Ideas to improve online education



  • Have students build stuff – CS students building real software, English majors write and self publish a book on Amazon, business majors start an online business.
  • Partner with researchers in related fields – Involve researchers from areas such as cognitive learning, teacher education to seek and apply latest findings from research labs into the real world.
  • Constantly experiment with new courses – Every field is advancing and converging with other fields at an ever increasing pace. Innovation in cross-disciplinary fields is the life blood of new industries. Teaching students how to learn and apply knowledge across domains should be prioritized to advance the overall progress of education.
  • Co-ed with industry – Pick the Googles of each industry and work with them to tie up students with a mentor who can guide them with career advice and coaching.
  • Create course content that is up to date – Most traditional education teaches outdated concepts and techniques that rarely apply to industry or contribute to employability skills.
  • A/B testing framework for course creators – Almost all consumer software that we use today is continually improved using A/B tests. There are plenty of sophisticated frameworks available for all major software platforms to easily incorporate such experimentation in the content creation process. Online education software should have such frameworks and tooling available for course creators to constantly experiment and figure out what works and what does not in teaching specific concepts.
  • Accreditation widely accepted – Not just to give students peace of mind, but to facilitate efficient recruitment and better mobilization of talent.
  • Gamification – Kids (adults also in fact) never have to be forced to play video games. Making education as engaging as a popular video game might be too high a bar, but there’s definitely huge opportunities to gamify and make education a lot more immersive and fun.
  • Focus on the developing world – High cost of education and opportunity cost is often the prohibitive factor stopping kids from attending schools and colleges in most parts of the developing world. Online education can be an order of magnitude cheaper than traditional schools and thus have a compounding effect in those countries to get more kids to learn.
  • Student collaboration across disciplines and geographies – One of the most important and underrated aspects of traditional education is lateral learning that happens outside the classroom. Online education should compensate for that by developing a collaborative model that encourages and rewards peer-to-peer learning.


Three types of founders

Successful founders –

  • solve a hard problem and
  • build something that adds real, measurable value to their customer

Unsuccessful founders who are smart –

  • solve a hard problem but
  • don’t know who their customer is or how to add value to them

Unsuccessful founders who are not smart –

  • look for easy problems (low hanging fruits) and
  • think everyone is their customer, thus end up building an irrelevant product

Ten ideas for Uber to experiment with

  • Affiliate program for third-party app developers to integrate Uber and allow users to book directly from their app. E.g. Yelp would have a “Get Uber” button to take the user to a restaurant they are looking at.
  • Flat fee service e.g. unlimited trips within a city (San Francisco) on Friday evening for $50.
  • Social feature that allows users to buy/pay for someone else’s ride e.g. when inviting friends to your party.
  • In-app video ads, that subsidize cost of a ride for the duration of the ride that the rider watches ads.
  • Pre-booking service for longer rides, taking the risk and uncertainty out of odd hours travel.
  • An exclusive women driver service, specially in cities that are not considered very safe for women, like Delhi.
  • Longer term cab rental service for e.g. at vacation destination where tourists usually book multi-day chauffeured cabs.
  • High speed wi-fi hotspots, could be super useful for riders to stay productive on longer commutes e.g. on pooled services for ~ hour long commutes to and from work.
  • Discounted rates on Uber for drivers when they travel.
  • Exclusive partnerships with regional gas companies to get discounted rate for drivers.


What I would love to see in messaging apps next

Personalized emoticons based on the profile picture of users.

Users upload their profile pics for a few different expressions – happy, excited, sad, disappointed, angry etc.

Or a much better UX would be that the app uses image recognition and sophisticated image manipulation algorithms to produce a variety of expressions from a base image.

And now each time a user adds an emoticon along with their message, the relevant expression from that emoticon gets picked up and is applied to their profile photo.

For example, a message like this –

Can’t wait to see you  :-)

Turns into –

Can’t wait to see you   giphy

Now wouldn’t that be awesome!



How good design could hurt your app

Think about this for a second – if your prototype or MVP is very well designed, it might in fact be trickier to get objective feedback from potential users. They are bound to get distracted by the beautiful design, and subconsciously base most of their feedback on how it looks/feels rather than to tell you if the core of your product is worthy and solves a real, genuine problem for them.

Design is important, but the timing to focus on it is more so.


How self-driving cars will change our world

  • Car ownership –
    • Economies of scale will make using an on-demand, self driving car an order of magnitude cheaper than today’s cabs/rental cars.
    • A side effect would be that owning a car will become much more expensive, since the cost of car production will be spread across fewer buyers.
  • Use of land –
    • The amount of space occupied by parking garages inside buildings, houses and on the streets will decrease and free up space causing changes in real estate prices, specially in big, dense cities.
    • Self-driving cars would also mean that road usage will be much more efficient than human drivers can achieve thus minimizing the need to keep expanding roads and highways.
  • Travel time – 
    • Travel time on highways and freeways could be improved dramatically by introducing dedicated self-driving car lanes allowing for much higher speed limits.
    • A real-time centralized road monitoring system in each city will be constantly recording car density on different roads and travel times to different routes. Such a system will be much smarter and powerful in analyzing and distributing traffic and constantly optimizing travel time across the city.
  • Safer roads – 
    • Self-driving cars would eliminate all the top causes of road accidents – distracted driving, drunk driving and speeding.
  • New types of users – 
    • Traveling for people who can’t themselves drive such as young, old and disabled will become much easier.
  • Car manufacturing and design –
    • Research and development in car industry will be much more focused on the software components (software’s eating the world!)
    • Car design process will change to prioritize function over form, since most cars will be owned by companies that operate fleets of cars rather than individuals concerned about style, color and other non-functional factors.
  • Car distribution – 
    • Most car buyers (~90%) would be institutional buyers such as rental car companies, cab companies and new types of franchises that emerge in the self-driving car world. Such drastic changes in the buyer profile would also hugely impact the car distribution and dealership market.
  • New competitors in the market – 
    • Google is already the biggest player in the market and unarguably the only player at such an advanced level. Given the complex research and interplay of hardware and software technologies that is needed, it’s hard to imagine any traditional auto manufacturers to reach this level without partnering or licensing from Google.
  • Effect on other modes of transportation – 
    • Other forms of transport such as airlines would lose business. Specially for short distances (~300 miles) travelers would find it cheaper, time efficient, comfortable and flexible to book a self-driving car rather than book a flight months in advance, travel to the airport ahead of time, wait at the airport, deal with flight delays, and travel in a cramped economy class seat.
  • Effect on adjacent industries – 
    • Other industries such as car insurance will be see drastic changes. Lesser people driving and owning cars would mean fewer accidents and fewer cars to insure.
    • Other services such as warranty companies, roadside assistance and mechanics would also see dramatic change in their business model and margins. Mostly because institutional car owners would benefit by having their own in-house maintenance and support teams, by spreading the fixed costs of such an operation over a much larger total pool of cars they would own.

10 types of people (based on skills) I would hire

  • Super connector – Has a large, ever growing network of different types of people. People to hire, to freelance, who would invest, who can be partners, who host events, who write popular blogs, who work in media, who are super connectors themselves.
  • Growth hacker – Loves to reverse engineer “virality”, who knows the growth curve of most successful startups, who is creative and unconventional, who has spent considerable time on forums where growth hackers hang out (, who regularly reads Andrew Chen, Sean Ellis, who is a mix of tech and marketing.
  • Chief of tech – Has years of experience learning and mastering their art (spent 10000 hours in the field). Loves to build and nurture top class engineering teams, is obsessed with code quality, world class architecture and building things for scale.
  • Black belt engineer –  Excellent at written and verbal communication and at understanding complex and vague requirements, loves software engineering, contributes to OSS, gives back on Stack Overflow, has their own pick of favorite programming languages and tools.
  • Master of all things front-end – Has experience building awesome UIs on different platforms, be it a WordPress site, a mobile app front end, a landing page, an email newsletter.
  • Design guru – Lives and breathes design, not just when they are working, but with everything they see and experience. Someone can talk about how the design of a door knob impacts usability, who has a awe inspiring portfolio on dribble, who knows about best practices on each platform but at the same time is not afraid to push the limits when needed. Someone who knows how to talk to users through design.
  • Business dev champion – Knows how to sell by not selling. Has coverage across different technology related businesses and can connect the dots to figure out hidden partnership opportunities. Can instantly come up with 10 ideas how Snapchat can make revenue, suggest 10 new verticals that Amazon should diversify into, 10 untapped tech markets that Google should explore.
  • Product maestro – Is a master at figuring out and doing everything it takes to build super engaging products that users love. Data juggler, customer advocate, tech background, design empathy, strategic mind, team player. Loves community building and is regular on Product Hunthacker news and other product communities.
  • Analytics ninja – A data geek. Someone who is on top of the industry standards when it comes to analytics and data. Who regularly reads awesome blogs such as Occam’s Razor, Hiten Shah and MixPanel.
  • Open role – Someone who is skilled at what they do, has a track record of having done amazing things, has bright ideas, and great conviction skills.


Why Facebook doesn’t need to monetize Whatsapp, ever!

After more than an year of the Whatsapp acquisition, there are no signs of Facebook monetizing the service. No ads, no subscription, no nothing.

So how will it ever generate revenue and justify the hefty price tag?

Well, I don’t think they need to generate revenue through Whatsapp itself at all.

The number of mobile users and usage has grown tremendously over the past few years and continue to do so. That is common knowledge. But in spite of all the growth, its still a zero sum game. The total amount of time spent on mobile/apps (total users x average usage) is a large number but limited nonetheless. Every time a new product or app hits success and millions of people start using it, it eats into the time people were spending on another existing product.

So, total mind share is a constant.

Put slightly differently, the amount of digital ad inventory is limited. No matter if its through display ads, banners, in-app, timelines, video, audio. The most engaging product with creative ad formats and relevant ad-targeting, will command advertiser’s dollars.

To provide an analogy, imagine there is a newspaper and its the only place where advertisers around the world can show ads.

  • The paper has 10 slots for ads and an ad reseller comes in and buys 7 of those 10 slots.
  • It then decides to NOT put ads but something else, something that readers love – puzzles, cartoons etc. on 4 of those slots.
  • Now there are only 3 slots remaining for ads and price for those 3 ad slots shoots up drastically, due to limited supply and lots of advertisers wanting to put their ads.

That is what Facebook’s revenue strategy is. Those 4 slots are Whatsapp’s mind share and the remaining 3 slots are Facebook itself 😉



The recipe for a successful tech product



How most people think successful products are built –

  • Think hard and come up with revolutionary idea
  • Hire well rounded team to execute
  • Build complete, well-designed product
  • Get a ton of users to discover your product through press coverage and “social media”
  • Raise money and hire more talent
  • Media, money & large team will auto-magically mean ever growing users & product usage
  • Start charging for product or put ads to generate revenue

How successful products are actually built –

  • Identify real burning problem of target user group
  • Connect the dots and formulate your hypothesis of what a good solution to the problem would be
  • Validate hypothesis & build on it through observation and feedback from users who have skin in the game (pre-selling)
  • Build the core of a solution that delivers enough value for which users are happy to pay
  • Iterate on v1 to achieve strong product-market fit
  • Continue developing deep understanding of the domain, user problems, and keep experimenting towards building a world-class, successful product

How to write better

  • Know your audience, and use vocabulary relevant to them.
  • Develop a tone, similar to how you would speak in person.
  • Organization is design – structure your sentences, use paragraphs/lists as appropriate.
  • Provide examples, real life instances. Its easier to understand with association.
  • Write less, write crisp.
  • Reread and edit after a break.
  • Summarize.