The Hats Our Kids Should Never Wear

Have you watched television lately? I recently switched back to the Basic Service Tier (BST) service after what seemed like an endless love-affair with Netflix, Amazon Prime, MXPlayer and more.... What seemed like an ad-free, choice-led experience on a streaming platform quickly morphed into multiple monthly payments for multiple services, each with juicy binge-worthy stories and an endless collection of accurate recommendations that I just couldn't wiggle out of. I was becoming a slave.

BST has been a change, one I hated at first but that I have grown to appreciate lately. I had quite forgotten what hate-spewing prime time news felt like. The one thing though, that has assaulted all sense of reason, much more than even the news (truly), was the advertising - this new wave of advertisements targetting parents suggesting that kids should learn how-to-code NOW.

In one particularly revolting ad, the proud parents of a young child were amused by potential businessmen (sic) fighting over investment opportunities in an app that the child had coded. An outraged milkman/postman expresses disbelief and something that seems like jealousy...

Misplaced pride?

Almost the entirety of the advertisement can make one cringe. The inanity of the storyline, the implausability of the success of the app, the absurdity of matrix-style fight sequences between businessmen... Perhaps the worst of it, though, is the disbelief offered by the lower middle class worker, right at the end. The advertisement is unambigous, "coding" is the newest way for India's elite to now rub their riches in the noses of the poor and hardworking.

This advertisement isn't the only one of its kind and neither is the company. Amir Khan's endorsement of another edtech company, in the advertisement below, encourages involvement with "interactive online classes" while signalling that parents do not really need to be involved with their children's learning actively anymore;

Not good enough for your kids?

Of course nothing could be further from the truth. Decades of educational research has proven time and again, that the real difference in educational outcomes for children, regardless of income or background related factors, is parental involvement. Learn and read with your kids and they will do well. Tuitions help but they don't replace a great home environment.

In this study, Hattie found the 'parent engagement effect' amounted to the equivalent of adding an extra two to three years' education over a student's school career. Parent engagement includes setting goals; encouraging good study habits; valuing enquiry and experimentation; and modelling enthusiasm for learning and reading.

This article isn't a tirade about values or parents' misplaced aspirations. In fact I quite agree that studying is overrated. The future of work is unpredictable, and most of us can say very little about what kind of jobs our kids will have, two decades from now. Equally true is the fact that Indian kids need to learn to question parental judgement more. Unfortunately the current obsession with coding isn't doing much for either the goal of pointless studying or guess-estimating the future.

I grew up largely without any career guidance from my parents. It didn't help. Stuck between being a half-writer, half-economist with no technical or professional "full-stack" skills I stradle too many worlds on the margins. However, to my parents' credit - I did learn how to think and also to code, in that order. Has coding helped me get ahead in life? No. Has the ability to think, no - but it has kept my belly full.

DeCoding Fears

What is coding? Coding - is a nonsensical word that means a lot of "nothing" to anyone who has ever trained in data-science or machine learning or any form of programming. In short it is not a skill in the sense that a profession equips anyone with a skill. By learning to "code" one learns to use a tool like learning to use a hammer. Such knowledge neither guarentees the mastery of a trade (you don't become a craftsperson because you can wield a hammer) nor does it give you enough insight to shape your career or usefully contribute to society.

"Coding" is the ability to instruct a computer to do something in a language it understands. That is it.

Now there is no harm in learning a language or a tool, afterall every trade has its tool and tools are meant to be used. The problem is that the current obsession with coding is built on a series of tall lies. These lies are designed to exploit the insecurities and anxieties that every parent carries inside themselves, about their childrens' future. Like streaming services, "coding-classes" for kids make parents feel like they don't have a choice. It is fundamentally disempowering, because, you feel like you don't have the option to simply opt-out.

I was first introduced to the coding-class frenzy by a former neighbour, who sent me message after message demanding that I sign up for a free demo class for my son.

"They'll learn together", "It is the future", "The class is free"..., then there was pressure from my collegues at work, "You know, my son, made this awesome app.... they even gave him a business card. He's just eight, I am sooo proud". Eventually my son's school also caught up. A parent posted "PROUD PARENT" posts on our shared e-classroom, featuring a bright and colourful certificate that declared that 'young Ms xyz, at only seven years of age, was a CERTIFIED coder'.

I've spent many a year holed up in front of my computer staring into the black abyss of Stata and the white abyss of Python (curly brackets and all). At this point, my mind, started to wonder what part of the joy-of-coding did I miss out on? I signed up for a demo class. I had also recently given up my job and so also signed up to be a teacher.

The recruitment ads were catchily done. The company was marketed as a "woman centric" space to be in, with the ability to earn up to 2 lakh rupees a month, flexibility and of course the ability to teach coding without knowing how to code.

The demo class for my child was a disaster. My son loved it, all he had to do was to press buttons and an app magically put itself together. The technical genius behind the whole "coding platform" was beautiful. The teaching was well below average, the teacher was a young (and obviously nervous) thing with clearly no knowlegde of code. There was no space for any questions to be asked - neither by parent nor by child. The instructor jumped from one hyperlinked project to another, spoke extraordinarily fast and clearly had no interest in engaging with the child.

For the next week I was bombarded by sales calls every single day right up to 9 p.m. On these calls, were fancy English speaking individuals, trained at selling dreams. I was told I could chose to buy a basic package of some eight or ten classes for my child which would certify him to be a game or an app developer. I was told that the course was endorsed and created by the very best of the tech-world starting from Sundar Pichai to Gates himself. Or, if I really wanted to make a success out of my child, I could go for the 48 classes programme -

The company would take my child to Silicon Valley and ISRO and make a veritable space scientist out of him, before he was 10! All I had to do was to either pony up half a lakh or a lakh and a half. Oh there were discounts, a whopping 10%.

I asked a few basic questions which remained unanswered - why weren't any questions allowed? Was there a curriculum one could see? What were the sources of the certifications being distributed and how credible were they?

I then "interviewed" to be a teacher. This was also a rather illuminating process. In order to qualify to teach all I had to do was to speak to a marketing person and tell them I wanted to be an "educator". They would then rate my in- front-camera performance (not my teaching ability, knowledge or ability to deal with children) which would, in principle, decide if I could or not be an educator.

Stage two of the interview process, involved "learning" (read memorising) a detailed instruction manual with scripted introductions, examples, doubt-resolution comments and so on plus an in-depth demo of the learning platform. Stage three (this is when I decided I had seen enough), the final round, would have tested my ability to regurgitate scripts and operate the platform.

Reviews of the company suggested that it offered flexibility to women, yet only night shifts were on offer. The women-centric work culture involved being allowed time-off when unwell, and the ability to reschedule classes if leave was granted. Impressive that the company had recognised that women are human beings too!

The most interesting part of the whole parent + teacher experience though was the platform itself. Like all players in the ed-tech space, this particular company, claimed to teach "coding" using its own proprietary platform.

Unfortunately, if you are sharp enough or care about not being taken for a ride, the demo class will show you that the platform is a bunch of front-end UI elements containing elaborate hyperlinks to actual code-snippets hosted elsewhere. That elsewhere is code.org (for this particular company), the menu of "activities" or apps (as sold to parents) that a child learns to code is really a pre-done series of programmes (already solved) that needs little other than a button to execute.

Now what is code.org? Take a look at their about page here, reproduced word for word below;

Code.org® is a nonprofit dedicated to expanding access to computer science in schools and increasing participation by young women and students from other underrepresented groups. Our vision is that every student in every school has the opportunity to learn computer science as part of their core K-12 education. The leading provider of K-12 computer science curriculum in the largest school districts in the United States, Code.org also created the annual Hour of Code campaign, which has engaged more than 15% of all students in the world. Code.org is supported by generous donors including Microsoft, Facebook, Amazon, the Infosys Foundation, Google and many more.

Here is what you need to know in a nutshell;

→ Code.org is free, especially for students. Anybody can use it and anybody can use it to teach others too.

→ It is a repository of thousands of games and app-like activities.

→ Code.org was set up to encourage kids from marginalised backgrounds to take interest in computing, not to teach them to actually "code".

→ Code.org has won awards and is endorsed by the greatest giants of the tech world.

So now here's the rub. Said tech company actually does not need to teach any kid to code. All it needs is skilled operators to deliver pre-determined scripts to parents, because, none of its buyers can tell the difference between code-like activities and code itself. It actively plagiarises and steals content from an award-winning platform and delivers this content as its own. It also claims recognition for its platform and company, derived from endorsement received for a completely different platform. Smashing isn't it?

How do you feel about this now? How does it feel to know that you have not only been cheated by an edu-tech firm that charges you, the parent, lakhs of rupees for something that your kids could learn for free - but also that it actively subverts material meant for the less privilgeed into a paid "product" for India's elite?

What Is In It For The Kid?

Apart from the terribly invasive marketing laced with lies, the truth is that this method of teaching "coding", regardless of where it comes from - is not coding at all. Twitter has been awash with sarcastic comments from parents horrified at how terrible this kind of marketing is;

At best these platforms provide children with what it means to experience the joy of coding, crafting or creating something, at express-speed completely packaged. To sell this, at a premium, along with gratuitous dreams of a gaurenteed future is not just stupid or unethical, it is downright evil.

To be clear here is why;

  1. The "languages" that one can code in today are NOT going to be same, not even in a remote way, in the future. When I learnt to code we began with actual programming languages such as BASIC or LOGO and graduated to C and C++, none of these are in use today. The basics of logic though, those endure.

  2. The future may not have apps. Nearly all future-tech reports talk about service-design or the Internet of Things (IoT) as the future not a gazillion apps. Unification is the future. By investing in teaching your kids to code "apps" - you're already sealing their fate, glued well and truly, to an era bygone.

  3. Coding does not increase scores in mathematics or the applied sciences. It is actually quite the other way around. If you understand mathematics, logic or the basic ideas of the physical sciences you are likely to be able to understand syntax of any kind. Similarly if you are good at maths, you are likely to be good at music or systematic art such as the south Indian tradition of Kolam. The common denominator here is not "code", it is an appreciation of logic and patterns.

  4. Learning-to-code classes are a waste of time and money. They increase screen-time for young kids, which has a documented poor effect, on learning levels in children of all ages. Early data from a landmark National Institutes of Health (NIH) study that began in 2018 indicates that children who spent more than two hours a day on screen-time activities scored lower on language and thinking tests, and some children with more than seven hours a day of screen time experienced thinning of the brain. Our kids, on average, already have thin brains as far as the evidence goes.

What then is the appeal of these coding classes for parents and children? They're different for sure.

For parents it is a clear case of FOMO - as parents we are terrified that the future will only need automatons who can spew syntax at lightspeed. This fear is not entirely unfounded. A large part of the problem has to be do with the path education and technology has gone down, especially in recent times.

Can You Future Proof Your Kids?

When there is an excessive focus on STEM or technological education, there is automatically less focus on the knowledge and mastery of theory and the idea of a social construct. Unfortunately what this leads to the ability to do, without the attendant ability to think through why something needs doing, critically.

An example of this is India's New Education Policy. The NEP has been criticised for several reasons, chiefly on account of its stance on laguage - but not enough, if at all, on its absolutely deficient idea of technology and evaluation. The NEP explicitly states for example, that it intends to use technology, especially AI, for the purposes of evaluation. The trouble with this idea, is of course, that it would be a disaster of epic proportions.

AI, as a science, has hardly developed to the point of a general-sort- of intelligence, that people seem to think we have arrived upon. The fact of the matter is that we have achieved only a very narrow form of AI thus far - and we are very very very far away from any sort of generic super intelligence. While machine learning algorithms continue to evolve, one has to remember what they are - algorithms.

The NEP, and all education, would do well to focus on encouraging a strengthened foundational learning of things like basic literacy and numeracy . Foundational learning is far more important tham learning how-to-code when in primary school.

Young people whom I teach data science to - often wonder why I begin by asking a) what is the shape of the data, can you describe the data-set? and b) what is the important part of a ML model? They learn soon enough. These two questions, in my humble view, illustrate all that is wrong with India's general understanding of tech and the unhealthy obsession with buying into any startup that uses the terms "Coding", "AI" or "Big Data" on its sales brochure.

Too much tech training without any knowledge of the WHYs and HOWs of a statistical procedure, like the current obsession with coding for kids, teaches students how to run large data-sets through a variety of analysis methods or algorithms, in the hope that one or the other will yield a pattern that makes moderate sense. In the social sciences we have a term for this kind of stupidity - it is called P-Hacking, this article on PLOS explains it thus;

There are two widely recognized types of researcher-driven publication bias: selection (also known as the “file drawer effect”, where studies with nonsignificant results have lower publication rates [7]) and inflation [12]. Inflation bias, also known as “p-hacking” or “selective reporting,” is the misreporting of true effect sizes in published studies (Box 1). It occurs when researchers try out several statistical analyses and/or data eligibility specifications and then selectively report those that produce significant results [1215]. Common practices that lead to p-hacking include: conducting analyses midway through experiments to decide whether to continue collecting data [15,16]; recording many response variables and deciding which to report postanalysis [16,17], deciding whether to include or drop outliers postanalyses [16], excluding, combining, or splitting treatment groups postanalysis [2], including or excluding covariates postanalysis [14], and stopping data exploration if an analysis yields a significant p-value [18,19].

In other sciences, the terms data-dredging or data fishing, data snooping and data butchery reflect the same idea, taking a sledgehammer and beating the patterns out of any data-set.

While there is no doubt that ML approaches and the Bayesian idea of more data leading to accuracy is fantastic, to compensate for lack of a theory or a research questions by force-fitting every model to every data set is reflective of a lack of imagination and the ability to think. Now imagine, if this kind of "AI", got to decide your child's grades? It might decide, for instance, that your child will never do well at art because she didn't paint the sky blue but shades of purple.

How would you programme AI for imagination?

Then there is of course the woeful knowledge of how a ML model actually works. What after the algorithm and the specifications? A big part of any ML success is to actually test and train the model for better predictions and results. How does one train a model? Using real-world data.

Perhaps you have been involved with some training in your life? Maybe a corporate training? What was it like? Was it flawless? Perfect? That seems unlikely. And so it goes with data-models. Real data is messy, models fail.

What Our Kids Really Need

What our kids really need is the ability to think and the ability to learn. More importantly they need to move away from ideas of instant gratification, funky apps and fancy certifications. We'd asked earlier on, in this essay, what is in it for the kids? This question is easy to answer. Kids love games, they love being given credit for work done and they love for it all to be easy. Tech platforms, such as the ones we have been discussing, are built to satisfy these desires. Every single of them.

After all creating something is fun and exciting. Play is a geat way to learn for this reason and truly, the best way to learn to code is play. Blocks, puzzles, sorting games, art, music even sport all help contribute to lateral and logical thinking. None needs extended screen-time.

At the heart of computational thinking, everyone's new favourite phrase, is problem-solving using a simple idea that is rather ancient. It is called a framework, for those of us in the humanities, this is the same as a theory. Computer scientists call this "abstraction". In systems thinking smaller problems are solved, by solving for the framework in which these smaller problems manifest. For example - STEM education used to have this problem before it got called STEM, it was algebra and C++ and it was boring. It was tedious. Then marketing happened. Suddenly mathematics was a STEM skill, in-demand, the path to a future cemented in the NEW. Nobody needed to focus on the tediousness of it anymore. It was GLITZY, SHINY and STEM.

The truth of the T of STEM though is that technology is tedious. At least real programming definitely is. This is where the satisfaction of being able to code comes from - anything that takes effort is satistifying once you succeed. The trouble with a platform created specifically for children - to teach code, is that it is already a coded product. Tailored to children, for children. At best these are code-simulators, they try to emulate the coding experience, but they don't actually teach children to code anything. There is no failure and success is too easy.

Following his demo class, my son wanted to create a spaceship game. I refused to shell out the 1 lakh, but I did introduce him to scratch and logo. To code his game, he had to know how to move an asteroid to the right - using a pre-built command called moveright. He didn't actually code from scratch. Somebody else did and it was beautiful code, it allowed him to understand what coding could potentially look like, at a higher level of abstration and beauty than making it, say in Python, would be.

Its a great intro - but if we start to believe that this is what real programming is like, we're setting up our kids for failure. There isn't an iota of real coding knowledge gained from this. One reason this is so - is because the whole process of programming simulation ignores the process of real programming, which is imagine, build, test, fail, document and investigate the causes of failure, iterate, repeat.

In that sense, if our kids were really learning to code, they'd be learning about life. But life isn't a hat you want your kid to don at five or seven, try letting them run in the sun, love a pet and play chess instead.

Cat lover. Peaceful bibliophile unless confronted by a fascist.