CS Education Week In NYC

All over the city this week, students in NYC’s public schools have been celebrating CS Education Week by doing events and hackathons to showcase their coding skills.

Through NYC’s CS4All program, over 1000 teachers have been trained to teach CS classes in their schools. That is over 500 schools to date. Over the course of the ten-year CS4All program, over 5,000 teachers will get this training so that all 1700 school buildings in NYC will have at least one CS teacher and many will have two, or three, or even four.

Most of these 500 schools, and many others around NYC, participated in CS Education this week. I was out in the schools along with my colleagues at the Department of Education, CSNYC, and the companies that support us, including Google, Accenture, and Alexandria Real Estate.

I met this eighth grader up in the Bronx at In-Tech Academy, a 6-12th grade school that specializes in STEM education and mostly pulls from the Kingsbridge section of the Bronx. He told me that he wants to be a game designer when he grows up. I told him he was well on his way and that he just needed to keep up his schoolwork and his excitement for coding and making things.

But it wasn’t just me out in the NYC public schools this week.

A bunch of Google engineers went out to the schools and helped with the hour of code. Google has developed a K12 CS Ed curriculum called CS First and Stephen Bloch was helping a student do a lesson from that curriculum.

The thing that most excited me this week was to meet all of the NYC public school teachers who have been trained under the NYC CS4All program to teach CS to their students.

This is a photo of a teacher named Ms Calise from Horace Mann, PS90Q in Queens, where a bunch of teachers have taken advantage of the CS4All program to learn how to teach CS skills to their students.

So, needless to say, this week has been very gratifying for me. CS Education is seeping into hundreds of school buildings in NYC and will continue to do so for the next few years until it is in every school building in NYC.

I am so thankful for the generous support of corporations and non-profits like Google, Accenture, Alexandria Real Estate, Hearst, AOL, Two Sigma, Wachtell Lipton, Math For America, Robin Hood, Hutchins Family Foundation, Paulson Family Foundation, and many many others, without whom this work could not happen.

If your company or non-profit wants to join this group and help bring CS to all students in NYC, please email me or leave a comment in this blog post and I will contact you.

#hacking education#Uncategorized

Comments (Archived):

  1. JLM

    .Stronger than an acre of garlic. Well played, Freddie Wilson. You are a mensch.JLMwww.themusingsofthebigredca…

  2. Richard

    I’m in the middle of reading the story of James Conant, warrior scientist.While it’s nice to hear of kids going into gaming, I hope teachers understand the uses cases of CS into the Core Sciences and have kids aim higher.

    1. sigmaalgebra

      Ah, IIRC, Conant was one of, say, the architects of the post-WWII US STEM progress. One of his ideas was to have so many sources for funding research and education that no one could turn them all off at once!So, no doubt at least partly due to Conant, we have an alphabet soup of funding sources — NSF, NIH, DARPA, ONR, NRL, JHU/APL, Air Force Cambridge, Army Durham, more Navy labs, lots of Department of Energy labs, etc.And, don’t tell anyone, long the leading US research universities got about 60% of their annual budgets covered by the 60% or so “overhead” from faculty research grants from NSF, NIH, etc. So, that’s where the English, Art History, French, etc. departments got their funding! And the bronze statue of Dean Markus Brody, the all-glass coffee shop, the new workstation computers for the faculty, the limo for the president, etc. And one way and another there was a lot of funding of the social sciences.Uh, did I mention, don’t tell anyone!

  3. Andrew Trenk

    What else can we be doing to engage the NYC tech community in CS education outreach? We’re capable of having a much larger impact if more people got involved. For example, the past few NY Tech Meetup newsletters didn’t even mention CS Ed Week.It would have been great if there would have been a big push to get people to sign up to volunteer at https://code.org/volunteer. Then the DOE could have made sure that teachers were aware that they could use that site to find volunteers. It’s too late for this year but maybe this could happen next year. I’m happy to help somehow if anyone wants to work on making this happen.

    1. Rob Underwood

      Andrew, as you and I have talked about offline, I agree with you. More volunteering by the tech community in NYC into NYC public schools would be great. I also think the schools are now in a much better position to make better use of volunteer time – for volunteers, teachers, and students – than maybe 2-3 years ago where visits were often more akin to career days (learn about what a programmer does). Now there is more meaningful co-teaching and mentoring to be done.That said, and as we’ve also discussed off line, coordinating, dispatching, and prepping volunteers as well as the schools they’ll visit is a big job. There are volunteer driven programs like ScriptEd and TEALS and a significant part of their operating budgets I’m guessing is to attract, engage, and prep volunteers, many of whom have no teaching experience, to be ready to be most helpful in a class or after school setting.So I go back to what I think remains an unfilled need in NYC: a single unified place for CS education related volunteer recruitment, dispatch, and prep. A volunteer database that can match volunteer directly to both schools and, in other cases, programs that need volunteers (e.g. TEALS, GwC).But that’s a lot of work and I don’t think that work – finding, prepping, and dispatching volunteers – can and should be itself by volunteers as I just don’t think it’ll get done, at least not well, especially as volunteers I think would rather be in schools rather than dispatching other volunteers. And it’s not just something technology can fix — there is some of this which is volunteer professional development/training. (And some orgs, like NYC Service, have proven impractical for this work).Anyway, you’ve heard this before as you’ve been trying to get this solved but just sharing here for others as every year now for 4 years a bunch of folks in NYC do a lot of ad hoc volunteer recruiting and dispatch during CSEd week and I think with a little organization and funding for that organization it could be done in a better way. But so far it’s been hard to get funders to see the value in funding this meta work of volunteer management compared with other ways to support CS Ed.

      1. Andrew Trenk

        I think the key thing here is finding a way to make a strong connection between the NYC tech community and schools. Most people in the tech community are capable of volunteering, but few have connections with schools, so they are not going to volunteer unless someone can connect them with a school.

    2. fredwilson

      Hi Andrew. Your enthusiasm for this work continues to amaze and impress me. Thank you for that. I think the tech sector needs to own this more than the DOE. I have a lot of respect for the work that the DOE does in this area but they aren’t great at tech industry outreach. Organizations like Tech:NYC and CSNYC can and should do more here. Folks like you and Rob have done a lot but we need to do more. I’d be happy to convene a meeting of the interested parties to discuss it

      1. karen_e

        Please consider me an interested party. Though not a coder, I’m well connected to the tech community in Boston + the Boston Public Schools + have facilitation experience.

      2. Mike Zamansky

        Hey Fred – you know I’m always up for this

      3. Rob Underwood

        Relevant update: bit.ly/2AkWNpR

  4. jason wright

    are those Chromebooks? love the concept. very democratic.

  5. Bryan Rosenblatt

    I’m an investor in Bitsbox, which is a monthly subscription product that teaches kids to code. (www.bitsbox.com). They are getting popular with teachers and their classrooms as well. They’re currently running a “buy one, give one” program with a childrens hospital but I’d love to see how this can be extended to more NYC students. ([email protected])

  6. maxnuss

    Is there a similar program in Chicago?

  7. sigmaalgebra

    Draft only, due to a server delay issue.See actual post just above.

      1. sigmaalgebra

        +Yes, looks, as I wrote, “impressive”.But also as I wrote, are they in for a severe, harmful “reality check” later?

        1. Rob Underwood

          Maybe. But it’s not really how I approach it, and I’m guessing may not be how our host and others working on this approach it either.It’s more about making a bet that CS could help students in a myriad of ways, some expected, some unexpected, and some still we can’t imagine yet — that students are likely (much) better off with it than without it, especially when it can be integrated into other subjects (e.g., chemistry students using Python and a library or two to do data analysis on experiment data).I do worry that the lower end, more easily commoditized/automated developer jobs may not be there 10-15 years in the numbers all of this work would imply, and so it’s really important we’re taking a broader view of CS that includes some concepts (e.g., algorithm analysis), a bit of theory, as well topics like as cyber security, network engineering, data science, etc., and not just teaching kids coding that implicitly and too directly assumes they are all going to grow up to be building web sites in Rails or apps in Swift as they probably won’t.Theme: try and direct the kids to where the puck will probably be, not where it is now.

          1. sigmaalgebra

            Right.Read again what I wrote and see that we are essentially in agreement except:(1) You are not sufficiently acknowledging the potential downside of some students who get hurt now from feeling discouraged and inadequate and/or get hurt later with a reality check that destroys their dreams. Broken hearts are ugly things.(2) You did mention, and I omitted, that, sure, the computing they are teaching need not be called, dumb terminology anyway, “computer science” but, much better, computer usage. E.g., when they teach driving they don’t call it automotive science or even automotive engineering or even engineering.So, instead, call it computer usage.Then emphasize how useful it is for looking up material on the Internet, writing papers, creating graphics, working with still images, video, and audio, generally creating media, good for homework and term papers in all subjects, including even English, e.g. get spell checking, maybe some good help with grammar, maybe some help with translation from French, German, Italian, etc., data analysis in chemistry, physics, biology, etc.FINE!Data science? Maybe I’ll type in some, e.g., short and sweet stuff on eigenvalues and eigenvectors, the polar decomposition, and factor analysis.E.g., factor analysis in a big hurry: For positive integers m and n, have some n questions and, then, get answers on each of the questions from each of m people. The number of questions might be n = 200. The number of people might be m = 100,000.Then notice that can do well for each of the m = 100,000 people with just 15 or so numbers instead of the 200.Commonly each of the 15 numbers is close to some intuitive meaning.So, put the answers in m x n matrix A. Let A’ denote the transpose of A so that we have n x n A’A which is symmetric positive definite. So, there exists n x n orthogonal Q and diagonal D = [d_ij] with d_ii >= 0 so that Q’A’AQ = DThen directly from that equation, the n columns of AQ are orthogonal.Without loss of generality, assume d_11 >= d_22 >= … >= d_nnThen typically in practice, the first 15 or so columns of AQ do well approximating, right, in least squares (easy exercise now), the whole A.So, in row i = 1, 2, …, m of AQ, the first 15 numbers do well describing person i.That’s just a start: Right away we can just read off a nice list of powerful, valuable best possible least squares properties, all close to a generalization of the Pythagorean theorem.The Q we get can commonly be regarded as something we have estimated statistically and can be used on answers from people not in our original m. Etc.Can do the arithmetic with Linpack or, no doubt, also R, Mathematica, or Python. Can call Linpack from Fortran or with a little work, C, C#, or Visual Basic .NET.This stuff gets to be important in marketing and was long a big part of educational statistics.That the 200+ shrinks down to 15 or so is curious, powerful, valuable stuff.Exercise: Prove or disprove: Anything can do with regression analysis with A can be done nearly as well with just the first 15 or columns of AQ.Oh, BTW, “Anything that calls itself a science isn’t.”

          2. Michael Ball

            Let’s also not forget, programming is only one aspect of CS.The current crop of CS curricula have a heavy portion of social implications, algorithms, and other fundamental topics that come with computing. Even students who aren’t programming in their future will benefit from a better understanding of a changing world.We teach kids science and social studies and art not because we necessary expect them to join those professions or apply that knowledge daily, but because it expands minds and hopefully enables them to be a better practitioner of their chosen fields.

  8. sigmaalgebra

    Impressive.But leading eager, gullible middle school students down a primrose path is neither very new or difficult.I’m not sure just how much good such K-12 computing will do. I’ve learned, used, and taught computing and now my startup is based on computing, right, so far software in 100,000+ lines of typing, with some new math, some old algorithms, and some new algorithms.Still, my guess is that with computing in K-12, a lot of students will feel left out and inadequate. And a lot of students will expect good starts, which will at most rarely happen, on good careers, which will at most rarely happen, and get discouraged and might give up on learning, careers, and themselves.It’s not nice to mislead young people by giving them false expectations. Maybe it looks like a big upside but then there can be a big downside.Maybe better examples are (1) the Dangerfield movie Back to School where his father had a business in the garment trade and his son went on to success in that business and (2) where Donald Trump’s father was successful in NYC real estate and Don went on to big success in real estate. In either case, K-12 or even college didn’t have much to do with it.So, with computing in K-12, there is a pier, and a lot of students will soon get discouraged or go the full distance, off the end, and all wet. I hope not too many students get hurt too much.I’ve seen schools, kindergarten, K-12, college, and graduate schools, badly hurt a lot of people. It’s not pretty.I’ve see some really good students get really badly hurt from discovering that they had been misled and dedicated to something unrealistic and even false.E.g., from the 1930s on, there was a big push that chemistry had the careers of the future. Movies showed white lab coats and bubbling test tubes. Then too soon the US was awash in chemistry students without jobs that even touched on chemistry.It is also highly relevant that for decades lots of big US capital forces have pushed flooding the US labor market with STEM field labor, e.g., took tax money from US citizens to pay for expensive college and grad school educations for immigrants when the US taxpayers were struggling to pay for college for their own children and where the immigrants without subsidies couldn’t afford even to live in the US on potatoes and cabbage. Bummer.Strong forces in Congress are still praising HxB programs for the “highly qualified” total upchuckable BS.Call immigration just what the heck it is — a new version of slave labor. For the associated jobs, native born US citizens need not apply.Computing careers have long been in a manipulated, managed market.Even if computing careers were not managed, bluntly, I’m reluctant to believe that there is much known about just what the heck to teach students in K-12 that K-12 has a chance of teaching and that has high promise of doing the students much good except just the 3Rs. Sorry ’bout that.E.g., if my startup is successful and I have to hire for software development, I will emphasize strongly in this order:(1) Ability to communicate clearly, especially in technical writing. Or it’s next to useless to write software if can’t document it to users and/or other programmers. The documentation is more difficult and important than the programming — in the end, by wide margins.(2) Good skills with some math, not arithmetic but high school plane geometry through college theorem proving courses in linear and abstract algebra.For computer science? Haw! Trivial stuff the useful parts of which I can teach in a few lectures in a week. Actually more than a good start on the whole subject is in the two posts I wrotehttp://avc.com/2017/12/do-a…http://avc.com/2017/12/un-s…Literally. Remember, I’ve taught computer science to ugrads at Georgetown and to grads at Ohio State. When I say “trivial,” I mean it.Software development? There are a few big but vague lessons, then the need is for good documentation (tough to find) on the basic tools. E.g., the Microsoft documentation on SQL Server is so bad that can struggle throwing guesses at a wall for a solid week just to get a “connection string” that works. No joke. That’s not math, science, engineering, or computing but just pure BS.And That’s All Folks!What’s being taught in schools in computing? Likely next to useless and hopeless. Sorry ’bout that.But, for the recent Hour of Code, I got sucked in and posted my contribution:http://avc.com/2017/12/do-a…That post touches on some of actual reality in computer science and computing. Actually, so does my posthttp://avc.com/2017/12/un-s…From those posts, I just hope no students get hurt.In interventions, there’s what is now a wise, old rule:First do no harm. Uh, maybe athttps://www.usv.com/blog/tr…some of the students will want to get a “reality check:”So, we may love the idea but until we see the implementation in a live setting (closed beta isn’t live enough for us) we do not feel like we can assess the social adoption issues that are so critical. We don’t care if it’s called beta or commercial code. As long as it has been in the market and we can use it, see how others are using it, look at internal weblogs, and external services like Comscore and Alexa, we can make a good assessment of the adoption curve.We are happy to talk to entrepreneurs long before they launch. In fact we spend more time doing that than pretty much anything else these days. We want to be able to watch a team take an idea from concept to execution. We learn so much from being able to do that. But until its launched, with tens of thousands of users (ideally hundreds of thousands) using it, we are unlikely to invest. On “social adoption issues,” not all consumer-facing Web sites are at all “social”. A serious entrepreneur would want a candidate BoD member to understand at least this much.At common, current add rates, that many users could result in $30,000 a month in revenues with next to nothing in expenses. Okay. That’s some darned good news.But, really that statement on “traction” is telling the students that, if they want to start a business in computing, they have to take it from nothing to a quite nicely successful business on their own with no funding from venture capital. The students need to know this fact of life. Maybe now we are starting to help the students.

    1. Twain Twain

      Next Tuesday, I’m at the Science Museum in London showing 13-14 year old STEM students how to make AI apps: Alexa voice apps (Scientists and Discoveries) and a vision recognition app.@fredwilson:disqus — What do female engineers do differently? We make sure lists of scientists in our app are as representative as possible.* https://uploads.disquscdn.c

      1. PhilipSugar

        This is my office (the carpet is the same in every conference room around the world: London, Dubai, Singapore, Shanghai, Sydney, LA, MN, and DE.This is why you start early.This is why you keep up the fight.Notice the tape line. We make sure that nobody gets excluded. You stay behind the line when the coach is coaching. (See my Son’s barefoot?) Also see the Mandarin on the wall.

        1. Twain Twain

          Thanks, Phil, you’re an inspiration. It’s easy to be indifferent and complacent and to fool ourselves, “Oh, my kids are okay, it doesn’t apply to me,” but that’s the wrong way to think about the world because the more hands there are to build the future, the better the future becomes.I’m now the only female engineer of 16 (mostly PhDs) on the Consciousness Prior AI project for Natural Language Understanding led by Yoshua Bengio, one of the three “Godfathers of Deep Learning” (the other two are Geoff Hinton of Google and Yann Le Cun of Facebook — and they focus on vision recognition).@fredwilson:disqus — When we talk about opportunity and discrimination, a lot of us aren’t aware that the logic and the language of the machines discriminates against entire communities in the population because it is “mindless AI”:* https://www.fastcompany.com…And Stanford and Google’s algorithms are major factors in propagating those biases.Motherboard: “The problem is the API labels sentences about religious and ethnic minorities as negative—indicating it’s inherently biased. For example, it labels both being a Jew and being a homosexual as negative.Google’s sentiment analyzer was not the first and isn’t the only one on the market. Sentiment analysis technology grew out of Stanford’s Natural Language Processing Group, which offers free, open source language processing tools for developers and academics. The technology has been incorporated into a host of machine learning suites, including Microsoft’s Azure and IBM’s Watson.”So that explains the MS Taybot fiasco.https://qz.com/1079284/antihttps://motherboard.vice.cohttps://www.washingtonpost….https://uploads.disquscdn.c…Yes, the system I invented is ahead of the market. No, it can’t be solved by Blockchain/ETH.I KNEW these problems wrt historical biases in the data and the inadequacies of the algorithms would surface, so I made a system to reduce those biases and to enable better insights and understanding on people’s language and perspectives.Re. opening up opportunities, we do what we can for small groups in our communities without realizing that, at scale, the algorithms are risk managing those communities out of their potential.Those systemic problems at the roots won’t change unless female engineers are there, shaping conversations and the code. https://uploads.disquscdn.c

          1. PhilipSugar

            I love it. My Dad had a degree in Operations Research…….and the company that bought us is in Montreal. Is that roughly where you are?

          2. sigmaalgebra

            > Operations ResearchGood grief: For a while I was Director of Operations Research at FedEx. Then I went for an applied math Ph.D. with much of the work close to OR. My two dissertation advisors ran something they called the “Operations Research Center”. To support myself and my wife through our Ph.D. degrees, for a while I took a job in OR for the US Navy. Yup, the work was several applications of optimization, stochastic processes, in particular continuous time, discrete state space Markov processes, game theory, etc.My best prof in grad school was a star student of E. Cinlar at Princeton, and Cinlar was long the Editor in Chief of Mathematics for Operations Research. When I was a B-school prof, I recruited a person for the faculty, and she soon published five papers in Math for OR in one year and was doing grant reviews for the NSF, one review of an application by a famous guy! A good mathematician, a student of Choquet in France (Choquet was a member of Bourbaki, a pseudonym for several mostly French pure mathematicians who started writing astoundingly polished books on the foundations of math) noticed her and married her.OR and, heck, just say applied math more generally has to be one of the most important keys to the future of computing if only because the progress is needed and available and for that progress there’s no serious competition for applied math. Call the work whatever, but when it’s done well it looks like math, as solid as pure math, but already with at least one good application!OR was, likely still is, a big deal, like a lot in applied math, e.g., filtering and control in high end EE, quite influential within 50 miles of the Washington Monument, especially for US national security.Recently athttps://www.jstor.org/stabl…I noticed the book review by Sephen B. Maurer of the bookEvar D. Nering and Albert W. Tucker, Linear Programs and Related Problems, Academic Press, San Diego, 1993, 578 pages.Good grief, I hadn’t known. Nering had been an E. Artin student at Princeton. He wrote a relatively advanced book on linear algebra — no wonder being an Artin student — and it was one of the several books on linear algebra I read very carefully. In the back was a chapter on group representation theory, and that was the topic of my ugrad honors paper. Also in the back was a chapter on linear programming, and that was my first introduction of the subject. Based on that, once I got a job offer in the IBM Chicago branch office to help some high-end oil refinery people make good use of linear programming. Had I taken the offer, I likely would have met R. Gomory and gotten deep into linear programming quickly. Of course, later Gomory was head of the IBM Watson lab and, then, the Sloane Foundation. Well, no wonder Nering was interested in linear programming — he knew A. Tucker at Princeton.Well, from that book review was a story about G. Dantzig and his work in linear programming. Apparently there was an organized push on three fronts: (1) Dantzig at UC Berkeley would push the algorithms and applications. (2) Koopmans at U. Chicago would push applications to economic theory. (3) A. Tucker at Princeton would push the theory.The guy who was the Chair of my Ph.D. orals committee was big in applied math and optimization for transportation, right, essentially OR, and had been a Tucker student at Princeton. He liked my dissertation; since he was a Member of the US National Academy of Engineering, my Ph.D. was in less doubt!Of course, one of the pillars of OR is optimization — some of the work is terrific.Actually, with a polished, high end presentation of linear programming, that is, with the math done as well as the best pure math, linear programming has some surprisingly powerful results, e.g., in what might be called computational geometry, but much more. E.g., IIRC, there’s a nice proof of Nash’s result by Lemke and based on the math of linear programming. Sure, linear programming knocks off the von Neumann saddle-point theorem in two person games as a finger flick. And since linear programming right away tried to make progress in integer linear programming, it was, thus, long the first to study the NP-complete problems. So, linear programming identified one of the biggest problems in pure and applied math today.Then, sure, in optimization, keys for the non-linear case are the Kuhn-Tucker conditions. That’s A. Tucker as above and H. Kuhn, both at Princeton. A lot of the math there, and it’s some nice math, makes crucial use of some results known as theorems of the alternative, and a super nice way to prove most of those is just directly from the math of linear programming.When I studied the Kuhn-Tucker conditions, I saw a question, couldn’t find an answer in the library or anywhere else, and took the question on as a “reading course”. Well, I got lucky and had some ideas and did much more than a reading course: I got a solid solution to the question. The question was, are the Kuhn-Tucker and Zangwill constraint qualifications equivalent? One direction was known; my work settled the other direction.Doing that, I discovered that for positive integer n, the set of real numbers R, and closed set C a subset of R^n, there exists a function f: R^n –> R so that f(x) = 0 for x in C, f(x) > 0 otherwise, and f infinitely differentiable. So, any closed set can be the level set of an infinitely differentiable function. Closed sets that add interest include sample paths of Brownian motion, the Mandelbrot set, and Cantor sets of positive Lebesgue measure. So the smoothest possible function can have a level set the most irregular possible. Curious. For that result, I got some attention around the department. The work took me two weeks — fast course! With that work, it seemed that I had a halo! Imagine that, me with a halo! Mind blowing!That result let me settle my question about the Kuhn-Tucker conditions.Well, some economists did take optimization seriously, both linear programming and, later, the Kuhn-Tucker conditions. For the second, there was a famous paper by Arrow, Hurwicz, and Uzawa. So, right, the first two got prizes! Well, in their paper they had a question they didn’t answer. My work did answer the question.Ah, operations research!I was up all night mud wrestling with Windows trying to solve some problems that I shouldn’t have, are slowing my work on my startup, but I’m chipping away at. I discovered some wildly curious behavior!For the future of computing, sure, operations research, if only due to its deep and first to see connection with the P versus NP question in computational complexity, will have an important role. But, really, again, what’s key is just applied math, if you will, the whole QA section of a research library together with powerful, useful new stuff!

          3. Twain Twain

            My brother has the MSc in OR and is a data scientist. I’m in London. People are taking part in the project from different places.