Equity Via Inclusion

We’ve been trying a big ambitious experiment in NYC over the last five years. We are training over 5,000 teachers to teach computer science classes in elementary school, middle school, and high school. We call it CS4All.

It sounds simple, but it is anything but simple. And it is expensive. We are funding it via public private partnership. I am leading the private part via a $40mm capital campaign. Talk to me about that if you are interested.

But here’s the thing. If you want to see engineering teams that are 50/50 male/female, and representative of our racial makeup (black, hispanic, etc, etc), if you want true equity and diversity in our workforces, then you simply have to do one thing:

Give the education/training/skills to EVERYONE.

And we are seeing the outcomes now that we are approaching our fifth year of this effort.

The NY Post (The NYC Mayor’s biggest critic) wrote this article about the performance and representation of young women in the AP Computer Science exams last spring.

Just 379 girls took the exam in 2016, compared with 2,155 last year, according to the department. That means that 42 percent of all city kids who took the AP exam were girls — compared with just 28 percent nationwide.


And the female students aren’t just sitting for the tests in far higher numbers — they’re conquering them at an accelerated rate as well.
In 2016, only 177 New York City girls passed an AP computer-science exam, officials said. In 2018, 1,266 earned the distinction.

https://nypost.com/2019/03/24/girls-are-acing-the-ap-computer-science-test-in-record-numbers-doe/

I am a huge fan of efforts directed at teaching young women to code, things like Girls Who Code. They are amazing resources for young women and they are part of why these numbers are moving in the right direction.

But I have always believed, and these numbers give me even more conviction, that the best way to get equity for everyone is to include everyone. Pretty simple really.

#hacking education

Comments (Archived):

  1. kirklove

    Right on!I recently attended a great STEM seminar at PS58 to get and KEEP young girls interested in STEM. The most alarming trend was by high school almost half stopped electing STEM pursuits. And by college that number dropped to almost 1%.We need to actively and consciously buck that trend. Provide it for all. Encourage all. And support all — all the way through.

    1. fredwilson

      Yuppppp. Turns out the women are just as good as the young men at this stuff, obviously, and that is a big part of keeping them interested.

      1. JamesHRH

        Sweeping generalization that, while possibly true, is no more evidence based than the status quo view.Some women are good at coding, just like some men are good at coding.Mostly because they like the work.Personality trumps gender.

    2. Michael Preston

      There’s a very promising program here called WiTNY — Women in Technology NY — that’s a joint venture between Cornell Tech, CUNY, and corporate partners to keep young women engaged from high school through college and beyond. At CSNYC / CSforALL, we’ve had numerous rock star WiTNY interns who have done work for us in data science, web dev, and design.http://www1.cuny.edu/sites/

    3. Richard

      It’s can be a bit boring !

  2. Mike Zamansky

    This is great. HUGE amounts of progress in what’s really a short period of time.I’d love to see more granularity in the stats though. Take the 2,155 test takers. Brooklyn Tech should have somewhere between 550 and 600 girls in 10th grade and they just made APCS-P a requirement. That means that one school possibly represents 25% of the test takers and a greater % of the new test takers and Brooklyn Tech was moving in the direction of requiring a CS course and APCS-P in particular independent of the CSforAll movement.It might not sound as good but I think the really impressive numbers are the number of schools that weren’t offering CS and now are even if it’s only for a few kids. That’s what shows we’re on the path to a big win.

  3. kidmercury

    if you’re a person who wants to see everyone given opportunity and freedom, but doesn’t require this to translate to equality of outcomes because you find scientific arguments for sex differences to be compelling, i just wanted to remind you what you already know but what society is increasingly unwilling to acknowledge: that you’re not a bad person, and your reasoning is not immoral.

  4. DJL

    I love the idea but do not think the “equality push” is always the right motive. For example, our school system (FBISD) is incredibly diverse. One of the most diverse in the nation. So you don’t have to force the diversity – it is already there. But because the students come from fairly well-off families, we don’t have any well-meaning millionaires willing to lead the charge to get the CS education. So there is no program in place.So ironically, you now have “poor’ kids in NYC getting access to life-changing resources that are not available to the kids in “flyover” suburbs.

  5. Kirsten Lambertsen

    Those numbers are a total wow! Looking forward to the day when inclusion doesn’t require a multimillion dollar public private campaign 😉 But things like CS4All are where it starts.

  6. awaldstein

    Too great!Yet another proof point that inclusivity breeds diversity and engenders change.Thanks for being a leader Fred.

  7. MrHatken

    I applaud your effort and suggest everyone needs some familiarity with Computer Science (as they do with Maths and English).However, personally, I would suggest perhaps a better way to spend everyone’s time, effort and resources would be to teach hypothesis testing, particularly using the fastest and lowest cost ways to test each hypothesis.IMHO, getting people thinking and acting more like scientists would be a major plus for all aspects of life and society but especially for entrepreneurship and startups, i.e. Lean Startup with Customer Development.Of course, they aren’t mutually exclusive.

  8. jason wright

    To round out and contextualise your underlying point, in what way are girls being denied equity in computer science?

    1. fredwilson

      societal bias mostly

      1. jason wright

        i still think that most issues of inclusion/ exclusion are a function of economic bias. It’s a circular argument though, and the debate centres around which comes first, economic bias or social/ cultural bias. If we change laws of inheritance (i.e. tax the inheritance of wealth to death) then we aid change. The priority should not be the private accumulation of wealth. It should be the creation and then distribution of wealth across society. If that isn’t happening then ‘society’ doesn’t really exist, and that sadly is where we are at.

        1. JamesHRH

          I think your bias is tops.Fred’s is second.And of course, the 50/50 split is artificial. It’s now a cliche that Scandinavian men & women have more gender freedom at work than anywhere else….. and choose traditional gender occupations @ traditional rates ( few male nurses, few female mechanics ).

  9. Pointsandfigures

    Cannot say I am surprised. Women are as good as men at many things and are just as capable. Good for them, and echoing @kidmercury:disqus , as long as outcomes aren’t guaranteed this is a great thing. Opportunity for all and freedom to choose.

    1. sigmaalgebra

      I don’t catch you making mistakes very often, but here you are clearly, badly wrong!!!! For yourWomen are as good as men at many thingsWell, literally, yes, but women, we’re speaking of generalities here, are MUCH better than men at lots of things.Leaving out the important, often even crucial, but subtle social and psychological stuff, women have better verbal talents and better manual dexterity. They are better with colors if only because men, at least of Western European descent, about 25% of them, are partially red-green color blind, that is, don’t see reds very well. In long term, say, weeks, demanding physical situations, women have better endurance; in simple terms, but there’s likely more to it, they start with more body fat and can keep burning it off as men just starve. Women are much less vulnerable to hardening of the arteries. I believe that girls and women have much better rote memories than men.My wife had much more sensitive sense of taste than I do or did. My experience is that girls and women are much better than men at going without sleep. Net, with the better verbal skills, manual dexterity, rote memory, and needing less sleep, my wife was unbeatable in the all night type up a humanities course term paper and get an A challenge. With better verbal skills and memory, she beat me like a rented mule in Scrabble.A quite bright girl just my age I knew as a college freshman, in the sixth grade got sick for two weeks and used the time to read all the Shakespeare plays and apparently understood them. I still can’t understand them if only because the language has changed so much and I’m so poor at guessing what was meant 400 years ago that I flatly can’t read them — they might as well been written in Russian, and there I don’t even know the alphabet. In K-12, when it was time to show that had memorized some poetry, she just arranged to be one of the last and did her memorization from the recitations of the other students! For some of the stuff girls do, there’s just no hope of competing!For a striking, GOTTA be convinced, watchhttps://www.youtube.com/wat…This is the J. Strauss waltz, in English, “The Blue Danube” played by some orchestra, likely in Vienna. But the video is also a ballet — can choreograph a ballet to a waltz? Apparently!Well, the stars of the show are four very pretty young women, about 16 or so. But along with them are four boys about the same age. Okay, okay, we can rush to the bottom line: The boys are there something like fence posts just to hold up the girls! But in more detail, the girls are gorgeously, elegantly graceful, and the boys, poor guys, are, uh, WOODEN, stiff. NO mistaking the difference. Yup, if look closer can see, e.g., that the wrists of the girls are limp, but there’s MUCH more to the differences than just that. Ballet is NOT easy and, boy or girl, have to be a good athlete, but, still the girls are DIFFERENT and totally the stars of the show.I’ve posted the link to this video before, but apparently the lesson is still needed. That video has some girls and some boys: See any differences or still believe that they are all the same, equal?Also can argue strongly directly from the ballet that these differences go back at least 40,000 years — I’ll leave the details as an exercise.There is more evidence: At some things, women are MUCH better than men.Yup, we’ve got some people who want suddenly to argue with, ignore, and reverse 40,000 years of Darwin’s work. That effort is dumber than AOC’s Green New Deal that the US Senate just voted down lots to zilch — first smart thing the Senate has done since when?AOC, is that a new spelling of ditsy, wildly overly emotional, irrational, self-destructive, where outrageous incompetence suggests helplessness, dependency, and, thus, a reproductive advantage, bimbo? I can only hope her father, an uncle, a brother, or a good husband will take care of her — she needs it.Yup, good with things!!!! The best case, my over the tops of the fences two states over, grand slam, never to be equaled in the history of this universe was when I was 14 and the girl was 12, by a wide margin the prettiest female I ever saw, then or since, in person or otherwise, her high school grad picture is in the UR corner of my screen — drop dead gorgeous. I would have fought with bare hands hungry lions and tigers to have protected that girl, and being so determined I might have won!But her mother’s vacuum cleaner didn’t work!!!! WHAT a soft ball pitch right over the plate, WHAT an opportunity! So, sure, no doubt she’d pulled on the line cord a few too many times. So, I got out my pocket knife and small key chain multi-bladed screwdriver, took the cover off the switch, cut off the broken wire, twisted in place good wire, reassembled, and pleased the mother and amazed the girl. No greater grand slam in all of history! Did I mention the prettiest female ever? GOD and Darwin totally had their way with me — I totally loved that angel.In some other things, I was much better than even the brightest females I knew. E.g., my wife was K-12, college, etc. a much, Much, MUCH, MUCH better student than I was and that if only because she was that much better compared to nearly anyone — Valedictorian, Summa Cum Laude,, Phi Beta Kappa, Woodrow Wilson Fellow, NSF Fellow, and more. She took great pride in that and for a long time got to look down on me.Finally in research, she lost it, and it was crushing for her: Without intending in any way at all, and with zip, zilch, and zero intention, to be competitive, I showed her how to do academic research publishable in some of the best journals: Just sit beside her on our bed while she relaxed watching TV and I did the research, occasionally asking her how to spell some of the words. It was crushing for her. E.g., as I have mentioned here before, in two weeks with the math of the Kuhn-Tucker conditions (KTC), I beat like “rented mules” two Nobel Prize winners in Economics that tried to use the KTC in mathematical economics. But she DID help me with the spelling! For my dissertation, before grad school I’d chatted with a prof at Cornell. He’d given me a 90 second lecture and, when my cab arrived, he gave me three words of advice. I thought about it on the plane, and when it landed I had the intuitive stuff for my dissertation — DARNED clever 90 seconds. Pro tip on research: I didn’t write down anything. Often in research SHOULD be able to see the important points just in your head.Using a terrific pure math course in my first year, independently in my first summer I did the fully precise, with theorems and proofs, math, and that was the research. The rest was writing some illustrative software, making it run shockingly fast [on the server I just built, 8 cores and 4.0 GHz clock, the code would run before could get index finger off the Enter key, with just one of the cores, but the code could easily have been written to make use of many cores, certainly all 8] and doing the writing, with my wife watching TV and helping me with the spelling again! Her struggles for her dissertation were so stressful that it was essentially what killed her. For me, the dissertation research was fast, fun, and easy, and that shocked her. She was MUCH better at spelling and LOTS more; I was better at research.Quite broadly a summary, that lots of people come to eventually if not right away, is that males are better with things and females, with people. On that, you can bet big bucks and take the winnings to the bank.Of course, we want to guess why: So, sure the guess is that going way back it was just crucial for men to be good with things and the women, with people. I’ll let you fill in the standard details about hunting and building versus homemaking and nurturing.Net, in a word, males/females, are different, in many ways, very different, already right at birth and likely earlier.More generally, much of this equality stuff is doomed to failure. Lots of people have been, are being, and will be hurt, and that’s a bummer.

  10. sigmaalgebra

    I’d say that for computers, computing, and teaching computer science there is a nagging, highly ironic point: Sorry, but for K-12 IMHO the best use of computers is for word processing to help the student learn to write, yes, technical material but, okay, also literary material.In particular, IMHO, and for the irony, the biggest bottleneck in computing is not coding but writing English, not writing code but writing English. So, in K-12 for progress in computing use the computers for word processing in English instead of coding in various computer languages.From all I have seen, all across computing, the industry is from inarticulate down to functionally illiterate and just in general unable to do well describing its work, either designs before coding or the final products after coding.Next, for the future of computing and the importance of computing for the students, looking ahead, and not necessarily very far, also more important than learning to code is learning mathematics, right, just the standard stuff from algebra, geometry, trigonometry, solid geometry, analytic geometry, calculus, linear algebra, and maybe some abstract algebra — a lot more important for computer science and computing than anything taught in computer science.For some early evidence, the computer industry is now getting all hot on what the heck? Yet another operating system, processor instruction set, programming language? Nope: The hot topics now are in machine learning (ML) and the current work called artificial intelligence (AI).Soooooo, you ask, where does the math come in? Here it is! Sit down for this one: Essentially all the value in ML/AI is from the core work which is essentially all math, essentially all of it stolen, uh, borrowed, from optimization and statistics.And, for doing better in both the methods and applications, the keys are, and the candidates are, instruction sets, quantum computing, operating systems, networking, optimization, statistics, and other math, and may I have the envelope please: Drum roll, please. And, “rip”, we have a three way tie!!! The winners are optimization, statistics, and other math.People might get at these topics via Python, R, Mathematica, or whatever, but that’s where the power and value are in both methods and, now, new applications.Yes, I know; I know: It’s totally obvious that the theory, main tools, and methods for an important application of computers comes from computer science. Totally obvious but nearly totally wrong.So, for computers, first use them for word whacking to learn better not how to write code but how to write English. Next, use the computers to help learning math, especially the word processing via Knuth’s TeX.Next, learn about the real world where the problems to be solved are.The coding? Trivial.Computer science? Well for some years it extrapolated from what is in Knuth’s The Art of Computer Programming. Then it hit the question P versus NP with a lot of the field grinding to a halt. Otherwise it is off borrowing from math.Q. How much math should a computer scientist know?A. More.Q. What should a person with high interest in computing pay attention to now?A. A good bet is D. Bertsekas, long at MIT, as inhttp://mslgoee.asu.edu/Medi…Yes, it is easy enough to notice that here for K-12 a still pretty good curriculum concentration is just the 3Rs.Disclosure: Nothing here is close to the crucial, core math of my startup.

    1. Peter J. Mills

      Interesting material, but surely it belongs in a blog of your own.

      1. sigmaalgebra

        What I wrote is on-topic — education in, for, computing.

        1. Peter J. Mills

          Equity and diversity is the topic.

          1. sigmaalgebra

            Well, above it happens that I’ve written about that, too!”Equity”? What a horrible, cruel, brutal thing to do to K-12 girls and college young women — HORRIBLE. They should be cared for. Equity, a HUGE step down.Besides, for the topic key words, Fred used “hacking education”.

  11. Ling Richard Yan

    I don’t understand why a 50/50 male/female ratio in software developers is a desirable goal. Could someone please send some light on this?Isn’t it possible that, at a time when (a) an equal opportunity of access to education is provided to both genders, (b) the societal perception that somehow discourages female participation in the engineering fields, and (c) employers of engineering jobs treat both genders fairly in recruiting process, there is still imbalance between the genders due to other reasons?I feel (a), (b) and (c) are already quite true today. And so if the gender gap still exists, then it’s a matter of different interests and choices between the genders.Evidence for the three bullets being true: Many women in science program have existed at the national, state and college level since at least as early as the 2000s. And computer science is one of those subjects where the quality of your work is easily checked against objective metrics; so hiring would also be more fair across genders.There are no campaigns to encourage men to become nurses or nannies, or to get women to become welders or plumbers. This is because there is a natural tendency for men and women to gravitate toward different careers. Same goes for the engineering fields.Separately, though, training teachers to provide CS education for students of low-income families is a great idea.

    1. JamesHRH

      It is a philosophical belief that superficial equity of outcomes is desirable.Symbolism driven.If I can’t see me (e.g., a gender fluid dark skinned person ) doing X, then I cannot / will not / am way less likely to even try to do X.Not everyone agrees, to put it mildly.

    2. fredwilson

      it is about opportunity

    3. Richard

      Consider it an experiment – it may prove to be highly beneficial or it may not. If you don’t test the hypothesis you’ll never know. We know very little about genetics today and how it correlates to invention. Let’s test things and find out. Who knows, maybe we find a needle in the haystack ?

      1. kidmercury

        we know there is a cost, in that the capacity for professional software engineers at top firms is not unbounded. i.e. google will not hire every software developer that applies. by setting outcome as the KPI, this sets the stage for outperformers in non-preferred demographics to be penalized. back in the day this was called racism/sexism, but i understand those terms can now only be applied to members of unprotected classes.the scenario i’ve presented is not hypothetical. it is very real, as illustrated by the harvard vs asian americans court case now going on.

        1. Richard

          I agree with you at the college level and in general, but in this situation (based on the demand for programmers) and on my experiences with the just how how pathetic the outcomes are with NYC Public Schools, this program (unfair or not) is a nonissue (for the time being).

          1. kidmercury

            sure. there is a difference, though, between the program, which i agree is awesome work, and how the program is measured at times (optimizing for uniform distribution in outcomes), which i think is essentially racism/sexism 1.0

          2. Richard

            Outcome points take some time to optimize. Probably should model it with a hazard ratio / survival analysis.

          3. sigmaalgebra

            Likely the biggest variable in K-12 performance is in the home. Send all complaints to Darwin c/o Mother Nature.

      2. sigmaalgebra

        I’ll borrow from a movie: “You are dealing with forces you cannot possibly comprehend.”More simply, there has long been productivity and economic advantage to have, borrow from another movie, women “be more like a man”. But, gee, wonder why, somehow that didn’t work so well?Or starting right away, as early as three years old, boys see little girls as pretty, sweet little things, maybe sweet like Mommie. The boys see themselves as small versions of Daddy, the girls as small versions of Mommy. The boys play with toys like Daddy uses at work; the girls play with dolls like Mommy works with.When you get in high school and college, you look at, want to date, buy dinner for, keep thinking about, what the heck, some 250 pound woman who can bench press twice her weight or some really cute, sweet, precious, darling, adorable, needs help rewiring a table lamp, 5′ 5″, 110 pounds, 38, 22, 38, long legs, small, straight nose, big eyes, great smile, LOTS of feminine gestures, high, sweet voice???? You tell me: You REALLY want the weight lifter?? Really???Gee, this is all really too complicated????

    4. Adam Sher

      50/50 obscures the mission of providing CS education to a school district that needs support. As Fred said above: Give the education/training/skills to EVERYONE. which is what he’s done and doing. (c) is indirectly refuted by research from Dana Kanze and others. Kanze’s research shows VCs ask men and women different questions, which leads to different funding outcomes (Kanze Huang Conley Higgins 2018). There is a lot of evidence that managers, hiring professionals, etc… just the same work by two different people differently based on their gender. Even humor in the workplace is judged differently by gender (Evans Slaughter Ellis Rivin 2019). Some combination of psychology and culture is impeding our ability to neutrally evaluate the competencies of candidates and employees. Much of this is behavior is easy to mitigate if one is aware.

      1. Vasudev Ram

        >Some combination of psychology and culture is impeding our ability to neutrally evaluate the competencies of candidates and employees. Much of this is behavior is easy to mitigate if one is aware.You nailed it. Awareness is the key, for this as well as any other issue. If you’re not cognizant of it, you cannot do anything about it. Bring issues to the surface.

        1. Lawrence Brass

          To implement inclusion we first must undersatand how exclusion works.In many cases what is needed is to remove the barriers.I think that equal access to education is the most important factor, tagging it as a “socialist” treat is so dumb.

      2. JamesHRH

        There axes of social interaction:- status- feelings- logicPeople have to balance all 3 to make excellent decisions.Most people don’t & some people can’t.

    5. sigmaalgebra

      > I don’t understand why a 50/50 male/female ratio in software developers is a desirable goal. Could someone please send some light on this?ABSOLUTELY I can “send some light”!!!It’s simple, really simple: Just like we need a 50/50 male/female ratio in the NFL, NBA, MLB, PGA, soprano roles in opera singing, and, especially, no question, motherhood!!!!Send all special concerns to Darwin c/o Mother Nature.Somehow I get reminded of the old TV ad closing with “It’s not nice to try to fool Mother Nature” IIRC followed by a big bolt of lightening.I tried for some decades to go along with that stuff about Mother Nature was wrong. I learned my lesson. Paid full tuition. At first I was following the traditional track, but I thought I’d found an exception. I hadn’t. I’ll never make that mistake again. And my children certainly won’t because my mistake cost me all my chances of ever having any children.I’ll put it to you this way: For women equality would be really bad, a really big step down, e.g., from the traditional “women need to be cared for”. That part of culture was a lesson obtained at no doubt horrendous, mind breaking cost. We don’t want to pay it again. But we are: Currently the birth rate is so low we’re going extinct, quite literally and quite rapidly. But Darwin is on the case: In a few more generations the fraction of weak, sick, dead limbs on the tree will be much smaller, and humanity can get back to being successful again.I’ve guessed, more and more yearly, that starting with Betty Friedan it was some darned clever Russkies who saw that just talking US women into hating traditional marriage and motherhood would do more damage to the US than the Russian military could ever do. So far that’s been the case. Add it up: How many humans have the feminists killed? More than Hitler, Stalin, Mao, Pol Pot? More than all of them together?Go ahead, keep talking this 50/50 stuff; in a few more generations there will be nearly no one around willing to listen.

  12. JamesHRH

    Your code is brutal.The IF statement is faulty even though the THEN YOU HAVE TO statement is bang on.

  13. Richard

    Any Spanish language instruction? The country has made a big bet on immigrants from our southern neighbors. A return on our investment is going to be dependent on improving the skill levels of these young kids. California not too long ago was 1st in primary education levels of high school kids and is now 48th!

    1. fredwilson

      yes, we teach comp sci in spanish, particularly in elementary school. i have attended some of those classes.

      1. Richard

        I’m not sure who the hipanic equivalent is to a Larry Ellison etc., but a poster or two of one would surely help younger Hispanic kids feel that getting into the field is within their reach.Also, be sure these skids understand opportunities still ahead. Once thing academics get wrong again and again is characterizing the game as if we are in the 7th innning rather than the 2nd.

      2. JamesHRH

        This is a great thing.But hopefully it’s supported w English.

  14. Guy Lepage

    But here’s the thing. If you want to see engineering teams that are 50/50 male/female, or representative of our racial makeup (black, hispanic, etc, etc), if you want true equity and diversity in our workforces, then you simply have to do one thing:Give the education/training/skills to EVERYONE.I applaud folks working on this but this should be handled by the board of education. No?k-12 is such an important educational development period for individuals. This is something that should be fought by all. Not just by those in tech. Strong k-12 education raises the economic output.It’s not just the ethically right thing to do… It makes great economic sense.

  15. Tim Horan

    Big congrats Fred.I know there’s still lots to be done on the equity front but well done to all involved for making such great progress in a relatively short period of time. Impressive.

  16. Richard

    I agree but given the demands for programmers and the cost to society if we don’t improve the productivity of our workforce, Spanish led programming is a short term fix, patching a pipe, that will save the house while we make the long term changes necessary.

  17. Matt A. Myers

    Whether teaching more of everyone leads to swaying more diversity into those fields is at odds with research, however I feel developing the skills of programming is good for developing logic gates in the brain – if anything to work towards a good foundation to help people detect and have integrity.

  18. Richard

    Who knows ? Maybe it will improve outcomes ! We test things all the time. Might as well get as much data from Freds work as possible. No debate on the importance of English, but – based on my time in Los Angeles – we have some work to do. If we can improve the productivity/outcomes of Hispanic kids – it’s a win win.