How code has changed the world - with Torie Bosch

foreign [Music] thank you so much for being here tonight it is a true honor I need to start this talk with a little bit of a confession which is that I cannot code and that might not be what you were expecting to hear tonight given the title of my book and the topic that we're discussing but it's true if you show me a screenshot of code unless it's something extremely basic like hello world or a birthday calculator the things you learn on the first day of a programming class I'm not going to understand it in fact I would wager a whole lot of you here tonight no code far better than I do nevertheless I've spent a pretty good portion of the last few years of my life devoted to the ways that code intersects with our lives and more importantly how humans produce code how they write it how it goes wrong how it goes right and how it affects all of our lives programming and the decisions around it undergird our cars our banking systems our financial systems which are our banking systems our schools criminal justice employment decisions public policy and social safety nets I think that we like to think that code is something almost clinical that it's pre-ordained it works that way because that's how it was meant to be but in code there is no meant to be it's a series of decisions whether you're a coding Wizard or like me you don't know your JavaScript from your python I hope that you'll leave here tonight with a better understanding of how each piece of software consequential or dot is made up of that series of decisions decisions that reflect humans biased messy short-sighted yet often brilliant thinking I think it's important for even the non-technical among us to understand that code is infused with human values and it's really critical for technologists to sometimes take a step back to think about how their work affects people and even listen to people who don't know how to code just once in a while as Lisa said you are not expected to understand this describes 26 lines of code both figurative and literal that have changed the world or reflect something about the process of creating code it covers the first pop-up ad a code error message that nearly aborted the first mission to the Moon a t-shirt with a snippet of code that made it technically violate U.S government export controls even Internet relay chat for those of you old enough to remember IRC like me it's written by technologists historians journalists human rights Advocates and others who work closely with technology and care about not just producing a product but how that product affects people of course the book has a rather provocative title one that I'll explain a bit in a bit and in a way I think it belies my personal belief which is that everyone should be expected to understand the basic way the code shapes our lives and how it's built and on the flip side that the tech industry has a lot to gain from adapting some Humanity's thinking I'd like to back up a bit now and tell you how I came to edit this book I said I don't know how to code and that's very true but it's not for lack of trying I first learned about the power of software as a teenager when I was about 14. from two things pop music and a near apocalypse we're going to start with the Backstreet Boys which says I'm sure what you were all expecting tonight yes the Backstreet Boys and you can't shame me more than I have shamed myself over this over the years when I was about 14 in 1998 I became a huge fan no one else in my life would admit to liking them I still think they were just hiding it so to find other people who would let me talk to them about the Backstreet Boys I went to this new place called the World Wide Web and there I found lots of people who would admit to liking the Backstreet Boys At least from the safety of their computer the coolest among them had websites websites hosted on sites like geocities Angel Fire and tripod and I decided I had to be one of the cool people I wanted my own Backstreet Boys site a Backstreet Boys fan fiction site no less now today if you decide to build a website fan fiction or otherwise God bless you if you do you get to use an editor that's what we call wysiwyg what you see is what you get and it works a lot like a word processing program if you want to build something you just hit that bold button but back in the 90s you had to do it yourself so I had to learn hypertext markup language or HTML and that's what's telling your browser what to do the wysiwyg editors do that for you so I bought books I scoured the internet and I learned that meant that tweet in teenage girls from around the world could come to my website where they could Feast their eyes on this sort of seafoam green marbled background with black text that broke all the accessibility rules I'm amazed that anyone bothered to read it and there I posted my terrible stories about Backstreet Boys falling in love with me and my friends and then I started posting other people's terrible stories about The Backstreet Boys falling in love with them and their friends and in time and I don't know why I insist on telling people this I created one of the biggest Backstreet Boys fan fiction sites online at the time but I really loved when I stepped back and thought about it how HTML gave me the power to give instructions not just to one computer but to every computer that navigated to my site if I wrote that I in little carrots the text would become italicized if I forgot to close that eye then all of the text that followed would be italicized if I forgot the r in the ahref part of a link that link wouldn't work and normally I would be linking to another person's Backstreet Boys fanfiction website per an agreement if I linked improperly that person would be angry because we would have agreed to link to each other as a side note I'm not the only person who learns some tech and online organizing skills from fandom I highly recommend Caitlin Tiffany's recent book everything I need I get from you how fangirls created the modern internet as we know it which is a little bit more modern that's about One Direction as I was teaching myself HTML something else was going on that was slightly more significant why 2K for those of you who might be too young to remember basically in the 20th century computer memory was very limited in fact your iPhone has more than one million times more memory in Ram than the computer that took the first mission to the Moon so an easy way for programmers to save space was to code years with two digits instead of four that meant that computers assumed every year started with 19. and that worked great in the 50s 60s 70s and 80s and then we got to the early 1990s and people started to realize that that could be a problem as we approached the year 2000. the big fear was that we would hit January 1st 2000 and computers would think we had gone back in time to January 1st 1900. and that would cause a whole lot of problems I think sometimes especially if you're a teenager it feels like the year 2000 was a technological Dark Age but even 23 years ago computers were really Central to our lives as fear about why 2K began to spread people were worried that electricity might stop working planes might stop Flying government payments wouldn't go out to people who needed them the banking system might collapse Doomsday Preppers started to stock up on food and supplies for the virtual end of the world or the end of the virtual world you might say now most people realize that the world wasn't really going to end but that didn't mean everything was going to be just fine so governments companies and individual programmers raced to update computers before we hit New Year's Day and in the end it mostly worked there are some problems when the calendar turned to the year 2000. most of them were fairly small scale so for instance some ATMs malfunctioned one of my favorites examples slot machines at a race track in Delaware in the United States stopped working well but there were some more serious problems as well here in the UK within the National Health Service people in a very small geographic area pregnant women who are being screened to see if their fetuses were at risk of down syndrome the screening program worked incorrectly it calculated their ages wrong and that meant that several women ended up finding out very late that their babies did have Down syndrome for most of us though things went along pretty much as usual and a lot of people thought or still even think that that meant that Y2K was over hyped that it was never really going to be that big of a problem but that wasn't the case Y2K didn't become a huge problem because programmers saw the threat was coming and they fixed it and in my mind that made programmers a hero between the Y2K heroics and my self-perceived Wizardry with HTML I thought computer science was the future and maybe part of my future too so when I was in 10th Grade what you would call year 10 I decided to take a c plus class the course catalog said that there were no prerequisites you didn't have to know a bit about coding to take it I thought I would Ace it and then I had the first day of class I was the only girl in a class of about 20 boys and all of them it seemed knew exactly what they were doing most of them had learned programming from their parents or from their dads and were just trying to add a new programming language to their repertoire my mother had actually coded a little bit in the 70s when she was a Management Consultant but she hated it and later I asked her how could you let me do this and she said I thought you would do better than I would and I showed her so I flailed I thought my HTML was background would help me and I couldn't understand what was going on now I understand the problem HTML under one definition is code it gives instructions to a machine but a better definition of code or a more modern one is that code has to compute something so HTML wasn't really code it was a really really basic idea that I didn't know I didn't know my class though was filled with people who did understand that so we didn't bother with that sort of introduction what are we doing here my teacher was a woman who seemed to think that I was single-handedly going to hold back all girls and women in computer science everywhere so she just sort of wrote me off pretty quickly sometimes when I was really stuck the boys would just fix my code for me very kindly they were all very sweet but it was clear to everyone in that room that I had no future in computer science so when I went to University I decided to study English which is the final Refuge of the non-technical eventually I became a journalist covering all sorts of topics and then eventually that led me back to Tech in 2011 I began working as the editor of future tense a partnership of Arizona State University slate magazine and new America which is a think tank in Washington DC the goal of future tense is to cover the intersection of Technology policy and Society and when the good Folks at slate ASU and new America told me about the job opening and asked if I was interested I said absolutely not I remembered that coding class how exactly was I supposed to help cover artificial intelligence and cyber security if I couldn't even limp through a c plus class But ultimately they told me and I think this was right that it was actually a benefit for me not to have a strong technical background they pointed out that the readership we wanted to reach was people who weren't necessarily experts the idea was to help citizens learn about the future to help them take part in the discussions about the technologies that were going to be so important to them going forward my non-technical background let me be an advocate for them if I didn't know what a computer scientist was writing then our reader wouldn't either and so I could ask that question for them so I began editing opinion and Analysis columns by academics and other experts who were interested in educating the public I also started to meet technologists who desperately wanted to work with somebody like me they wanted they told me to try to break out of their bubbles this is around 2011 2012 when as you may remember the hype about technology was huge Tech was going to fix everything and I was meeting people who worked in Silicon Valley and told me that they were starting to really get worried that computer scientists were only talking to each other they weren't actually listening to what the rest of us might want from Tech they just wanted to solve problems without thinking about the human in 2019 a colleague of mine at slate began reading the coverage of the coding error that contributed to the crashes of Boeing 737 Maxes in Indonesia and Ethiopia it was a terrible tragedy more than 300 people died my colleague thought that it might be Illuminating to collect other examples of places in which a tiny coding error had huge consequences huge negative consequences I thought it sounded great too so I volunteered to lead the project which would end up becoming both a series on slate and then this book so to get started I sent a lot of emails I surveyed almost 200 technologists historians journalists human rights Advocates and others to find out what error in code did they think it had the biggest consequences we also asked them to provide a little snippet of the code if they could and well technologists can be pretty blunt and they quickly told me that I was so misguided on this they told me that yes if we looked for small errors in code with big consequences we'd find other examples some of the other famous examples are that in the early 1980s the Vancouver Stock Exchange software had a tiny rounding error that ended up making the exchange lose a pretty significant amount of its value until someone realized what was going on then the exchange had to shut down for a little while to fix it a couple years later in a more tragic example there was a radiation machine used to treat cancer patients a coding error meant that it gave patients a far too large dose of radiation and at least three people died as a result but in a grim way these errors the Boeing 737 Max the rounding error in the Vancouver Stock Exchange the Iraq the uh the radiation machine they were all sort of the same small error big problem and the message is just be careful in the end that doesn't really tell us that much new what would be far more valuable I realized was to mix those sorts of errors with more figurative lines of code and when I expanded the scope of what we were looking for the experts had a lot more ideas including internet lingo famous comments and code the birth of a coding language the structure behind Wikipedia and all sorts of other things digging through those suggestions took months really thought-provoking months and the end result was a big package that we've covered published on slate in October 2019 called the lines of code that changed everything and the spine of this package was a list of 36 blurbs written by both experts and journalists at slate about all sorts of significant coding moments beginning in 1725 with the birth of the punch card or binary code essentially we also published a handful of essays about topics that were a little bit hard to boil down into 200 words so there was an example about the binary about binary code the gender binary and the huge effort that it's taken to update large institutional databases many of which were coded decades ago to reflect more than just male and female another looked at hello world and the spark of joy that comes when you code your first successful program a spark that I don't identify with but I really wish I did it sounds fantastic we also had one from Slate's excellent Vice Tech vice president of Technology who wrote about how it's so difficult for programmers to code not just for next week but for years into the future an idea that I'll be coming back to a lot tonight now if you've ever found yourself a little riled up by a list of the best 50 songs of the past 50 years you know that the purpose of an online list is not really to be definitive it's to get you arguing and this list did that so we had lots of technologists some of whom whom did not respond to my initial email saying why didn't you include this or I disagree about that and the online conversation kind of helped us decide that we could make a book out of this the good Folks at Princeton University press agreed and nearly four years later here it is thanks in no small part to Hallie Stebbins the Fantastic editor that I worked with at Princeton University press and it features beautiful illustrations by Kelly cheddler which you'll see some more of tonight the book has a whole lot in common with the original package but also several new ideas and new writers and of course it's expanded and now I think it's 26 essays that can offer something to both non-experts and experts alike now I'm going to go through some of the examples from the book and the way they can help all of us understand a little bit more about how good intentions in programming can have bad consequences and why it's so important to bring Humanity's thinking even from failed computer scientists like me to software engineering our first story comes from Ethan Zuckerman back in 1997 a couple of years before I would become a Pioneer in the highly competitive world of Backstreet Boys fan fiction Ethan was a young programmer at tripod tripod was one of a handful of services that would allow users to create their own personal web page which was a pretty new idea the services were free that we're free because as now the services could add ads to your website now the online advertising ecosystem was very different then today if you look at a pair of shoes online they will follow you around for weeks back then you couldn't even really buy shoes online and if you could they certainly would not be able to Target you or follow you that way much to advertisers dismay they had no idea where their ads were going to appear and in fact even now that's a difficult problem every now and then you'll hear about say a controversial YouTube influencer who says something that creates a firestorm and then advertisers get very angry when they realize that their brands are appearing as pre-roll on that person's YouTube page so 1997. unfortunately if you give people the opportunity to create whatever they want online some people aren't going to be terribly wholesome in fact even Backstreet Boys fan fiction got pretty racy at times which was horrifying to me at the time so 1997 Ethan zuckerman's boss at tripod tells them they have a big problem an auto company has discovered that their ads are appearing on a website devoted to a rather adult recreational activity Ethan's boss asked him to find ways to sell ads while ensuring that brand managers wouldn't send him outrage screenshots of their precious Brands appearing alongside offensive content so Ethan had an idea he wrote the line of code that you see here this code was inserted into the top of every home page hosted by tripod and when it loaded in your web browser it would launch the page you wanted to go to but also a second window and that second window included some tripod branding some tools to navigate between user generated pages and critically a 200 by 200 pixel ad this was probably the world's first pop-up ad and as Ethan puts it and you are not expected to understand this the result was Presto plausible deniability the pop-up was his solution to the thorny problem that the internet still faces how can advertising support user-generated content just a couple of weeks later tripod's leading competitor geocities copied Ethan's code and used it to launch a similar website as Ethan points out this didn't exactly require sophisticated corporate Espionage you could just hit view Source in your browser and see it that's how lots of JavaScript programmers learned their trade that's how I learned a lot of HTML not long after that Ethan writes he saw a pop-up ad with only an ad in it no separate branding or navigation tools and then came pop-up ads that would move around the screen as you tried to chase as you tried to close them which was fun for like three seconds then there were pop-up pop unders that would appear underneath your web page so you'd think you had made it through on scathe until you closed the window and found the papa unders and then there were Cascades of pop-ups just one after another until as Ethan puts it you would wish you had never heard of the internet in the first place the worst was when you encountered basically the opposite of the problem Ethan was trying to solve you went to a perfectly wholesome looking page only to have a rather explicit pop-up ad attack you now some of you may have lived in an internet pretty free of pop-up ads because browsers are very good now at um at keeping them from well popping but for those of us old enough to remember the pop-up ad was like a little bit traumatizing I think we're all a little bit scarred by them Ethan learned about that Collective trauma in 2014. that year he wrote an article for the American Magazine the Atlantic about what he called the original sin of the internet offering users free content in exchange for monitoring our attention and surveilling us it was a very long essay that included that little shortest side you see there I wrote the code to launch the window and run an ad in it I'm sorry our intentions were good a very Savvy editor at the Atlantic notice that little nugget and realized that that was something people would really want to read about so in addition to the long article she published a very short item on the website with just a little that part about the ad and an extra quote or two from Ethan so now there are two articles running on the atlantix website the first is a 4500 word thoughtful exploration of the uh the Internet Ad ecosystem and the second was about 260 words made up of two or three quotes which one ended up getting a whole lot more readers the apology ricocheted across the internet it even appeared on late night talk shows mostly it generated Rye laughs most of us have recovered from this drama but some people were genuinely Furious Ethan has experienced actual death threats and swatting as a result of this in case you don't know swatting is when someone will falsely call police and say there's an active hostage situation going on so that the police will go and try to stop it it's incredibly dangerous Ethan understandably is pretty sick of talking about the pop-up ad but he agreed to revisit it for this book for one reason since he apologized many technologists and Business Leaders have told him that he's taking on far too much blame here as they put it his code wasn't exactly groundbreaking it was clever but if he hadn't done it someone else probably would have or at least something quite similar so does that very likely scenario Ethan wants to know mean that no one was actually responsible for the pop-up ad that it was just capitalism not the action of a single person Ethan could take that out but he's someone who feels a great deal of responsibility he says that the problem with refusing to accept personal responsibility for making the internet a worse place is it means that no one is responsible for making the internet a better place the internet is built on the system of surveillance capitalism monitoring us all turning us into data and then both selling that data and using it to advertise to us if the emergence of surveillance capitalism is diffuse then it's a force of nature it's Unstoppable and incomprehensible but if it's actually the sum of small decisions made by self-interested actors then it's understandable and maybe it's reversible today in a sort of penance Ethan's working at the University of Massachusetts and is trying to create a version of social media that doesn't watch you predict your behavior and Target adds it to you but I wonder even if the pop-up ad wasn't inevitable was it unethical of Ethan to code this it would have been really difficult for him to imagine how a single pop-up would spread through the internet causing endless amounts of frustration and even spreading malware it's a demonstration of how decisions that end up having big ramifications don't seem all that important in the moment encoding and elsewhere Ethan wasn't really trying to transform the way Internet ads were displayed he just wanted to keep from losing that auto company as a as a sponsor the difficulty of long-term thinking and code brings us to our next example of a programming decision that have consequences for decades to come but this one is far more serious this example demonstrates how computer systems far from being somehow objective reflect human values biases and blind spots and you are not expected Charlton McElwain who is a historian of technology and administrator at New York University writes that in the early 1960s as the black civil rights movement raged across the streets there was a quest to build a more racially just and Equitable Society and that happened right alongside the Computing Revolution soon those two would fuse in a software system to help police departments collect crime data and then use it to determine where to focus crime fighting efforts it's a decision that would end up deeply affecting our society even today in the early 1960s IBM tops the list of the world's leading Computing companies it not only innovated new computer hardware and systems but new ways of thinking about the computer's role and utility in everyday Society in 1965 in IBM's 1965 annual report the company president defined the computer as essentially a problem-solving tool the company's marketing Representatives didn't peddle pre-packaged products instead they asked leaders in every leading industry what problems they faced and how IBM could solve them and while IBM labored to Market New computational solutions to social problems uprisings were happening across the United States one of the most famous of these took place in 1965 in the Watts neighborhood of Los Angeles after one too many incidents of police brutality at the hands of the Los Angeles Police Department there were six days of unrest hundreds of police officers flooded the streets 14 000 National Guard troops were deployed 34 black residents were killed and thousands more were injured and more than 40 million dollars in property damage occurred just two and a half weeks before the Watts Uprising president Lyndon Johnson formed the president's Commission on law enforcement and the administration of justice the commission's goal was to study the causes of and find solutions to America's crime problem in the spring of 1966 following the Watts uprisings Johnson added the Science Tech and Technology task force to the commission to introduce new computational solutions to crime the task force was led by Saul I gas a mathematician and operations research Pioneer who had been managing all of IBM's Federal research projects and the central focus of the task force became the police beat algorithm a snippet of which you can see here it was designed to help police departments divide up Municipal areas and determine how to deploy resources in those areas doing so relied on two primary components the availability of crime data and two sorry one primary component the availability of crime data and two key computational techniques norming and waiting norming refers to analyzing the data to determine the normal and aberrant levels of criminal activity both across Geographic areas and across specific populations of criminal suspects such as white people versus black people waiting in this instance was a means to rank the severity of different crime so for example homicide rape burglary larceny and auto theft were all weighted with the same score of four signifying the most severe forms of crime but these weights were all arbitrary or even biased as Charlton writes the geographical areas were weighted by the preponderance of crime committed within its boundaries the crime data the statistical Norms weights and geographical configurations All Figured into the police beat algorithm so in one respect the police beat algorithm was developed to address a problem that framed black people primarily those who were poor and lived in urban areas as predominantly responsible for crime and as a result the problem that needed to be solved the police beat algorithm was therefore geographically determined to locate isolate and Target black and brown communities for police profiling and surveillance and if they were more policed more crime would be caught and that meant that they were just taking the history and projecting it on to the Future creating a reinforcing cycle in which they just kept being policed more and more essentially this was a proof of concept but it was actually implemented in 1968 in the Kansas City Missouri Police Department and it's through that system that we first saw the disparate impact that the police beat algorithm had the algorithm LED police to make the Tactical decision to concentrate on East Kansas City which housed most of the city's black citizens and ultimately the police beat algorithm became thousands of similar systems that now include facial recognition mobile surveillance and other tools and the same Logics and assumptions that permeated the police beat algorithm now show up everywhere from local law enforcement to International Security today when we talk about law enforcement algorithms people really like to invoke Minority Report the 2002 Steven Spielberg movie in which somehow A system can determine to the second when a crime will occur and who will commit it in fact when I was editing technology coverage I made a rule that you were not allowed to invoke Minority Report because I just got really tired of it I also got really tired of Terminator references and artificial intelligence we can talk about that later if you'd like but the problem not just me being getting bored with it is that I think that the comparison confuses people in the movie AI Works basically as though it's magic we really don't know exactly how it happens that the system made up of these three people with these things on their head know exactly when crime is going to happen but in reality we know exactly where that ability comes from as Charlton describes it comes from that mapping the past and projecting it onto the future except more so but when we talk about AI as though it's magic it makes it much harder to actually regulate it it's also worth knowing that one of the few good things about the police beat algorithm is that we do have the source code we know how it works we knew that homicide and auto theft were weighted in the exact same way today most of these systems used by law enforcement are proprietary and they don't release that source code so we have no way of knowing exactly how these systems whether in fact they're in criminal justice or other way or other systems that affect our lives actually work and that's I think a real problem so now we have one more example of code going wildly wrong in the late 1980s a handful of computer scientists began voicing concern about what could happen if a rogue program were released on the network but no one was prepared when it actually happened as journalist Katie Hafner writes this is Robert Tappan Morris and at 8 30 in the evening of November 2nd 1988 Robert was a 23 year old computer Cornell University Computer Science PhD student and he accidentally Unleashed just such a program it was really intended as a harmless hack just to prove that it could be done his program was supposed to copy itself from computer to computer and then hide in the background in as many machines as possible so that it couldn't be detected Morris used several bugs he found in the Unix source code and designed the program to lead the list of users on a Target computer and then it would run through their names permutations of their names and common passwords incidentally in 2022 the maker of a virtual private Network released a study that found that the most common password in the world was still password the number two most common password was one two three four five six and the number three for people who wanted just a little bit more security one two three four five six seven eight nine if you were using any of those passwords no shame I think that shame is one thing that holds us back in cyber security but please change it use a password manager they're beautiful so back to Robert Morris when his program guessed the password his it would then sign on to the computer and masquerade as a legitimate user and copy itself onto the remote machine not only did he have no intention of crippling the internet but he had actually tried to devise a way to keep it from propagating too swiftly once the program entered a computer it would signal its arrival to the machine and ask whether that machine had already been invaded if the answer was yes the two copies would talk to each other flip an electronic coin and decide which one which should stop running but and this is where Robert got into trouble what if someone discovered the intrusion and tried to trick the incoming program into believing that it was already on the machined approach which would access essentially vaccinate the computer against the intrusion thinking like a chess player Robert Morris decided there had to be a countermeasure against potential vaccination his solution was randomization that is that one in N times the program would enter a computer and command itself to run on the target machine regardless of whether a copy was already there and then he went to dinner unfortunately the number that Morris chose for n was seven and that was a fatal miscalculation the number should have been Higher by a factor of at least a thousand because the number was so low the result was that sometimes dozens or hundreds of copies of the program would appear in the same machine by the time he got back from dinner to check on the program's progress Morris realized that he was in deep trouble it had spread wildly out of control he even panicked and called a friend at Harvard and asked him to post this message on the network the friend posted the I'm sorry and the way to address it his friend noted that he hoped it was a hoax but it most certainly was not a hoax computer managers stayed up all through the night trying to fend off the virus which was later reclassified as a worm for the patents in the room it hopped back and forth around the internet people were also frantically trying to make their compute their Network safer because they were worried that this was just the first in a new Salvo of attacks at one of the Computing facilities at UC Berkeley someone taped a sign that read Center for Disease Control on the door within hours the program had crashed thousands of computers across the United States at universities research institutions and unfortunately for Robert U.S government Morris ended up being charged with a single crime under the 1986 Computer Fraud and Abuse Act a law that was itself interestingly inspired partially by the 1983 Matthew Broderick movie War Games in that movie Matthew Broderick breaks into a federal computer Morris became the first person charged under the provision of the law that made it illegal to gain unauthorized entry into a federal computer he was convicted but the judge decided to be lenient because it came became pretty clear that Morris really hadn't intended to cause this much damage so he was sentenced to community service he ended up going back to school not to Cornell this time he went to Harvard got his PhD in computer science he made a fortune in the Computing industry and he became a tenured professor at MIT and has never spoken publicly about what happened now I don't want you to get the idea that my book is entirely about things going wrong there is a really fun chapter on the Roomba and how it learned how to navigate your living room I can't have a Roomba because my cats lose their minds but anyway in another delightful chapter David Castle um the chapter in fact that gave the book its title David Castle talks about code comments Now comments are basically little lines in the code that are not directed to a computer they're directed to people to programmers they're succinct explanations for why the code is doing something in this spot sometimes that might even be sort of code to yourself to remind yourself of why you did something this way sort of like trying to help yourself read your handwriting many years in the future now normally comments are pretty prosaic you know they sometimes they're usually just something very boring but sometimes they can offer something a little bit more like a clear joke or they can unintentionally offer something more and that brings us to 1975 and a routine chunk of code for the sixth edition of the Unix operating system a Unix operating system programmer Bell Labs had written you are not expected to understand this and 50 years later those seven words live on but instead of warning off readers the enigmatic comment intrigued them and it became famous in its own right it became showing up began showing up on t-shirts and sweatshirts and even baby clothes as an expression of pride of technical Mastery that we understand something that the rest of you do not that I do not except as Dennis Ritchie one of the original programmers who wrote the Unix code later wrote that was not what was intended at all The 1975 team simply felt that the snippet of code in question was addressing a particularly obscure problem in 2004 he wrote you are not expected to understand this was intended as a remark in the spirit of this won't be on the exam rather than an impudent challenge but maybe the comment has remained so popular despite that misunderstanding because it speaks to a general fondness and warm appreciation for those glimpses of Personality that sometimes sneak into computer code in 2014 silicon Valley's Computer History Museum announced that they'd received permission from Microsoft to publish the historic source code and comments for the original word for Windows released in 1990. one Microsoft programmer begins a comment by first dubbing the section a gruesome hack after a particularly technical flourish another programmer added the comment coded in line because we're God there is an awful lot of profanity still is but there are other comments too that become sort of cherished reminders of a momentary glow of humanity in this world of unforgiving logic programmers may offer up a humble moment of self-deprecation or warnings about strength too far into exotic and Arcane Solutions and sometimes they do both at the same time when eunuch's Pioneer Roger Faulkner died in 2016 a fellow programmer remembered one of Faulkner's famous comments this is a horrible cluge or cludge depending on how you want to pronounce it a cludge is sort of like a a sloppy fix like MacGyver using duct tape it is vile it is swill if your code has to call this function your code is the same another chapter of the book looks at the Apollo guidance computer used during the first lunar Landing and as authors Ellen stofan and Nick Partridge Wright the programmers notes to one another in the code are tremendous there's an instruction to reposition an antenna that's labeled crank the silly thing around in order to safely continue off to see the wizard there's another sequence that ends with come again soon and a block to control instrumentation displays was filed under pinball game buttons and lights another section that made it into the final version was marked temporary I hope hope but the code also reflects the state of the nation in the world at the time there's an ignition sequence that's called burn baby burn which one young programmer later said was an allusion to the Black Power movement so even as they were preparing to go to another world the fight for justice was at the Forefront of their minds there's a really important meta message hidden in these silent libraries of intellectual Endeavors which is that it's people who write these programs and more importantly that it's communities of people who write them and preserve them there's a stereotype of a lone genius kind of hacking away but it's an outdated pop cultural idea codes written collectively though we very rarely see it that way at least those of us who don't work on it on a day-to-day basis every comment is an implicit acknowledgment of that code and all the careful caretakers who may someday have to revisit it and they may revisit your code Far later than you expect on Saturday I talked about this book at the Oxford literary festival and I made a really cheap joke about Y2K I said that programmers had fixed Y2K but they also caused it and one reason I felt that way is that as I mentioned my mother spent a brief unhappy period of her life working in programming and as she was learning she noticed the two-year four-year problem two-digit four-digit year problem and she asked an instructor hey is that going to cause problems in the year 2000 and his response apparently was that's going to be someone else's problem she told that anecdote especially in like late 1999 you know as a way to demonstrate programmers passing the book and I took it in that spirit and that's why I made the joke on Saturday but as the lovely gentleman who came up to me after the uh the address told me he had been a programmer in the 70s himself and it had never occurred to him that 30 years later anyone would still be using the software that they wrote and that's a great point because the truth is we don't just use 30 year old code we use 60 year old code when the pandemic hit and unemployment benefits were expanded in the U.S people flooded State unemployment websites those websites were built on databases that were written in Kobo which is a language that as is described and you're not expected to understand this was born in 1959 so in 2020 we were using a programming language from 1959 a language that nobody learns today in fact the situation became so dire that the state of New Jersey released this plea begging for retired cobal programmers to come back to work and help them fix the system similarly as Meredith Broussard writes in the book when same-sex marriage was legalized in the U.S it required changing thousands of databases this is a process that was informally and delightfully called why too gay these were databases that had been set up to only allow marriage between a man and a woman and so changing the law meant that the databases had to be changed as well Meredith phrases this beautifully computer systems are not just mathematical they're socio-technical and they need to be extensively and regularly updated just like humans you are not expected to understand this has a lot of big ideas but the one I really want to want to leave you with is that you can't look at code in isolation which is a lesson that I think many in big Tech still have not quite learned as this book was nearing publication in the U.S Elon Musk officially took over Twitter at the time I was still working at slate on technical topics which meant that musk also took over my life as he would regularly tweet things at 8 pm on a Friday and then I would have to try to cover it one of those times was when he said that he only wanted people who were hardcore to stick around at his new Twitter many of musk's initiatives were and probably are I don't pay quite as much attention now focused on code on this prioritizing code above all as this pure distillation of technical expertise at Twitter 2.0 he said there would be far less focus on humans and far more emphasis on code journalist Casey Newton reported that the day after musk took over Twitter Engineers were told to print out 50 pages of code they had written in the previous 30 to 60 days soon after that though a slack message went out presumably after Executives realized that it's probably not a very secure thing to have everyone print out 50 pages of code and they were all told to actually shred them and instead bring over their laptops to show off their code and a few weeks later another request went out must ask people who actually write software to email him a bullet point summary of what their code commits had achieved in the past six months along with screenshots of the most Salient lines of code and when I saw that request I asked my colleague Greg lavali who was Slate's VP of Technology what he thought and he laughed at the idea then I made him write about it for slate he told me that there are huge problems with the idea of looking at a screenshot of a piece of code and then using it to judge an engineer's ability which in a lot of ways goes back to the first criticisms of the project that led to this book looking at code and isolation is not really going to tell you very much if you just asked for the code you're not getting the context for why it was written at the time and who was writing it but perhaps more importantly as Greg wrote in slate code is written by teams and musk was asking for presentations from individuals musk has also fired many of the people from the softer side of Twitter those who work on government policy on Democracy on human rights and on fighting misinformation so at a time when the world needs more humanist thinking and Technology Twitter is now going in the other direction one of the risks of focusing on code more than humans as Ethan Zuckerman pointed out when we did an event in Los Angeles recently is that the very process of programming may be changing very soon you've probably heard about chat GPT the generative artificial intelligence system that you can ask to write a short story about clowns in the manner of George Orwell or ask for romantic advice which seems like a terrible idea and yet people are actually doing it but you can also ask it to write a small simple program and usually it'll work now prediction in technology is a really dangerous game but this suggests that it's possible that in a decade or so programming jobs could look very different than they do now but even if an artificial intelligence is writing code I would argue it's still going to be human in many ways remember how the police beat algorithm and its predictions were reflections of human biases boiled down into code and then projected forward again generative AI like chat GPT is built on humans specifically what we've written online we don't know exactly what trained chat jpt but I would bet it was an awful lot of Reddit for instance that means that even when written by an AI code is a reflection of humanity of our values of our assumptions it represents our priorities and these biases and it governs our lives in ways large and small so I think it behooves all of us to keep in mind whether we write code or not that code is human and humans need to understand code now that doesn't mean that you have to understand what any given line of code means but I think that to be an active participant in 21st century life you are expected to understand the power the code wields and how it's come to be thank you [Applause]

Channel: The Royal Institution

Views: 16,552

Keywords: Ri, Royal Institution, royal institute, history of coding

Id: hg-puN5GEFs

Length: 55min 28sec (3328 seconds)

Published: Thu Jun 01 2023