to and they took Linux 294 00:25:47,620 --> 00:25:52,060 and modified it to make it work. Google built a giant cluster out there. 295 00:25:52,060 --> 00:25:56,350 Amazon's cloud computing initiative, what they've done with Linux is just amazing, things 296 00:25:56,350 --> 00:26:06,420 that they couldn't do with any other operating system because they didn't have access to the source. 297 00:26:06,440 --> 00:26:09,520 I remember when I was 17 or 18 and, you know, 298 00:26:09,520 --> 00:26:18,871 long hair, Iron Maiden t-shirt, living with my parents, full of acne and insecurity, and 299 00:26:18,871 --> 00:26:26,300 my brother came to stay with me for two weeks. And, of course we hated each other. I was 300 00:26:26,300 --> 00:26:31,060 using Windows one day, and was getting frustrated with it and he basically said: "Why are using 301 00:26:31,060 --> 00:26:35,590 that thing? It's Mickey Mouse." And he was a sys-admin, and he said you want 302 00:26:35,590 --> 00:26:39,280 to use this thing called Linux. I had never heard of it. 303 00:26:39,280 --> 00:26:44,110 I said: "What's that?" "It's Unix." Which I had heard of is this big complicated thing 304 00:26:44,110 --> 00:26:45,720 that people with beards tended to use. 305 00:26:45,820 --> 00:26:47,680 I worked in a bookshop at the time and he said you 306 00:26:47,720 --> 00:26:49,880 can get it on the back of a book and it's all free. 307 00:26:49,880 --> 00:26:55,410 And so we went and got it, and he installed it for me, and it literally took two weeks, 308 00:26:55,410 --> 00:26:59,950 and he actually used the soldering iron because there was something wrong with my hard drive. 309 00:26:59,950 --> 00:27:01,840 He had it all open, sprawled all over the thing. 310 00:27:01,840 --> 00:27:07,030 And when it was finally on there, it's all text, and he just left me the username and 311 00:27:07,030 --> 00:27:11,720 the password to login with, and then he left the following day. And I was stuck there with 312 00:27:11,720 --> 00:27:17,760 this huge book, and this post-it note, and I logged in and I had no idea what was going on. 313 00:27:17,760 --> 00:27:20,660 So I opened up the book and I started reading. 314 00:27:20,660 --> 00:27:25,720 And a lot of the technical mumbo-jumbo I didn't understand, it was too technical for me, but 315 00:27:25,720 --> 00:27:30,600 I remember reading the introductory section about community and how the software was put 316 00:27:30,600 --> 00:27:36,820 together by the volunteers all over the world who share their contributions into one pool, 317 00:27:36,820 --> 00:27:41,380 and I just thought that was fascinating. At the time the Internet was fascinating, 318 00:27:41,380 --> 00:27:45,720 but the Internet was very much something you consume content from, you went to websites, 319 00:27:45,720 --> 00:27:49,700 but with this you were interacting with people and you are creating something together and 320 00:27:49,700 --> 00:27:54,880 I thought that was really interesting. The mechanics of how this thing was created 321 00:27:54,880 --> 00:28:00,390 could pave the way for something even better to be created to solve the problems that were 322 00:28:00,390 --> 00:28:03,850 present in the system. The opportunity for people to work together 323 00:28:03,850 --> 00:28:09,130 around a common mission, a common goal, is what's made all of this possible. 324 00:28:09,130 --> 00:28:15,460 You can't easily go into a computer store, buy a computer and Linux is just there. You 325 00:28:15,460 --> 00:28:20,610 can do it in some places, Dell will sell you Linux pre-installed on their machines, but 326 00:28:20,610 --> 00:28:25,940 the common case is not that. So even to get into Linux, normally you'd 327 00:28:25,940 --> 00:28:30,590 have to decide to install your own operating system. And at that point, most people say: 328 00:28:30,590 --> 00:28:33,800 "What, install my own operating system? No thank you." 329 00:28:33,800 --> 00:28:39,880 Linux has got this potential, this great ability to grow very quickly, with everybody working 330 00:28:39,880 --> 00:28:46,000 on it. And Linux is a great example of open source, and it's probably one of the best 331 00:28:46,000 --> 00:28:50,500 originating ones, but that's spreading to a large number of things. 332 00:28:56,520 --> 00:29:01,460 The web is built mostly on Linux. Free and powerful software creates a market for people 333 00:29:01,460 --> 00:29:03,760 who can customize it for others. 334 00:29:12,060 --> 00:29:14,300 I started off as a print designer, and the 335 00:29:14,300 --> 00:29:18,760 web thing kind of came about, I knew absolutely nothing about programming, and really never 336 00:29:18,760 --> 00:29:22,190 wanted to. I decided to kind of jump in the game, decided 337 00:29:22,190 --> 00:29:27,720 to jump in with PHP, kind of learn that. These open source platforms kind of came about, 338 00:29:27,720 --> 00:29:31,100 WordPress being one. And you jump in there, sure there's a little 339 00:29:31,100 --> 00:29:35,060 bit of a learning curve, but if you have a question about anything, and I never claimed 340 00:29:35,060 --> 00:29:38,670 to be the greatest programmer whatsoever, but if you have a question about anything, 341 00:29:38,670 --> 00:29:40,880 or how to do something, the answer is out there. 342 00:29:40,880 --> 00:29:45,950 Their Codex, their manual, is super searchable, all the information is there, that allows 343 00:29:45,950 --> 00:29:51,460 a person like myself, who knew nothing about coding, to sort of jump right into that arena, 344 00:29:51,460 --> 00:29:57,300 and actually now just build some pretty robust, amazing sites. 345 00:30:06,040 --> 00:30:12,880 A kernel, graphical interface, and web browser, are the basic software for a desktop computer. 346 00:30:13,180 --> 00:30:18,080 But for everyone from secretaries to CEOs, the office productivity applications are the 347 00:30:18,090 --> 00:30:20,610 most important. There are billions of documents out there 348 00:30:20,610 --> 00:30:25,710 in a file format so arcane that only Microsoft Office can fully understand them. LibreOffice, 349 00:30:25,710 --> 00:30:30,120 which is based on OpenOffice, is the best free application suite, but it's still unknown 350 00:30:30,120 --> 00:30:32,400 to the average person. 351 00:30:32,400 --> 00:30:36,630 Microsoft has some assets like Office that 352 00:30:36,630 --> 00:30:44,450 have stayed strong, and there's so much room for innovation in those. So, you know, 353 00:30:44,450 --> 00:30:50,360 that's probably the most amazing asset that Microsoft has is that, globally the way people 354 00:30:50,400 --> 00:30:57,040 get productivity with Outlook, Word. You know, they've maintained that position, and there's 355 00:30:57,040 --> 00:31:01,160 so much more that they are doing, and you know I get to push a little bit and share 356 00:31:01,160 --> 00:31:04,700 ideas in my part-time work. 357 00:31:11,960 --> 00:31:14,520 Microsoft Office is critical to the Windows ecosystem. 358 00:31:14,520 --> 00:31:17,490 Without Office, Windows is pretty useless. You can play solitaire on it, of 359 00:31:17,490 --> 00:31:22,480 course nowadays you can browse the web, but Office brings so much more, particularly for 360 00:31:22,480 --> 00:31:25,400 a business, for analytics. You can not just write letters to your family, 361 00:31:25,400 --> 00:31:30,170 you can fill in government forms, you can model the world. You can take reality and 362 00:31:30,170 --> 00:31:35,270 crunch it, in number form in your spreadsheet and see, what we should do? Lots of spreadsheets 363 00:31:35,270 --> 00:31:39,360 are used to make business decisions. "Should we make more black cars or green cars? Which 364 00:31:39,360 --> 00:31:41,340 is selling, what should we do? Should we do a marketing campaign?" 365 00:31:41,340 --> 00:31:45,650 We can get very excited about the wobbly windows, and the funny effects, and the glass stuff, 366 00:31:45,650 --> 00:31:49,020 but at the end of the day, Office is where the value is. 367 00:31:49,020 --> 00:31:53,250 So from a code perspective, LibreOffice is around half the size of the Linux kernel. 368 00:31:53,250 --> 00:31:57,300 I would argue from a complexity perspective we are significantly more complicated. 369 00:31:57,300 --> 00:32:01,060 If you go to universities, they'll teach you courses on how to implement a kernel, or how 370 00:32:01,060 --> 00:32:06,310 to implement a compiler, and that's well understood. It's less well understood how to implement 371 00:32:06,310 --> 00:32:10,860 an office suite, and it's significantly more complicated, and the code is much more coupled. 372 00:32:10,860 --> 00:32:14,640 Linus Torvalds would say that the Linux kernel is kind of a done problem. "Go and hack somewhere 373 00:32:14,640 --> 00:32:16,920 else." LibreOffice is an extremely challenging codebase, 374 00:32:16,920 --> 00:32:21,940 but it's an extremely good one for students and people wanting to get involved with professional 375 00:32:21,940 --> 00:32:24,970 development. Vendor lock-in is a big problem because what 376 00:32:24,970 --> 00:32:30,320 happens is, you start a business, you install a ton of Windows machines, you put Microsoft 377 00:32:30,320 --> 00:32:36,070 Office on those machines, you save your documents in Microsoft Word format, and you are always 378 00:32:36,070 --> 00:32:39,760 going to need Microsoft Word format to open them again. Which means you are always going 379 00:32:39,760 --> 00:32:42,810 to have to pay for it. The problem with a proprietary format is really 380 00:32:42,810 --> 00:32:47,760 one of vendor control. That vendor gets to choose exactly what goes into the standard, 381 00:32:47,760 --> 00:32:51,440 and what goes into the standard is really some kind of silhouette of their office software. 382 00:32:51,440 --> 00:32:56,400 So what ends up in there is all of this detail around the edges of what is perhaps 20 years 383 00:32:56,400 --> 00:33:01,950 of accumulated baggage. It looks like this vast mound of un-navigable stuff that just 384 00:33:01,950 --> 00:33:05,721 ends in there if you're not careful. OpenDocument is a better format because it 385 00:33:05,721 --> 00:33:09,880 reuses existing standards and was created by a diverse community. 386 00:33:09,880 --> 00:33:14,040 If you look at the genesis of ODF and DOCX, I think they're both rather different. 387 00:33:14,040 --> 00:33:19,250 So ODF is really the OpenDocument format. It was produced by multiple vendors, working 388 00:33:19,250 --> 00:33:24,090 on producing a real standard that can be used for multiple implementations, so I think there's 389 00:33:24,090 --> 00:33:27,711 a degree of compromise in there, there's a degree of working together with other people 390 00:33:27,711 --> 00:33:30,790 actually producing suites, and actually changing the standard. 391 00:33:30,790 --> 00:33:34,121 That's critically important, that we open the market for office software so other people 392 00:33:34,121 --> 00:33:37,920 can provide suites, and they can compete against each other in the common good. 393 00:33:37,920 --> 00:33:42,030 LibreOffice grew out of some frustrations with the OpenOffice project. There were many 394 00:33:42,030 --> 00:33:46,030 problems with that. I think, it's well described as a problem for every solution. 395 00:33:46,030 --> 00:33:49,460 When Oracle acquired Sun, we were very hopeful that things would improve. Sadly, that didn't 396 00:33:49,460 --> 00:33:55,010 occur, and so we created LibreOffice to really actually start doing those things that we 397 00:33:55,010 --> 00:33:59,820 knew we needed to do to make LibreOffice more relevant, improve quality, to make it more 398 00:33:59,820 --> 00:34:02,700 viable, and better inter-operability too. 399 00:34:09,280 --> 00:34:13,380 Free software is a real enabler of local business because, 400 00:34:13,380 --> 00:34:17,180 not only are you going to need help migrating to LibreOffice, some help, and Collabora 401 00:34:17,180 --> 00:34:22,260 provides those services and has partners that do that, that brings money into the local economy. 402 00:34:22,260 --> 00:34:26,840 Local companies can exploit niches and provide services that are just not possible 403 00:34:26,849 --> 00:34:30,069 for a large company to provide. We have some amazing volunteers that have 404 00:34:30,069 --> 00:34:33,839 contributed, just fantastic features that people see in the millions and appreciate. 405 00:34:33,839 --> 00:34:38,319 From a community perspective, that's relatively easy, that's the easy part. Branding is 406 00:34:38,320 --> 00:34:44,260 a different problem. We have a significant brand problem getting the word out about LibreOffice. 407 00:34:47,900 --> 00:34:52,019 Donations to the foundation responsible for LibreOffice are much smaller than the amount 408 00:34:52,020 --> 00:34:57,760 of money Microsoft receives for Office. What LibreOffice needs is more money and programmers. 409 00:35:05,080 --> 00:35:09,280 Neil Armstrong may have been the first one to step on the moon, but Margaret Hamilton 410 00:35:09,460 --> 00:35:13,020 was the lead engineer for the software that got him there. 411 00:35:13,020 --> 00:35:18,410 Back then, programming was considered secretarial work, mostly done by the organized women, 412 00:35:18,410 --> 00:35:22,570 while piloting of the dangerous aircraft was done by the more expendable men. 413 00:35:22,570 --> 00:35:26,920 Today, there are special events to teach programming, but it isn't part of the daily classroom. 414 00:35:26,920 --> 00:35:32,220 Math is taught with pencil and paper, as it has been for hundreds of years. 415 00:35:32,360 --> 00:35:36,180 Computers can be excellent teachers because they allow children to explore and work at 416 00:35:36,180 --> 00:35:41,280 their own pace. Python is a programming language that 8 year olds can start writing, and then 417 00:35:41,280 --> 00:35:43,520 spend a lifetime mastering. 418 00:35:45,160 --> 00:35:47,900 Few schools teach Python today. Instead, many 419 00:35:47,900 --> 00:35:54,860 kids are taught visual languages they will quickly outgrow and aren't used in the real world. 420 00:35:57,460 --> 00:36:00,400 Using Python and the Sage Math Library, we 421 00:36:00,400 --> 00:36:03,340 can create a new form of math and science education. 422 00:36:11,880 --> 00:36:16,060 It's weird because every single thing you do as an educator involving mathematics at 423 00:36:16,070 --> 00:36:20,660 least is opened except possibly the software that you use to illustrate something. 424 00:36:20,660 --> 00:36:24,900 If you are in a class and you're showing something or describing some theorem or some bit of 425 00:36:24,900 --> 00:36:29,420 mathematics and a student asks you: "Why this?", "Why that?" "What's going on?", "Who did this?", 426 00:36:29,420 --> 00:36:33,150 every single question you can answer but if they ask you how does that program draw that 427 00:36:33,150 --> 00:36:37,349 picture, compute that integral, or whatever, then if you are using a proprietary system, 428 00:36:37,349 --> 00:36:41,910 you are kind of shut off from finding out what the answer is, which seems really strange. 429 00:36:41,910 --> 00:36:46,099 Some people like me are extremely annoyed by this and don't want mathematics to be like 430 00:36:46,099 --> 00:36:50,829 this for the kids of the future. And other people are perfectly happy with it. And so 431 00:36:50,829 --> 00:36:54,940 there is this sort of conflict. It's really sad when the field for doing 432 00:36:54,940 --> 00:36:58,910 mathematics is not very level, where there are very smart people all over the world, 433 00:36:58,910 --> 00:37:03,890 kind of equally distributed, but the access to resources for kind of stupid reasons aren't 434 00:37:03,890 --> 00:37:06,810 equally distributed. In mathematics, it's all about trying to 435 00:37:06,810 --> 00:37:11,269 understand something more deeply. You have some basic thing like some fact, and you want 436 00:37:11,269 --> 00:37:14,750 to understand it. The more you understand it, you get really excited, your pulse quickens, 437 00:37:14,750 --> 00:37:18,950 and you just feel like you're accomplishing something if you understand that fact more and more, 438 00:37:18,950 --> 00:37:22,390 and by having access to the source code that lets you compute with some mathematical object, 439 00:37:22,390 --> 00:37:25,460 you can dramatically deepen your understanding of that object. 440 00:37:25,460 --> 00:37:29,400 You can add little bits of extra code in there and see what happens, it's really, really 441 00:37:29,400 --> 00:37:32,809 fun. Maybe they really want to look under the hood 442 00:37:32,809 --> 00:37:37,180 and see what's really going on with the code. There's some very easy way, when you 443 00:37:37,180 --> 00:37:42,240 are using Sage, to see the underlying source code that's being run when you ask for a calculation. 444 00:37:42,440 --> 00:37:46,720 "Freedom to tinker" is an expression that you'll hear a lot. You know, I want to be 445 00:37:46,730 --> 00:37:50,500 able to open up a device and see how it works and play with it, and make it better. 446 00:37:50,500 --> 00:37:54,170 Kids were getting consoles, you know, it's a black box, you sit down and play games: 447 00:37:54,170 --> 00:37:58,222 "You should be a consumer of media rather than be a producer of media, rather than someone 448 00:37:58,222 --> 00:38:01,490 that makes things." And there's this great movement about, sort 449 00:38:01,490 --> 00:38:06,990 of in the free software movement about making things and tinkering. That's really what makes 450 00:38:06,990 --> 00:38:08,700 the engineers and scientists of tomorrow. 451 00:38:08,700 --> 00:38:10,289 We can hack into this mainframe. 452 00:38:10,289 --> 00:38:15,980 I love that video games have finally folded in on themselves. We've gone from the action used to be: 453 00:38:16,000 --> 00:38:23,720 "Slay the dragon", now it's: "Let's go into a basement and play with computers and servers!" 454 00:38:23,730 --> 00:38:26,670 We often see in schools, that people teach 455 00:38:26,670 --> 00:38:31,299 programming or IT, but it's actually just typing into a word processor or spreadsheet, 456 00:38:31,299 --> 00:38:34,910 you're not actually learning the key skills you need to program, and so you don't learn 457 00:38:34,910 --> 00:38:39,030 to understand computers, which are so vital in the modern world. Everything's driven by 458 00:38:39,030 --> 00:38:41,440 software, if our children don't understand it, we're doomed! 459 00:38:41,440 --> 00:38:45,630 I've always been an entrepreneur. When I was younger, I was selling candy door to door 460 00:38:45,630 --> 00:38:49,650 and all that type of stuff. And really, figuring out programming and learning web development 461 00:38:49,650 --> 00:38:55,279 from the very start was: "Woah, I can build businesses by myself you know, just sitting 462 00:38:55,279 --> 00:38:59,300 in front of a computer." The idea that the Internet gives us the ability to connect to 463 00:38:59,300 --> 00:39:02,940 anybody now. You can build on top of that platform, 464 00:39:02,940 --> 00:39:06,720 you can create your own business, and the way you do that is through programming. It is 465 00:39:06,720 --> 00:39:11,161 very easy to do. And that's why a lot of these Hackathons and these startup weekends and 466 00:39:11,161 --> 00:39:14,339 all these events that are now. These kids get together and they'll just spend, 467 00:39:14,339 --> 00:39:17,660 you know, a few hours working on something, and something can come out of it. 468 00:39:17,660 --> 00:39:21,650 In the digital age, communities can create things together without being in the same 469 00:39:21,650 --> 00:39:26,240 location, but where's the fun in that? Conferences, hackathons, and other events 470 00:39:26,240 --> 00:39:31,000 are taking place all over the world, and they are a great way to meet people and exchange 471 00:39:31,000 --> 00:39:34,130 ideas. We're at MHacks, this is the 3rd event that 472 00:39:34,130 --> 00:39:37,009 we've held. We call it the most epic college Hackathon. 473 00:39:37,009 --> 00:39:41,359 Programming in the classroom doesn't necessarily teach you all the stuff that you need to know 474 00:39:41,359 --> 00:39:45,800 to actually go out and either get a job or contribute to a project. What you learn in 475 00:39:45,800 --> 00:39:50,609 the classroom is: "This algorithm is faster than that one," and stuff like that, and that's 476 00:39:50,609 --> 00:39:55,520 definitely useful, but to actually be able to go out and make things, is totally different. 477 00:39:55,520 --> 00:40:01,420 Hacking for me is, has nothing to do with cyber-hacking or hacking someone's Facebook 478 00:40:01,420 --> 00:40:06,299 or information, hacking for a hacker, like those who come to MHacks, is all about being 479 00:40:06,299 --> 00:40:11,850 able to learn very quickly, being able to produce a solution to a problem really, really, 480 00:40:11,850 --> 00:40:15,900 quickly, and basically hacking together a solution to a problem. 481 00:40:15,900 --> 00:40:20,880 The most fun thing about programming for me is just the amount of power that I feel that 482 00:40:20,880 --> 00:40:25,040 I have control over: everything on the Internet. You know, I see a bunch of data, and I can 483 00:40:25,040 --> 00:40:29,119 scrape it, and I can use it, I can analyze it. I can put it on a website, and put that 484 00:40:29,120 --> 00:40:34,100 out to display to people. I can write whatever I want to and it costs me nothing. 485 00:40:34,100 --> 00:40:39,400 With programming, I can literally just open up my computer and start learning whatever 486 00:40:39,480 --> 00:40:44,200 I need to learn, and create whatever I want to create. It's really hard for me to find 487 00:40:44,200 --> 00:40:50,640 a limit to the things you can do just with your own, curiosity and drive. 488 00:40:50,640 --> 00:40:56,360 Programming is kind of seen as an elite thing, and it shouldn't be, because programming 489 00:40:56,369 --> 00:41:02,240 is going to be akin to driving in the future. Everything has a computer now, and that's 490 00:41:02,240 --> 00:41:06,570 not going to go away. So you need to know how to use your stuff, how to use your tools. 491 00:41:06,570 --> 00:41:11,730 A couple of years ago I tried learning C++ on my own, and it was actually really difficult, 492 00:41:11,730 --> 00:41:15,890 I didn't know what I was doing. But then I started trying Python, and it's super simplistic 493 00:41:15,890 --> 00:41:20,710 which I was actually really surprised with. So, I feel like, now that I'm learning Python, 494 00:41:20,710 --> 00:41:23,440 I might actually be able to get into programming. 495 00:41:23,440 --> 00:41:27,430 I like Python a lot because it's free as in libre, 496 00:41:27,430 --> 00:41:32,640 free as in speech, so there aren't restrictive licenses attached to it. It's 497 00:41:32,640 --> 00:41:37,270 not platform specific, so you can, write it on Linux, and then run it on Windows. 498 00:41:37,270 --> 00:41:42,000 And, you know, all my friends use different operating systems, so it's super easy to collaborate 499 00:41:42,160 --> 00:41:44,960 regardless. And it's an interpreted language, so it's 500 00:41:44,960 --> 00:41:51,859 easy to debug, because you can run a few lines in a terminal, and call methods individually 501 00:41:51,859 --> 00:41:56,289 without having like to compile, and make sure all of your stuff is in order for the entire 502 00:41:56,289 --> 00:41:58,920 thing. You can just run what you need to. 503 00:42:01,280 --> 00:42:04,100 Python is a very general, mainstream language, 504 00:42:04,109 --> 00:42:08,750 certainly one of the top 5 or 10 most popular interpreted programming languages, and you 505 00:42:08,750 --> 00:42:13,040 can do a wide range of things in it. You can write a web server, you can write 506 00:42:13,040 --> 00:42:17,690 a game, you can write code for doing mathematics. You can do a very wide range of things, and 507 00:42:17,690 --> 00:42:22,140 it's not really aimed at any one particular domain, it's very, very general, which is 508 00:42:22,140 --> 00:42:27,990 great because it attracts a huge range of users, and it attracts a wide range of different 509 00:42:27,990 --> 00:42:31,760 people and companies that like to contribute to it, making it better. 510 00:42:35,280 --> 00:42:40,579 So I did program when I was 8, 9, 10 years old. I spent a lot of time, you know in Fred 511 00:42:40,579 --> 00:42:43,430 Meyer, in their computer section, which they don't have anymore, programming in BASIC 512 00:42:43,430 --> 00:42:45,990 and learning to program in fact from other kids. 513 00:42:45,990 --> 00:42:50,930 The language we used then is very similar and immediate like Python is now so I can 514 00:42:50,930 --> 00:42:55,839 easily imagine 8 year-olds today using Python in a similar capacity especially via their 515 00:42:55,839 --> 00:43:00,400 cellphones or over the web or whatever, and they really have a full, very powerful, very 516 00:43:00,400 --> 00:43:03,080 general purpose programming language at their disposal. 517 00:43:13,960 --> 00:43:19,340 Python is a programming language where you tell a computer to do something. 518 00:43:19,359 --> 00:43:26,059 Computers are dumb, and you have to tell them what to do, you have to make sure everything 519 00:43:26,059 --> 00:43:29,369 is in order, and you always have to double-check your work. 520 00:43:29,369 --> 00:43:36,460 We've learned how to write little pieces of code and make them one big thing. 521 00:43:36,460 --> 00:43:43,359 Programming can be hard because there can be a lot of lines of code, and it takes time 522 00:43:43,359 --> 00:43:49,140 to type all of them. You can make all sorts of shapes, or you can 523 00:43:49,140 --> 00:43:55,600 even draw cars and a bunch of different things, you can do anything basically that your mind 524 00:43:55,600 --> 00:44:01,609 can think of. We do programming because it's the most interesting 525 00:44:01,609 --> 00:44:12,900 thing you can do. It's cerebral, and you actually build something practical. So, it's not like 526 00:44:12,900 --> 00:44:17,260 playing chess, there's an end result, and you build something big and interesting. 527 00:44:17,440 --> 00:44:22,791 I would do this even if I wasn't paid, but the fact that it's important enough to other people 528 00:44:22,800 --> 00:44:27,980 that they want to pay me to make sure that I can do it full-time, I'll take it. 529 00:44:35,260 --> 00:44:40,040 Many people are frustrated with current technology. The fight for freedom continues to the next 530 00:44:40,060 --> 00:44:45,440 generation of people and devices. The trends are not positive. 531 00:44:45,680 --> 00:44:51,380 Digital Rights Management or DRM controls the use of content. DRM restricts DVDs and 532 00:44:51,390 --> 00:44:57,990 Blu-Ray, e-books, cable TV, web video, and even coffee machines and tractors. DRM is 533 00:44:57,990 --> 00:45:02,680 fundamentally insecure while adding costs and difficulties for both sides - consumers 534 00:45:02,680 --> 00:45:05,550 and creators. There's nothing free about Android, as we 535 00:45:05,550 --> 00:45:09,770 certainly have asserted in a number of cases, there's an intellectual property royalty 536 00:45:09,770 --> 00:45:14,220 due on that, whether they happen to charge for their software or not is their business decision. 537 00:45:17,240 --> 00:45:19,400 The Android ecosystem took off because of 538 00:45:19,410 --> 00:45:24,540 all the free software already available. Even so, everyone including Google builds a lot 539 00:45:24,540 --> 00:45:28,890 of proprietary code on top. Therefore, Android customers complain about 540 00:45:28,890 --> 00:45:35,000 the unwanted add-on bloatware, bugs, and discontinued support for older devices. 541 00:45:35,000 --> 00:45:41,589 The Central Intelligence Agency has lost control of highly classified documents and computer 542 00:45:41,589 --> 00:45:48,269 code that describe top secret techniques for intercepting phone calls, reading encrypted 543 00:45:48,269 --> 00:45:54,060 text messages, and turning televisions into eavesdropping devices. 544 00:45:54,060 --> 00:45:58,520 The Wikileaks Vault 7 release showed that governments and hackers can take control of 545 00:45:58,529 --> 00:46:04,730 nearly every device on the planet. If a cyber-weapon is stolen, it can be exploited by anyone until 546 00:46:04,730 --> 00:46:08,260 it's fixed, which can be years later. 547 00:46:10,580 --> 00:46:12,580 Since the creation of the automobile, people 548 00:46:12,580 --> 00:46:19,260 have dreamt of cars that can drive themselves. Any camera can record pictures of the road, 549 00:46:19,260 --> 00:46:22,980 but it takes artificial intelligence to interpret and safely navigate it. 550 00:46:22,980 --> 00:46:27,140 Creating autonomous vehicles that can handle bad weather and unexpected pedestrians, is 551 00:46:27,150 --> 00:46:31,720 a difficult challenge. Most corporations and public universities are working in secret, 552 00:46:31,720 --> 00:46:36,190 re-inventing the proverbial wheel. We tried to interview Ford and GM engineers 553 00:46:36,190 --> 00:46:41,630 to ask basic questions, but many details are confidential. This secretive culture impedes 554 00:46:41,630 --> 00:46:48,190 progress, and leads to less reliability. The good news is that it requires extremely 555 00:46:48,190 --> 00:46:52,769 simple software to let us take a nap on highways in good driving conditions. Companies like 556 00:46:52,769 --> 00:46:58,170 Delphi Automotive demonstrated this years ago. 557 00:46:58,170 --> 00:47:10,880 As robotic cars become reality, flying taxis and other fun toys will appear. 558 00:47:10,880 --> 00:47:15,080 Vision and control software are the biggest challenge in building a robotic butler or 559 00:47:15,099 --> 00:47:20,390 maid that can cook and iron clothes. With a few thousand dollars for hardware, 560 00:47:20,390 --> 00:47:24,660 you could build a plastic robot as well-trained as Alfred from Batman, and even able to learn 561 00:47:24,660 --> 00:47:29,130 personal preferences. Software and technology are rapidly changing 562 00:47:29,130 --> 00:47:33,600 many aspects of life, but it has personal implications for healthcare. 563 00:47:33,600 --> 00:47:38,140 Everybody here goes to a radiologist, and they take an X-ray, and so forth, and they say: 564 00:47:38,150 --> 00:47:41,630 "Oh, you are fine." I would much rather have the computer analyze 565 00:47:41,630 --> 00:47:45,869 the radiology picture and then advise the doctor, and I would pay a premium for that, 566 00:47:45,869 --> 00:47:47,829 because that doctor is going to get a better result. 567 00:47:47,829 --> 00:47:54,309 The reason I mention radiology is that the computer vision stuff is better now than human vision 568 00:47:54,309 --> 00:47:58,089 In an of itself, that's a pretty big deal. 569 00:47:58,089 --> 00:48:04,150 What do you do with vision, aside from exist, right? Dermatology, radiology, things which 570 00:48:04,150 --> 00:48:09,619 involve vision skills, these are relatively automatable, and that doesn't mean the dermatologist 571 00:48:09,619 --> 00:48:15,130 goes away, but it means that he or she becomes efficient, does a better job, and is more 572 00:48:15,130 --> 00:48:19,420 accurate. So you should expect that in the next decade, 573 00:48:19,420 --> 00:48:25,079 you'll use tools with humans that make them more efficient, make you safer, and more productive. 574 00:48:25,079 --> 00:48:28,220 I think it's speculation what happens after that. 575 00:48:28,220 --> 00:48:33,170 I did all this research about the ways that software is reviewed on these devices, and 576 00:48:33,170 --> 00:48:37,440 what the process is at the FDA. I found that in fact that the software is not really being reviewed 577 00:48:37,440 --> 00:48:44,080 that much. Now, anyone who is a software developer knows that all software has bugs. 578 00:48:44,400 --> 00:48:48,080 Things that I found about the FDA that were quite, I say shocking but it's a terrible 579 00:48:48,080 --> 00:48:54,950 pun, were that the FDA doesn't actually review the software on most of it's implantable devices 580 00:48:54,950 --> 00:48:59,089 that it reviews, it only reviews the software if there's a problem. 581 00:48:59,089 --> 00:49:03,940 They don't also request the software because they're not reviewing it, so they don't maintain 582 00:49:03,940 --> 00:49:09,450 a repository of it. So in like a disaster scenario, the FDA or the government 583 00:49:09,450 --> 00:49:14,380 doesn't even have a copy of it. In a catastrophic situation, if you're relying on one company 584 00:49:14,380 --> 00:49:18,430 to fix your heart device if there's a problem, if it's vulnerable, then you are really out 585 00:49:18,430 --> 00:49:23,160 of luck. If it's free and open source software then anyone can see the source code, and anyone 586 00:49:23,160 --> 00:49:25,920 can make a patch. What we have in medical device industry, or 587 00:49:25,920 --> 00:49:31,099 at least with my pacemaker, is the worst of both worlds. We have devices that have closed 588 00:49:31,099 --> 00:49:35,859 software, it's proprietary software so people can't review it broadly, but we also have 589 00:49:35,859 --> 00:49:39,819 no encryption. So most of these devices are broadcasting 590 00:49:39,819 --> 00:49:45,559 signals all the time and there is no encryption protecting us from having other people hack 591 00:49:45,559 --> 00:49:47,869 into them. This was a think tank that actually were able 592 00:49:47,869 --> 00:49:51,700 to hack into the device using just regular, over the counter equipment. They were be able 593 00:49:51,700 --> 00:49:55,940 to deliver unwanted shocks, they were be able to put the device in test mode to run down 594 00:49:55,940 --> 00:49:59,240 battery, and these devices are only good as their batteries are. 595 00:49:59,240 --> 00:50:05,520 It's pretty terrifying how vulnerable these devices are and it's this really kind of this 596 00:50:05,560 --> 00:50:10,050 perverse result, because you think that it's more vulnerable if you release the code for 597 00:50:10,050 --> 00:50:14,319 people to review. It's a little counter-intuitive, but it's actually the reverse. We should publish 598 00:50:14,319 --> 00:50:18,829 the code so everyone can see it, and make sure that it's safe and good, but we should 599 00:50:18,829 --> 00:50:23,359 put real encryption on it, real protections, real passwords, so that we know that these 600 00:50:23,359 --> 00:50:29,140 devices are not vulnerable to attacks. I want to be able to review the source code 601 00:50:29,140 --> 00:50:34,220 on my device. If I'm not a programmer, I want my friends to be able to do it. I want 602 00:50:34,220 --> 00:50:38,250 the think-tank to do it. I want the university to it. I want the world to be able to look 603 00:50:38,250 --> 00:50:41,359 at this software and make sure that it is safe. 604 00:50:41,359 --> 00:50:48,259 But that same logic not only goes for implanted medical devices, it goes for our cars, it goes 605 00:50:48,259 --> 00:50:53,570 for our voting machines and our stock markets, anything we are relying in our lives and for 606 00:50:53,570 --> 00:50:58,849 society should be reviewable by everyone. Life is very complex. It has evolved over 607 00:50:58,849 --> 00:51:03,960 millions of years, developing all kinds of systems to keep us alive. There are many different 608 00:51:03,960 --> 00:51:08,599 analysis techniques in order to investigate a living being, and they all generate a huge 609 00:51:08,600 --> 00:51:10,920 amount of data to be analyzed. 610 00:51:18,000 --> 00:51:21,119 Biology has become as much of a computational 611 00:51:21,119 --> 00:51:24,460 science as a traditional lab science that it used to be. 612 00:51:24,460 --> 00:51:29,270 Once upon a time, people used to spend their entire PhDs studying one gene. Nowadays we 613 00:51:29,270 --> 00:51:34,619 have these high throughput methods that allow us to investigate, a whole genome, 10s of 614 00:51:34,619 --> 00:51:39,220 thousands of genes, in one experiment. On the other hand, various labs, people generally, 615 00:51:39,220 --> 00:51:44,559 tend to focus on one particular angle, one particular technology, for instance they can 616 00:51:44,559 --> 00:51:49,480 do proteomics, or genomics, or metabolomics, whereas in order to really understand a complex 617 00:51:49,480 --> 00:51:51,153 problem, you need to look at it from many, many angles. 618 00:51:51,153 --> 00:51:57,930 Imagine you are waging let's say a war against a disease, you need different types of weapons. 619 00:51:57,930 --> 00:52:07,220 So it's absolutely crucial to have resources that gather data of various types in one place. 620 00:52:07,220 --> 00:52:13,580 The National Center for Biotechnology Information, or NCBI, is the largest resource for information 621 00:52:13,589 --> 00:52:19,160 about health sciences. Inside the NCBI are archives with lots of data about many species 622 00:52:19,160 --> 00:52:23,829 of life. The NCBI, being a government agency, that 623 00:52:23,829 --> 00:52:31,359 is able to provide a variety of great access to scientists in a variety of different, pieces 624 00:52:31,359 --> 00:52:39,269 of publications, they have access to genomic and proteomic data, and the best part about 625 00:52:39,269 --> 00:52:45,980 all of this is that it gives this, idea a sense of community, so that anything that 626 00:52:45,980 --> 00:52:52,579 gets discovered can be publicly available for other scientists to look at, to examine 627 00:52:52,579 --> 00:52:58,359 and to further their own research. It's hard to believe that just two months 628 00:52:58,359 --> 00:53:08,059 ago this virus was unknown to us. We have already learned so much about it. We know 629 00:53:08,059 --> 00:53:14,920 it's DNA, we know it can be transmitted from one person to another, we know that those 630 00:53:14,920 --> 00:53:22,700 most at risk are older people and those with underlying health conditions. 631 00:53:22,840 --> 00:53:29,440 But there is still a lot we don't know. We don't know the source of the outbreak. We 632 00:53:29,440 --> 00:53:36,210 don't know what it's natural reservoir is, and we don't properly understand it's transmissibility, 633 00:53:36,210 --> 00:53:42,339 or severity. To defeat this outbreak, we need answers to 634 00:53:42,339 --> 00:53:49,130 all those questions. And there are tools we don't have. We have no vaccine to prevent 635 00:53:49,130 --> 00:53:59,120 infection, and no therapeutics to treat them. To put it bluntly, we're shadowboxing. 636 00:53:59,360 --> 00:54:06,180 We need to bring this virus out into the light, so we can attack it properly. 637 00:54:06,180 --> 00:54:13,780 That's why on February 11 and 12, we're convening a global research and innovation forum. 638 00:54:13,789 --> 00:54:19,210 The World Health Organization, they've just asked for 500 million dollars for research 639 00:54:19,210 --> 00:54:25,700 into the virus, which of course is wonderful, I hope they get it. We really need to get 640 00:54:25,700 --> 00:54:29,530 on top of viral infections. We're just in the Middle Ages in a lot of ways as far as 641 00:54:29,530 --> 00:54:32,740 viral infections is concerned. Anti-viral drugs are very limited. 642 00:54:32,740 --> 00:54:38,500 How can we possibly make things better, how can we improve the research speed and the 643 00:54:38,500 --> 00:54:41,799 accuracy of the results? The first issue that one needs to deal with 644 00:54:41,799 --> 00:54:46,950 is that of standards. If several thousand people generate data in different formats, 645 00:54:46,950 --> 00:54:50,680 nobody will be able to use that. So it is very important for the community to have some 646 00:54:50,680 --> 00:54:54,760 standards that specify how the data should look like. 647 00:54:54,760 --> 00:55:00,609 Another important step that community could take, is for major journals to require people 648 00:55:00,609 --> 00:55:04,950 to make their data available when a paper is submitted for publication. This is very 649 00:55:04,950 --> 00:55:09,369 important because, again this is human nature, the tendency of most PIs 650 00:55:09,369 --> 00:55:13,960 is to hang very close to their data. So it's is very important so results can be verified, 651 00:55:13,960 --> 00:55:19,609 and most importantly, could be re-analyzed, in conjunction with other data sets. 652 00:55:19,609 --> 00:55:25,230 I think researchers could gain a lot of insight if some of the clinical trial data 653 00:55:25,230 --> 00:55:30,809 were available for additional mining or interpretation and so forth. 654 00:55:30,809 --> 00:55:36,670 There is an ongoing debate about FDA approval process. On the one hand, you have to be very 655 00:55:36,670 --> 00:55:41,500 careful, you cannot, give a certain drug to millions of people potentially without testing 656 00:55:41,500 --> 00:55:47,910 it thoroughly. However, the current process can be extremely painful and extremely long, 657 00:55:47,910 --> 00:55:53,239 especially in those cases where there is compelling evidence that a particular drug candidate 658 00:55:53,239 --> 00:55:56,430 works. In some situations, the treatment can be found relatively rapidly, 659 00:55:56,430 --> 00:56:01,569 and it's a shame that treatment not to be made available to people who need it 660 00:56:01,569 --> 00:56:08,549 just because the drug has not been approved yet. Patients with terminal conditions, terminal 661 00:56:08,549 --> 00:56:17,660 illness, should access to experimental treatment immediately that could potentially save their lives. 662 00:56:17,660 --> 00:56:24,059 People who are terminally ill should not have to go from country to country to 663 00:56:24,059 --> 00:56:32,089 seek a cure. I want to give them a chance right here at home. It's time for Congress 664 00:56:32,089 --> 00:56:42,840 to give these wonderful, incredible Americans the Right To Try. 665 00:56:42,840 --> 00:56:47,739 Right To Try applies only to patients with terminal conditions. The FDA needs a modern 666 00:56:47,739 --> 00:56:53,420 approval process, that also lets anyone with a debilitating disease try experimental cures. 667 00:56:53,420 --> 00:56:59,360 We should reform patent laws, prohibiting them for math, software, or biological genes. 668 00:56:59,360 --> 00:57:04,780 Technology doesn't exist in a vacuum. Bad laws can stifle progress, and encourage monopolies 669 00:57:04,789 --> 00:57:10,859 and corruption. 670 00:57:10,859 --> 00:57:14,930 The power of community applies far beyond the Internet. Every year new movements are 671 00:57:14,930 --> 00:57:20,280 created that change the world. They can also help travel to the final frontier. Because 672 00:57:20,280 --> 00:57:25,060 of the high cost of rockets, the community surrounding space exploration is small today. 673 00:57:36,300 --> 00:57:40,320 All systems are go. We're about 90 seconds from the launch of Space Shuttle Discovery. 674 00:57:40,340 --> 00:57:53,360 Go to go for launch: Guidance, FIDO, Bravo, GNC, Effects, Eagle, ECON, Payloads. For the customer, 675 00:57:53,560 --> 00:58:02,600 BPS, Ginko, Booster, Surgeon, PDRS. Okay, lets see, we'll wait on GNC. 676 00:58:02,820 --> 00:58:07,660 Discovery's onboard computers have primary control of all the vehicle's critical functions. 677 00:58:07,660 --> 00:58:23,500 20 seconds. T-17 seconds and counting, 15, 12, 11, 10, 9, 8, 7, 6, main engine start, 678 00:58:24,120 --> 00:58:32,780 2, 1, booster ignition, and liftoff of the Space Shuttle Discovery returning to the Space Station, 679 00:58:32,780 --> 00:58:35,600 paving the way for future missions beyond. 680 00:58:44,360 --> 00:58:46,720 57% of rated performance reducing the stress 681 00:58:46,920 --> 00:58:49,040 on the Shuttle as it breaks through the sound barrier. 682 00:58:49,280 --> 00:58:53,420 Discovery is already 1.5 miles in altitude, 1.5 miles downrange. 683 00:59:10,560 --> 00:59:12,560 You want software wars? I can tell you about 684 00:59:12,580 --> 00:59:17,359 software wars. You think software is important to NASA? When I went in there, it was at the 685 00:59:17,359 --> 00:59:24,150 end of the Mercury program, which didn't do much, into the Gemini program which was, among 686 00:59:24,150 --> 00:59:29,779 other things its primary purpose was to develop rendezvous. 687 00:59:29,779 --> 00:59:33,460 So you can imagine you're trying to shoot something up and have it find something else 688 00:59:33,460 --> 00:59:39,710 in space, hook up with it, so people can get out, and go back and forth or connect, which 689 00:59:39,710 --> 00:59:44,380 was the heart of the Apollo design. Well that was all about software. 690 00:59:44,380 --> 00:59:49,980 With the Shuttle, it got worse. On Gemini it was a small little computer, it got quite 691 00:59:49,980 --> 00:59:54,039 a bit bigger on Apollo, but the computer was still nothing like you have in your pocket 692 00:59:54,039 --> 00:59:57,789 right now, which is kind of staggering when you think we went to the moon and back, and did all 693 00:59:57,789 --> 01:00:02,661 the navigation, all of the deliberations, control of the burns, was all done with a 694 01:00:02,661 --> 01:00:07,519 computer was no better than your iPhone. In fact, it was way inferior to your iPhone. 695 01:00:07,519 --> 01:00:11,249 You had to have boosters which you could light the fuse on, and they would take off, but 696 01:00:11,249 --> 01:00:16,170 software was where it was at. Some of the biggest knock-down, drag-out meetings 697 01:00:16,170 --> 01:00:23,150 that I was at, and contentious issues, involved software: how to configure systems, how to 698 01:00:23,150 --> 01:00:29,420 respond to emergencies, how to decide if there was an emergency, and it was paramount to 699 01:00:29,420 --> 01:00:34,369 what we did and what we accomplished. But then I got into simulation. "What's that 700 01:00:34,369 --> 01:00:39,539 about?" Computers, everything about it was computers. The hardware involved was almost 701 01:00:39,539 --> 01:00:43,690 incidental. It was just to take a cockpit, put it on some actuators so you could jiggle 702 01:00:43,690 --> 01:00:48,599 it around, pipe some sound in, and put some instruments in so that the guy thinks he in 703 01:00:48,599 --> 01:00:51,970 spacecraft, but outside of that is where it's all happening with software. 704 01:00:51,970 --> 01:00:56,479 It was pretty far-fetched to go to the moon, when you think about it. In my day, I experienced 705 01:00:56,479 --> 01:01:02,099 what I thought was almost a crazy mandate. We did it, we put a man on the moon, we got 706 01:01:02,099 --> 01:01:11,599 him there, got him back alive, and had some narrow misses, and we were proud of what we 707 01:01:11,599 --> 01:01:16,440 had done. When I was doing space missions, we were always 708 01:01:16,440 --> 01:01:22,369 limited by how much we can send up. You've got little tiny instruments, on a big satellite, 709 01:01:22,369 --> 01:01:27,829 that goes on a gigantic rocket, so it all comes down to what you can launch. We were 710 01:01:27,829 --> 01:01:31,769 always looking for more. Most of my interaction with NASA was on 711 01:01:31,769 --> 01:01:36,600 the space elevator, a concept that's been around since 1960. You have a string, a cable, with 712 01:01:36,600 --> 01:01:41,529 one end attached to the earth and the other end up in space. Because the earth is rotating, 713 01:01:41,529 --> 01:01:46,220 the upper end is basically thrown out away from the earth. If you spin a ball on a string 714 01:01:46,220 --> 01:01:50,809 around your head, it sticks out away, and then you can climb the rope to space. 715 01:01:50,809 --> 01:01:57,339 The modern version of that is, it's a cable, made of high-strength materials, and it's 716 01:01:57,339 --> 01:02:02,980 got cars or climbers that are attached to the cable that go up, and can go to orbit, 717 01:02:02,980 --> 01:02:07,259 and deploy satellites. This is very different from the rockets, the 718 01:02:07,259 --> 01:02:13,390 rockets carry all their fuel, it's very heavy, very expensive. To go from rockets to the 719 01:02:13,390 --> 01:02:19,119 space elevator, you'll reduce cost of getting into space by hundreds to thousands of times. 720 01:02:19,119 --> 01:02:23,099 So instead of ten thousand dollars a pound, it would cost you a hundred dollars a pound 721 01:02:23,099 --> 01:02:29,640 to get into space, and you can have massive quantities. It's like a large railway going 722 01:02:29,640 --> 01:02:34,290 from earth to space. It changes the entire aspect of space. Space 723 01:02:34,290 --> 01:02:39,171 is no longer this inaccessible location, you can go there, you can do what you want, you 724 01:02:39,180 --> 01:02:42,820 can make it into something that you want for the future. 725 01:02:45,100 --> 01:02:49,339 I wanna go into space, I want other people to be able to. We just can't do that on the rockets. 726 01:02:49,339 --> 01:02:54,920 I saw in Arthur Clarke's book about the Space Elevator, and then I heard comments like, 727 01:02:54,920 --> 01:02:59,029 "It simply can't be done." That's a statement, that gets some people interested in doing 728 01:02:59,029 --> 01:03:04,710 it, and that's where I started looking into the space elevator and found we could do it. 729 01:03:04,710 --> 01:03:09,719 When I was 5 years old I watched Neil Armstrong step onto the surface of the moon, I don't 730 01:03:09,719 --> 01:03:15,329 know, it just instilled a huge passion in me for engineering and science, and exploration, 731 01:03:15,329 --> 01:03:21,820 I've been kind of crazy about rockets since I was a little boy. 732 01:03:21,820 --> 01:03:27,579 Single entities have tried, and they haven't made it: NASA, Virgin Galactic, we had SpaceShip 733 01:03:27,579 --> 01:03:32,900 One, they were great, but they're also very limited. So you need to broaden it, you need 734 01:03:32,900 --> 01:03:36,349 to get lots of people, you need to get interest, there needs to be returns. 735 01:03:36,349 --> 01:03:43,099 It's sort of an open source for space, to get people in, doing the modeling, doing hardware 736 01:03:43,099 --> 01:03:47,299 development, doing the research, getting the funding. And it all needs to be in open source, 737 01:03:47,299 --> 01:03:50,280 it needs to come in from a lot of different places. 738 01:03:50,280 --> 01:03:55,450 From a technical standpoint you could easily build a space elevator in about ten years. 739 01:03:55,450 --> 01:03:58,729 You know, if you really get down and said: "We wanna build this thing," you did sort 740 01:03:58,729 --> 01:04:02,950 of a Manhattan Project kind of thing where everybody's focused on it, you bring in who 741 01:04:02,950 --> 01:04:07,390 you need to, and you just do it. You can probably have a space elevator up in about ten years, 742 01:04:07,390 --> 01:04:15,329 from there it would grow fairly dramatically. I think the space elevator can be accomplished. 743 01:04:15,329 --> 01:04:20,229 I think it will be momentous, I think it will be beyond anything we'd ever imagined that 744 01:04:20,229 --> 01:04:26,150 we would try to do. There's no show-stoppers per-se, it's just technology, taking it a 745 01:04:26,150 --> 01:04:30,700 step at a time, and figuring out what needs to be fixed, what needs to be designed. 746 01:04:30,700 --> 01:04:35,989 The Space Elevator is estimated to cost 10 billion dollars to create. NASA could take 747 01:04:35,989 --> 01:04:40,880 it on with their existing funding, if they cancelled projects made obsolete by the elevator. 748 01:04:40,880 --> 01:04:46,411 Many space elevator enthusiasts think strong nanotubes are impossible to manufacture, even 749 01:04:46,411 --> 01:04:50,839 though they carry around a supercomputer that runs on batteries in their pocket. Unfortunately, 750 01:04:50,840 --> 01:04:54,860 NASA hasn't asked commercial industry for help producing nanotube threads. 751 01:04:54,880 --> 01:04:58,900 For fastest results, we should create the elevator designs in public. Harnessing the 752 01:04:59,100 --> 01:05:05,280 intellectual power of universities can accelerate the process. Once the first elevator is completed, 753 01:05:05,500 --> 01:05:09,840 we could build a tremendous base on the moon within a year. You could arrange a trip from 754 01:05:09,840 --> 01:05:11,780 your phone. 755 01:05:16,840 --> 01:05:21,480 This open source model isn't necessarily making a better world, it's making a better product. 756 01:05:21,490 --> 01:05:25,769 This a way, a really good way, to create software. Now what you make that software do is something 757 01:05:25,769 --> 01:05:30,519 totally on up to you. You can make software do bad things, you can make software do good 758 01:05:30,520 --> 01:05:34,400 things, and we're putting it out there without any restrictions on its use. So hopefully 759 01:05:34,400 --> 01:05:37,900 Linux will create a better world, but again we are just creating software. 760 01:05:38,000 --> 01:05:43,759 I know from our sources deep inside those Silicon Valley institutions, they genuinely 761 01:05:43,759 --> 01:05:49,229 believe that they are going to produce artificial intelligences that are so powerful, relatively 762 01:05:49,229 --> 01:05:55,460 soon, that people will have their brains digitized, uploaded to these artificial intelligences, 763 01:05:55,460 --> 01:06:02,130 and live forever in a simulation, therefore will have eternal life. It's like a religion 764 01:06:02,130 --> 01:06:05,280 for atheists. I suck at that, honestly, because I feel like 765 01:06:05,280 --> 01:06:11,060 there's so much more I probably can do with my phone and computer like combined together, 766 01:06:11,100 --> 01:06:14,800 but I don't take the time out to learn those type of things. 767 01:06:14,800 --> 01:06:17,580 A lot of what's going to determine the "software 768 01:06:17,589 --> 01:06:23,569 wars" between open source, free software and proprietary, is basically going to be economics. 769 01:06:23,569 --> 01:06:27,950 It's been widely publicized that the Navy has been asked to cut 25% of IT spending. 770 01:06:27,950 --> 01:06:32,369 You are going to see there come a point in time where the balance is going to have to 771 01:06:32,369 --> 01:06:37,269 be much more on the free and open side, as supposed to what we have now which is all 772 01:06:37,269 --> 01:06:41,049 proprietary, where things are re-invented each time, and charged-for each time. 773 01:06:41,049 --> 01:06:44,690 As taxpayers and citizens, that's what we really should be pushing our government to 774 01:06:44,690 --> 01:06:50,229 think about is, what are the economic benefits of using, and not only using but contributing 775 01:06:50,229 --> 01:06:56,009 back to free and open source projects, so that they can become more self-sustaining 776 01:06:56,009 --> 01:06:59,719 even more than they are today. But I think in the open source or free software, 777 01:06:59,719 --> 01:07:06,430 it's changed because it's quality. I don't think people realize that companies and individuals 778 01:07:06,430 --> 01:07:11,579 are starting to work together in open source and it's becoming very reliable, very secure, 779 01:07:11,579 --> 01:07:18,579 and very capable set of functionality. Because it's free, it's really empowering. 780 01:07:18,579 --> 01:07:24,749 The ability to download some free software, open up the source code for it, be able to 781 01:07:24,749 --> 01:07:30,460 learn from it, and build upon it yourself with no restrictions. It just unlocks infinite 782 01:07:30,460 --> 01:07:34,000 creative potential. One of the problems that people face in this 783 01:07:34,000 --> 01:07:37,910 area is that they think they want to create something new. You know, "I want to become 784 01:07:37,910 --> 01:07:42,339 Linus Torvalds." My advice to you is to join other people, 785 01:07:42,339 --> 01:07:45,989 because the real goodness in free software comes not from doing things by yourself, but 786 01:07:45,989 --> 01:07:50,359 by doing things together. I don't think you can stop that. I think the 787 01:07:50,359 --> 01:07:54,739 Microsoft's out there, and the other proprietary companies out there would like to see it stopped, 788 01:07:54,739 --> 01:07:59,339 because it threatens their profit model, it threatens their control, not only of money, 789 01:07:59,339 --> 01:08:03,440 but of information. There's nothing more valuable than being able to control information, 790 01:08:03,440 --> 01:08:08,690 not just the flow of information, but information itself! 791 01:08:08,690 --> 01:08:14,460 I think the Wikipedia's out there, and the other projects out there, that are not just 792 01:08:14,460 --> 01:08:18,700 liberating the information, but gathering the information, and disseminating it, and 793 01:08:18,700 --> 01:08:23,880 keeping it free, are ultimately just going to eclipse the proprietary model. 794 01:08:24,020 --> 01:08:31,480 Source code is so complicated, any interesting project, even a small project, is so complicated 795 01:08:31,480 --> 01:08:40,400 that I claim that no developer can really understand the problem anymore. It really 796 01:08:40,400 --> 01:08:46,020 boils down to that. Yes, science spread through the world, because 797 01:08:46,020 --> 01:08:53,210 it works. You can't do complex things in a hidden environment where you try to control 798 01:08:53,210 --> 01:08:59,040 it because the control will basically limit your technology. 799 01:08:59,040 --> 01:09:04,710 But whoever says that Linux is a cancer, or Linux is Communist, is just uneducated and 800 01:09:04,710 --> 01:09:12,810 stupid. It's that simple. It's like saying science is Communist, or science is a cancer. 801 01:09:12,810 --> 01:09:20,690 That's completely idiotic, it's beyond stupid. If you don't spread your ideas in the open, 802 01:09:20,690 --> 01:09:26,251 if you don't allow other people to look at how your ideas work, and verify that they 803 01:09:26,251 --> 01:09:32,040 work, you're not doing science. One of my favorite quotes from history is 804 01:09:32,040 --> 01:09:36,660 Isaac Newton, who was apparently not a very nice person but he was certainly smart, 805 01:09:36,660 --> 01:09:42,390 he said: "If I had seen further, it has been by standing on the shoulders of the giants", 806 01:09:42,390 --> 01:09:49,609 where the whole point, is in order to reach greatness, you have to be able to build on 807 01:09:49,609 --> 01:09:56,740 what has gone before you. And so we will decide who wins based on what 808 01:09:56,740 --> 01:10:01,190 people use, and what the general population of users use. 809 01:10:01,190 --> 01:10:05,990 It's about people having the ability to control their own destiny when it comes to technology, 810 01:10:05,990 --> 01:10:09,790 and that's the most important thing. And that's what open source has given us, and that's 811 01:10:09,790 --> 01:10:16,670 what's challenged the industry as well. I mean, Microsoft are a different company because 812 01:10:16,670 --> 01:10:20,950 of open source. 100 years after the transistor was invented, 813 01:10:20,950 --> 01:10:26,620 proprietary technology continues to dominate many aspects of society. Billions of dollars 814 01:10:26,620 --> 01:10:32,000 are wasted, while freedom and cooperation are hindered. The more we share scientific 815 01:10:32,000 --> 01:10:35,760 information, the faster we can solve the challenges of the future.