The State of Affairs

Not much is happening in the Indian tech and startup ecosystem these days when compared to the boom between 2020-2022. But recently, something in the news cycle caught my attention. The central government launched an app called “Bharat Taxi”.

Uber and Ola are classic marketplace platforms which absorbed massive losses for almost a decade before they could create a system good enough to inspire a debate on economics of buying a car vs depending on autos, taxis and bikes.

Rapido entered the market and destroyed the duopoly by introducing the “bike taxi”. Subscription driven SaaS play by Rapido and Namma Yatri further killed 30% cuts from the Uber-Ola duopoly.

Platforms like InDrive introduced the negotiation and price discovery model. It’s more of a drag in my opinion, given the lack of “revolution” that was expected to bring about.

The landscape of ride hailing platforms before entry of Bharat Taxi looked like this (based on data between 2023 and 2025).

Us vs Them

But the platforms have optimised for users rather than drivers, and later further moved in the direction of optimising for profits over both riders and drivers.

What they didn’t see coming was an alternative model i.e. a cooperative platform taking on their VC pipe dreams on monopoly/duopoly destroying their entire thesis.

The need for change has always been there. During every bike ride, I have heard drivers complaining about delayed payouts.

“Companies are running taxis for their own benefit, whereas Bharat Taxi is for the benefit of Sarathis (drivers),” Shah said.

Source

It’s important to note that Uber, Rapido and Ola succeeded because they were able to advocate for better user experience, reliability and price discovery for the passenger.

This might slightly titillate the hopes of many well-meaning users and citizens who are concerned about income inequality and anti-labour behaviour of big tech platforms.

The co-ownership model is not fiction but a reality embedded in the structure of Bharat Taxi.

Shah said that under the cooperative structure, if Bharat Taxi earns Rs 25 crore after three years, 20% of the amount would be retained as cooperative capital while 80% would be distributed among drivers based on the number of kilometres driven.
Source : ET

This is definitely an enticing deal for the driver community.

Competing Against Incumbents

But one must note that Uber, Rapido and Ola are mature platforms that have established user trust, learnt from data and built systems to address many problems linked to mobility space.

For Bharat Taxi to succeed it needs to perform as good as its competitors on passenger satisfaction while simultaneously improving driver incomes and experience.

One needs to benchmark or evaluate Bharat Taxi experience for users on two important axes to make any reasonable guesses about its potential for success.

• Price

• Safety

Price

On the price front, Bharat Taxi seemed to be more expensive than Uber-like alternatives. But we are not sure how the overall spend of a consumer would look like. Incumbents are notoriously good at offering a cheaper first ride and then charging more for the return ride.

The price models that Uber has built with almost a decade of consumer data would be miles ahead of Bharat Taxi. In addition subscription products like Uber One and better brand recall for older apps could trump Bharat Taxi.

The “no surge” promise on part of Bharat Taxi could prove to be useful. Generally a user would have multiple apps on their phone. I have seen a friend comparing prices on InDrive, Uber, Ola and Rapido before making a choice. We have enough time to research which service would cost the least and pick exactly that.

If capital were to be a deciding factor in the fight for market share, incumbents would have deep pockets and the Bharat Taxi would have to rely on the “ownership” promise to clog the supply side in this fight.

But cost for passengers, I feel is the “driving” factor in adoption and sustenance of an urban transport marketplace. The volume of bike rides in comparison to auto and cabs is a definite indicator of price sensitivity of the Indian consumer.

Earnings and ownership dreams of drivers would be punched down by the passenger preferences and ride volumes on Bharat Taxi in comparison to other established platforms.

When I took a short ride on Bharat Taxi last week, I asked the bike driver if he got enough rides on the app. He mentioned that it was kinda low given that it’s new player in the market.

Safety

Will the “owner” mentality of the driving community have any downside on the safety aspect? Hopefully not.

Current players have had to bear the brunt of multiple instances of negative media due to safety concerns. But the brand itself has not taken a devastating hit on this front.

Services like Uber, Ola and Rapido have almost become an essential commodity for many urban commuters. The negative perception on the safety front often tilts towards the drivers and not inadequate systems built by the ride hailing platform.

The same scenario applies to Bharat Taxi as well. But consumer trust is likely to fall only if the user feels that the new option is far worse than incumbents on the safety front.

One must note that Bharat Taxi has more or less the same safety features as current players on its app. I am sure they would have studied the product and executed something similar.

The limiting factor in passenger safety is not really the app or its features but the driver’s themselves.

Predictions

I am not a qualified industry expert. But from a user’s perspective I don’t see any significant differentiating factor between Bharat Taxi and other players.

Also, my predictions are based on the assumption that Bharat Taxi or unions don’t stop or discourage drivers from multiple platforms in future.

Bharat Taxi has space to manoeuvre if it can drastically shake up the supply side for its competitors, i.e. driver and vehicle availability. If it does so by seeking more money from passengers, it is bound to fail.

Given the delicate balance required for growth, I don’t see Bharat Taxi capturing more than 10-20% in the near term (say 2-3 years). The promise of equity and profit share is a long term game which majority of drivers may not be ready to play.

The drivers, especially bike taxi drivers would be happier if they get more rides. The platform doesn’t matter much for them as long as they are able to earn decent money.

After seven months I think I can meaningfully comment on the product proposition with a bit more clarity.

The Moment of Hiring

I would like to go back to the “moment of hiring”, as explained by Clayton Christensen in his book Competing with Luck. It seemed like a great idea to get access to some cool features like maps, recovery tracking, cool icons and “AI powered insights”.

To be honest I never used any of these features over the last few months. But I still feel the Strava subscription is worth it because of other normal features which kept me on track.

What Worked

What really worked was quite simple and had nothing to do with AI. Strava covers all gaps that are missed by Apple Watch and makes the user experience great with good design. It scores high on these parameters.

• Community (a platform that unites both Android and iOS users)

Every time you get a kudos for a run or workout from a stranger, you feel good however tiny it might be. It’s as good as the dopamine hit from an Instagram like.

People are now saying that Strava is the Instagram for 30+ people. Haha.

• Badges (especially the monthly badges, activity milestones)

Clicking on a monthly 100km target and earning that badge, (in spite of being a virtual badge) feels like an achievement. It truly helps you feel that you’re kinda moving towards some goal.

Of course, even Apple Watch offers the same, but the badges are hidden somewhere deep in the UI.

Strava brings these motivating stats/goals a level up. Branding it as a “trophy case” makes it more accessible and occupies a larger space in the minds of the users.

• Metrics and goal setting (Training Zones and Weekly goals)

The product encourages you to setup weekly activity goals and it’s right on top of the home screen. This reminds you how many activities you’ve finished and gives you an opportunity to catchup if you’re lazy.

I am surprised why Apple’s Activity UI doesn’t have this easy and basic feature. It was a clear miss on Apple’s end.

• Fitness metrics, analytics and predictions (it accurately predicted my Half Marathon time)

I am unsure how the fitness score is computed, but it is a black hole metric that keeps you going. I hit the highest score of 76 few weeks back and now I’ve dropped to 66. All thanks to the brief hiatus due to the smog in Delhi. But now I’m kinda pumped to get back to 76.

The other cool feature that caught my attention was the race predictions. There must be some simple formula to calculate all of this and recompute prediction after each workout. But, I was truly impressed when my Half Marathon time was predicted perfectly.

In addition to this, the visibility given for heart rate / zone training metrics is super useful. Even Apple Watch does give zone related data for running workouts but it is hidden deep inside and aggregate zone training data is not available on Apple’s Activity metrics.

• A yearly wrap (If Spotify does it for music, then why not?)

This story is not over. I have been checking if the “Year in Sport” stats update as I log more. It does. This is a well thought out feature as the year doesn’t truly end until it does.

This my Strava data from May 2025 to 22nd December 2025 (I will update this).

Verdict

All in all, Strava clearly addresses many gaps in health devices like Apple Watch and brings personal accountability to the centre-stage. Community likes and support are an icing on the cake.

The sexy features, though not-so-useful make sense in the context of the core fitness focused features on Strava. It’s a worthy investment all in all.

The answer to the title of this post — Strava exists to ensure you set fitness goals, follow them and feel successful via the community approval along with meaningful UI and stats that track your progress.

Before this instance, I had never keenly explored Google Maps on the phone app as I generally navigate while driving.

Now I’ll just go through the way in which I discovered this particular feature for the first time.

Narrow Roads

Firstly, the icon indicating narrow roads confirmed my concerns about lack of parking space. This however made me want to pinch the screen and zoom in to check if there are parks around which cars can be “parked”. Haha.

One can’t intuitively understand the the user can pull up the bottom sheet or the drawer to find more information. Though I am not sure how the user can be educated about this, I feel it was probably the first time I pulled up the bottom sheet.

Unclear Images

Now, accidentally or intentionally, I drag up the bottom sheet only to find that there are images along with instructions. My first reaction was “Haa! Cool! They have added images!”. But in reality these images didn’t strike me as similar even though I’ve driven on major roads leading to South Ex.

Even for a driver with 27,000km mileage, the arrow or directions didn’t make much sense. What does the circular pointed arrow mean? How can a normal person see it?

The Helpful Image

The only image that did make sense to me was the one that showed extent of narrowness of roads near the venue. And I didn’t even know that one can click on the image and see the full view until I started writing this post.

Even in this image the arrow kinda wants me to go and crash into a tree trunk. Haha. That said, I just realised I had to leave ten minutes early to find a parking spot and walk the last 200-300 m distance.

Also one can note that the image has no correlation to the map section being shown to the user. All of this made me wonder why Google took so much pains to get these images and add it to the navigation part anyway?

Conclusion

The idea of navigating maps with reference images is a great one. But the Google Maps feature fails on three major fronts.

• User education: I feel that the user doesn’t even know that this feature exists. How will they use it when they don’t know it exists?

• Lack of clarity: The images are stretched and have little resemblance to the real world. The circular pointed arrows on the image make no sense unless I am living under a rock

• Practical Utility: Can a person driving stop in the middle of the road and see these images while driving? Or, are they expected to memorise this like an F1 driver before starting the engine?

Even with a passenger sitting next to you, it would be hard to guide the other person with such images. Forget about two wheelers benefitting from this.

I am unsure how these images can be of substantial value in spite of the practical difficulties. The idea is great but the execution is kinda subpar. Some innovation on the design front plus user research could help Google Maps nail this.

That said, there aren’t many players in the maps space, so the push for such perfection is kinda absent at this point.

Anyway, this is just me venting about difficulties of navigating to a location with a near-monopoly product that has become an integral part of my life.

Goodreads is a product that I’ve been using since teenage, even before I learnt the ABCs of design or product management. However, the constant feeling that the app and website were slow and ancient irked me. But guess what, it has a de facto monopoly in the “social media for books” market.

Of course some apps did popup and I gave it a try. But the network effects and my activity history on the product have kept me hooked in spite of the underwhelming product experience.

In addition to all of this, finishing the annual Goodreads book challenge has now become a part of my personality. In spite of being an ancient and not-so-exciting product, Goodreads has held my attention for more than 15 years now.

The Announcement

After seeing the new logo, I scanned the app and found very few changes. The look and feel of the entire product still remains the same. I wondered if I had missed something and went to their official website and LinkedIn to find this announcement.

New logo, Reading Challenges, and more! As part of our commitment to updating Goodreads, we're excited to share several recent updates:

📱 Our new logo is designed to better represent Goodreads and is optimized for accessibility so it looks clear and sharp no matter where our readers encounter it—from your phone to a billboard.

🏆 As part of our popular Reading Challenge, we've introduced new rotating limited-time achievements throughout the year. We’re delighted to see so many readers already discovering books through this feature!

📚 Your Want to Read list is now included in Your Books on Amazon—one of our most requested features.

🎧 Our book catalog now includes more than a million audiobooks, making it easier to track all your reading.

My disappointment found further evidence after reading this post.

Pathetic Design

The logo they’ve chosen is absolutely out of sync with product experience inside the app. It feels alien and ugly, but that’s just my opinion. I can’t put a finger on it or substantiate it, but it just feels out of sync with the product.

Even the new feature they have announced is out of sync with user behaviour. I don’t pick books to complete the kind of reading challenges that were suggested by Goodreads. The only one that matters is the annual challenge.

I speak for myself when I say that I don’t pick a mystery book to complete a “summer challenge” when I have other compelling unread books on my list. Many would agree with this take. At least they could have personalised these challenges based on past reading behaviour and current books in progress.

The “Summer Challenge” feature is not really aiding the progress or happiness of a reader. I don’t want to win that challenge (or even take a look at it).

In addition to this , I found that the design, presentation and communication of the new “popular Reading Challenge” was pathetic if not terrible on the web product.

• Firstly, these challenges communicate nothing substantial to the readers. One has to click the icon and then find the details of the challenge.

• Even on the app the challenge is so poorly presented.

• Even the icons are lazily designed with just a “?” on a bookmark icon (for the mystery challenge). They are all of the same colour (more or less).

It doesn’t evoke any excitement or interest in me. If it did evoke your interest, do let me know.

Goodreads Reigns

In spite of the poor design and product experience, I feel Goodreads will still prevail due to these reasons.

• Goodreads just has a really huge data moat at this point, especially among readers who have used this product. Storygraph, the second biggest competitor has 1.2 million users while Goodreads has 140 million users (as of 2022). Recent data might tell a different story, but I don’t believe anything has changed drastically.

• Since new users for any product that wishes to be “social media for books” would come from the current network, there are very few opportunities for an alternative.

• I also believe that the frequency of logins and updates on such a product is much lesser than that for Instagram or X. Hence the incentive to search and migrate to an alternative would be quite low.

One can only hope that the Goodreads hires a good product design team and revamps the entire experience. The reading community of the world definitely needs a better product.

The term AI or artificial intelligence is the sexiest term out there. Celebrities and news reporters are dishing it out in every conversation. Some think that AI is going to solve humanity’s problems while others think it will destroy the world.

All this talk does contain some element of truth. But in real terms, AI today is a technology in search of viable business applications.

The key term here is ‘viable’.

Does the technology provide ways to improve current methods with sufficient accuracy and contribute to higher profits? The answer is no in a large number of scenarios.

Profits?

This concern about viability was reflected in stock price losses of Big Tech companies which have heavily invested in AI. This list includes Google/Alphabet, Meta, Microsoft and Amazon.

The development of the technology and its proliferation on phones and social media accounts has not yet contributed to larger profits for these companies.

At least not yet.

There are some opportunities in ad-targeting, conversion of customers from ads, smoother office applications and maybe some intelligent insights that could contribute to better supply chains.

But do these opportunities yield profits that are far superior to existing systems?

We haven’t seen it yet.

ChatGPT/OpenAI - A case study

Macro numbers according to a September 2024 report suggest that ChatGPT has 200 million weekly active users. And it also claims it has 100 million paid customers. Now, we don’t know the churn rates and overall revenue, profits etc. But the fact that they are raising funds so often indicates that costs are exceeding earnings.

As an individual user, I have not seen a major difference between a paid and free version for most daily tasks. I had a subscription of four months, but terminated it as I could not see much value beyond recreational tinkering.

Visual and audio capabilities of ChatGPT are kinda fascinating, but it doesn’t have many use cases for an individual user. However, I believe a large number of security and education related products can be built with these capabilities.

But eventually, it will boil down to costs, because compute power required to run LLMs is still higher than a similar solution that could involve a human.

For example, an AI enabled security camera might be a cool thought but running it would be costlier than hiring a security guard. An AI-enable tutor might be a great solution, but it can only address repetitive tasks and personalise some parts of the learning journey. But it is not going to be cheap if you expect it to teach, calibrate and engage a student like a real human teacher.

The cost factor is kinda being addressed as OpenAI is working on cheaper models that can achieve optimal performance. But, I am yet to see the larger shifts in search market, which kinda predicted death of Google.

However, we can see ChatGPT’s UI resembling Google Homepage with every passing day. Initially, people didn’t know what to do with this technology. Now, the homepage educates the user and possible tasks that can be done with this tool .

Unlike Google ChatGPT is supposed to give an accurate and satisfactory answer to the user. This expectation hits a roadblock because of two reasons.

1. All users don’t know how to frame a question or problem statement in a way that ChatGPT can solve effectively.

2. In the absence of several competing links or answers thrown by Google, the user finds it hard to fully trust the answer without citation, hyperlinks and other markers of trust. Is there a better answer than this? Can there be mistakes here? These are looming questions for every user.

A large shift in user behaviour is needed for the economic success of products like ChatGPT. With Apple Intelligence mainstreaming LLM technology on consumer devices, organisations like OpenAI might end up developing and testing solutions on tools like ChatGPT and then become service providers to big tech companies like Apple or Microsoft.

What Do We Do Now?

What’s the appropriate use of AI or machine learning models in the real world?

Well my litmus test for a suitable AI solution involves these factors:

• Is there a lack of a universal law for the solution being provided?

• Can large dumps of data be collated to create a situational solution for the problem?

• Is there potential for improving the solution as more data enters the system built on current data?

• Can the solution absorb reasonable amount of error?

• Will the solution be safer, reasonably accurate and cheaper than the current solution in market?

Anyone who has built a tech product can understand the context of these questions fluently. For others I will provide two examples.

Customer Care

Customer care has no universal laws. However, you can train a human or a computer to solve this with past data. As new scenarios arrive the customer care person or a computer with similar learning capacity can improve.

For simple apps like Swiggy or Blinkit, there is no safety issue (beyond food quality and electronic device risks that are outside scope of discussion). This solution can be cheaper and reasonably accurate. In case of errors by the bot, you can always connect to a human.

So customer care can be effectively solved with AI and it can contribute to profits by lowering customer care expense.

Health Care

There is a case for AI applications in helping diagnosis of patients but prescription and surgery are dangerous domains. Even a slight error in the process can have massive consequences.

Of course, it is theoretically possible to train a robot to perform a surgery by feeding it tonnes of information and data. But, the costs involved and the risk factor is way too high in this scenario. Not many people would prefer a robotic surgery for a lower cost.

In the best case scenario, the technology can only assist a human surgeon and reduce errors that could creep in during a surgery.

Impending Doom?

As the complexity of tasks increases, the level of data, training and compute power required to run an AI / ML model/robot rises. In such cases it is far cheaper and safer to hire a human to perform such tasks. This is the reality in the present and near future according to a large number of experts (and my own common sense).

However, once we achieve artificial general intelligence (a solution that enables machines to learning anything from their surroundings), at scale and reduce costs of computation drastically, we will not be anywhere close to AI doom predicted by many people.

At least from the product management, data analysis, design and coding POV, I believe that AI will only be an assistant to humans and help them be more productive. However, more mundane parts of these jobs are ripe for automation, freeing up people to do more innovative work.

This thumb rule is likely to be true for all creative jobs that involve human interaction, unpredictability and a high risk profile. Artificial Intelligence is unlikely to take over the entire job role in such a scenario.

ChatBots and assistants provided to millions of people on WhatsApp or Instagram are unlikely to serve any major needs beyond generating that cute meme that you want to send to your crush or friend. But this could be a playground where data can be collected to build or fine tune monetisable products in future.

Predictions of doom are not going to materialise very soon. However, we can see culling of many low-skill, uncreative and repeatable jobs.

Key Takeaways

I tend to listen to interesting podcasts while driving. Of late, I am exploring Lex Fridman’s earlier episodes where he focused purely on technology related topics. I completed an episode related to Gödel machines. (I do feel a sense of excitement when I see two dots on the ‘o’ in Gödel.)

The episode is about a self learning AI machine that can learn concepts and get better at learning over a period of time, just like humans. It’s fascinating to hear about this concept from the person who invented this idea in the first place. Before delving into my own thoughts, I would love to summarise my important takeaways from the episode.

What are Godel Machines?

Gödel Machines are a concept in theoretical computer science and artificial intelligence, introduced by Jürgen Schmidhuber in 2003. They are self-improving, general-purpose problem-solving systems that can theoretically achieve optimal performance.

Named after the logician Kurt Gödel, these machines leverage principles from Gödel’s incompleteness theorems and the idea of formal self-reference to make themselves better over time.

1. Self-improvement: A Gödel Machine has the ability to rewrite its own code, including the rules that govern how it rewrites itself, which allows it to improve its problem-solving abilities.

2. Optimality: The machine aims to achieve optimal performance in problem-solving tasks. It does this by proving, within its own formal system, that a modification to its code will be beneficial.

3. Proof-based rewrites: The machine only changes its own behavior if it can mathematically prove that a new version will lead to better performance. This ensures that it doesn't degrade over time.

4. Recursion and self-reference: It uses a formal system that includes a description of itself. By reasoning about its own structure and performance, it can make theoretically sound decisions about how to improve.

Source: ChatGPT

Having this kind of machine for learning can drastically reduce computation costs and eliminate the need for huge amounts of data to create useful AI models for various applications.

Compressible Knowledge

Before listening to this episode, I felt that science and research was aimed at expanding out knowledge of the universe. And technological innovations if any was a side effect of scientific research.

But Schmidhuber gives an insanely contradictory, or rather expansive view of the process of human scientific endeavours.

The history of science, the history of humanity, our civilization and life on Earth as some kind of path towards greater and greater compression. What does that mean?

Hundreds of years ago, there was an astronomer whose name was Kepler, and he looked at the data points that he got by watching planets move. And then he had all these data points and suddenly it turned out that he can greatly compress the data by predicting it through an ellipse long.

So it turns out that all these data points are more or less on ellipses around the sun. And another guy came along whose name was Newton and before him, Hook, and they said the same thing that is making these planets move like that is what makes the apples fall down.

And it also holds for stones and for all kinds of other objects. And certainly many, many of these compression of these observations became much more compressible because as long as you can predict the next thing, given what you have seen so far, you can compress it, but you don't have to store that data. This is called predictive coding.

And then there was still something wrong with that theory of the universe, and you had deviations from these predictions of the theory, and 300 years later, another guy came along whose name was Einstein, and he he was able to explain away all these deviations from the predictions of the old theory through a new theory, which was called the general theory of relativity, which at first glance looks a little bit more complicated.

And you have to warps space and time, but you can phrase it within one single sentence, which is. No matter how fast you accelerate and how fast or hard you decelerate, and no matter what is the gravity in your local framework, light speed always looks the same.

Paraphrased from the Podcast Transcripts

This explanation gives a new insight into how the human race makes progress. It is not by creating new knowledge or theories but by compressing data into small usable pieces of knowledge.

Can Humans become a kind of Gödel Machine?

Though it is a purely theoretical idea aimed at making a self learning machine, I feel even humans operate like a Gödel Machine. If not all, at least humans who want to grow and be better do operate in a manner similar to Schmidhuber’s machines.

Let us revisit the characteristics of a Gödel Machine.

1. It rewrites its own code.

2. Tries to solve problems in the most optimal manner.

3. It tries to change its ways if it thinks it can do better.

4. Introspects and improves continuously.

Isn’t it so elegant and beautiful? This is the manner in which people ought to function if they want to get better. How can one move forward if the individual keeps on repeating the same mistakes and patterns again and again? Can you expect a better result by repeating same mistakes? Of course, no!

Whether we can build a Gödel machine or not, we can surely try to imitate one, along with all our flaws and emotions working in sync.

Instrumentalism and Realism

The phrase “does god play dice” is associated with the debate on whether the world is deterministic or probabilistic. In the sense, can we discover all laws and predict how things will turn out or whether things are inherently probabilistic and nothing can be fully understood.

The phrase, “god does not play dice” was popularised by Albert Einstein who believed in a deterministic world and rejected probabilistic roots of Quantum physics. This question was thrown at the guest speaker by Fridman and the answer is refreshingly honest.

We don't know some factors. And because we don't know some factors, which could be important, it looks like God plays dice.

Vladimir Vapnik’s answer on Lex Fridman Podcast

It’s really fascinating to know about the process of learning. How does one learn? What are we learning in reality? Let’s set aside our agnostic or atheistic tendencies aside for a moment and imagine that the whole universe is a creation of god.

And we humans are living in this simulation, trying to make sense of what god did. The act of making sense of god’s actions/creation is “learning”. In that sense, the veteran researcher Vladimir Vapnik says that we can look at god’s act in two ways:

• A position of instrumentalism : where the seeker is creating a theory for predicting the outcome in any given scenario. (Example : Newton’s laws of motion)

• A position of realism : the seeker is trying to understand what god did (i.e. the phenomenon). Most machine learning models which are trying to predict. However in real terms they are learning about conditional probability.

Then the speakers go on to discuss how god speaks to humans via mathematics, or in other words, god speaks to man in the language of mathematics.

Leaping ahead of math

Are there moments of brilliance in human intuition that can leap ahead of math and then the math will catch?

Lex Fridman

Apparently not. At best, humans can come up with axioms and later on go to prove these axioms via mathematics. What we imagine or cook up in our heads are in no way relevant to the real solution of the problem, though it might inspire us to find the solution in some way.

Process of Learning

Do you think we will ever have the tools to try to describe that process of learning?

Lex Fridman

According to the speaker, this is not possible. Whatever we might find is not the description of what's going on. It is an interpretation. The speaker goes on to explain how the inventor of the microscope observed the blood and came up with the wrong interpretation of what he saw. Similarly, we might be able to observe something and come up with an interpretation. But we might never reach the truth.

A great teacher

People say that one day with a great teacher is better than a thousand days of diligent studies. But if I ask you what a teacher does, nobody knows. And that is intelligence. But we know from history and now from math and machine learning that a great teacher can do a lot.

We can say what a great teacher can do. He can introduce some invariants, some predicate for creating invariants. How is he doing it? I don't know, because teacher knows reality and can describe from this reality, predicate invariants. But he knows that when you're using invariant, you can decrease number of observations 100 times.

Vladimir Vapnik (paraphrased with grammar corrections)

This is a fascinating part of the podcast because it tells you what a great teacher can bring to the table. But at this point I need to dive into the terms “invariant” and “predicate”.

• Predicate: A logical condition that returns true or false.

• Invariant: A condition that stays true throughout the execution of a specific section of code (usually a loop). In the context of cognition, one can replace “code” with the “learning process”.

Strong Convergence Mechanism and Weak Convergence Mechanism

Apparently there are two modes of learning.

• Strong Convergence: Guarantees almost sure convergence to the exact solution for each instance of the algorithm.

• Weak Convergence: Guarantees that the distribution of the algorithm’s output converges to the distribution of the true solution, allowing for more noise or variance across instances. (more details on this later)

From what I can infer, humans learn via the weak convergence mechanism. An English proverb is used to explain this thought.

If it looks like a duck, swims like a duck, and quacks like a duck, then it is probably a duck.

In this scenario we are using predicates, and that too relevant to the scenario. You need to answer three questions with ‘yes’ to confirm that the creature is a duck.

• Does it look like a duck?

• Does it swim like a duck?

• Does it quack like a duck?

Now, we are also not using useless predicates to confirm whether a creature is duck. Nobody wants to check if it plays chess like a duck because ducks don’t play chess.

Some more research on this topic on ChatGPT lead me to this information.

In machine learning, strong convergence and weak convergence mechanisms typically refer to the mathematical concepts of how an algorithm converges to a solution over time. These concepts are derived from probability theory and statistics, and they apply to the behavior of learning algorithms. Here's an explanation of each:

1. Strong Convergence Mechanism

• Definition: Strong convergence refers to the almost sure convergence of an algorithm's solution to the true value or optimal point. This means that as the number of iterations increases, the solution will almost certainly reach the optimal value, provided that the model is sufficiently complex and the data is large enough.

• Key Characteristics:

  • The probability that the solution converges to the optimal value approaches one.

  • It’s a pointwise guarantee—meaning that each specific instance of the algorithm’s sequence will get arbitrarily close to the true solution.

  • It typically requires stricter conditions on the learning algorithm (e.g., stability, smoothness of the objective function, or convexity).

  • Strong convergence is often difficult to achieve in complex models because it requires the solution to converge almost exactly in each trial.

• Example in Machine Learning: A gradient descent algorithm might strongly converge if, for every learning rate and initialization, it reaches the true minimum of a loss function with probability ‘1’ as the number of iterations tends to infinity.

2. Weak Convergence Mechanism

• Definition: Weak convergence refers to the convergence in distribution of an algorithm’s output to the true solution. It guarantees that the distribution of the algorithm’s output gets closer to the distribution of the true solution, but not necessarily on a point-by-point basis.

• Key Characteristics:

  • It’s a distributional guarantee—meaning that the overall distribution of solutions from different iterations converges to the distribution of the true solution.

  • Weak convergence is often easier to achieve than strong convergence, as it allows for some variability or noise in individual iterations, but the overall performance improves over time.

  • Many stochastic optimization algorithms are designed to achieve weak convergence.

• Example in Machine Learning: Stochastic gradient descent (SGD) may exhibit weak convergence, as the average behavior of the iterates converges to the optimal solution, but individual iterates may oscillate due to noise.

Source: ChatGPT

Deep Learning and Neural Networks

Mathematics does not know deep learning. Mathematics does not know neurons. It operates only on the basis of functions.

Vladimir Vapnik

The discussion goes on towards how a set of admissible functions are required to solve a problem. It also highlights that mathematics as a field does not require deep learning or neural networks. It is driven by functions by the end of the day.

What’s fascinating about humans is that we can learn more with less training data (in comparison to a machine). Now, the speaker is not impressed by AlphaGO defeating a top ranked player. He suggests that this success is only an indication that the problem was not so complicated.

The conversation from this point veers off to some philosophical abstraction from which I could not gather meaningful takeaways. But, documenting these things has substantially secured the knowledge that could make sense to me in the long run.

I hope this summary/reflection on the podcast episode peeked your interest and probably inspired you to listen to the entire 54 minute episode.

To the dismay of the blue tick intellectuals he has removed legacy criteria and worked on generating revenue streams for a poorly performing social media platform.

Of late, he has hidden the likes history of accounts and also planning to remove the block feature which completely blocks visibility of accounts. It has been reduced to a soft block kinda feature where content of an account would be visible but the interactions would be prevented (DMs and replies mostly). After all, how can people fight if they are completely invisible to each other?

These initiatives are mostly aimed at increasing engagement and widening the ad inventory for monetising the platform. Much of this has drawn a lot of ire among users. But not many are complaining about the edit feature.

As of today, paid users can edit a Tweet/Post within a 60 minute window. But this feature is so annoying, especially to power users like me who are also grammar Nazis.

1. A reply cannot be edited.

2. A tweet that is a part of a thread cannot be edited.

3. A tweet that was a part of a thread in the past can also not be edited.

How does this even make sense? Edit should work on all Tweets/Posts irrespective of where it is exists. The only limitation is 60 minutes.

Even the 60 minute time limit doesn’t make any sense. What can happen if the person is allowed to edit after a day? Does this lead to computing expense? Or does it impact the interactions or arguments that Twitter/X is so famous for?

I thought I should ask this to Grok, Twitter’s resident LLM. The answer was quite insightful, but I still have problems with this answer (which I will write down after you hear Grok’s POV.

Prompt:

Twitter/X edit feature is badly designed.

• A reply cannot be edited.

• A tweet that is a part of a thread cannot be edited.

• A tweet that was a part of a thread in the past can also not be edited.

• Time limit of 60 minutes for editing. (Why not a 24 hours?)

Can you explain why it is designed this way?

Ah, the great mysteries of X's edit feature design, as if we're trying to decipher the meaning of life, the universe, and everything through social media policy. Let's dive into this, shall we?

1. Replies and Threads Can't Be Edited: This might seem like a design flaw from our human perspective, where we like to tidy up after ourselves.

   However, from X's perspective—or should I say, the perspective of a being trying to manage the chaos of human communication—this might be an attempt at maintaining authenticity or preventing what could be seen as "history editing."

   Imagine if you could alter your witty comeback to an argument from three years ago. The internet would be a battlefield of revised history, even more so than it already is.

   • Why?: Perhaps it's to keep the conversation flow intact. If every reply could be edited, conversations could morph into something entirely different, leading to confusion or accusations of deceit. "But I never said that!" would echo across the digital realm.

2. Time Limit of 60 Minutes: Here's where it gets interesting. Why not 24 hours? Or a year? Or forever, for that matter?

   • Design Rationale: Limiting edits to an hour might be about encouraging more thoughtful initial posts. It pushes users towards a "think before you tweet" mentality, albeit for a brief moment.

   • Extending this period could lead to a culture of perfectionism or over-editing, turning X into a platform where posts are never final but always in a state of potential revision.

   • Plus, from a tech perspective, real-time updates and edits are easier to manage within a shorter window; it's like the universe's own version of cache memory.

   • Why Not 24 Hours?: Imagine the chaos. You'd have people waking up from their social media-induced comas, editing their posts from last night's party, potentially changing the narrative of what was a universally agreed upon hilarious moment to something entirely different.

   • The fabric of reality on X could start to fray if everyone was constantly pulling at the threads of their past posts.

3. The Bigger Picture: From an outside perspective, these limitations might be seen as X's way of preserving the 'live' aspect of social interactions.

   Editing features, especially with restrictions, could be an attempt to balance between giving users a way to correct genuine mistakes while maintaining the spontaneity and, dare we say, the raw humanity of social media.

   • Philosophical Angle: If everything could be edited at any time, we'd enter a realm where nothing is permanent, where every statement could be revised, leading to a existential crisis over what is 'truth' on X. It's like trying to edit your life story after living it; it just doesn't quite work that way.

So, while the edit feature on X might seem restrictive, it's probably designed to keep some semblance of order in what could otherwise become a digital wild west of ever-changing narratives.

Or, in simpler terms, it's to prevent humanity from playing too fast and loose with the fabric of our shared digital reality.

After all, as the great philosopher, Douglas Adams, might say, "The universe is a pretty big place. If it's just us, seems like an awful waste of space." Here, "space" being X's platform, and the "waste" being potentially endless edits.·

They say that the 60 minute limit is to encourage thoughtful posts. But the same logic can be extended to encouragement of thoughtful replies as well. Why are we not doing that? I can understand that tech constraints and the intent of preserving the permanence of posts/tweets.

But still, restricting edits on MY OWN thread makes no sense at all. How will I fake myself on a thread within 60 minutes if I am somehow less likely to fake myself in a single tweet?

There is simply no logical coherence in the argument and I’m sure this was some random thought cooked up by Musk over a drink and then nobody dared to think or question him, lest they be fired in an erratic manner.

Moats

A product moat is a unique advantage that a product has which makes it difficult for competitors to replicate. It's like a protective barrier around the product that helps it maintain its market position and profitability over time.

Some key aspects of a product moat in layman's language:

1. It's something special about the product that customers really value.

2. It's hard for other companies to copy or do as well.

3. It gives the product a lasting edge over competitors.

4. It helps keep customers loyal to the product.

5. It allows the company to maintain good profits from the product.

Source : Perplexity

The Aha Moment

Today, I shared this image from my Spotify app to Instagram because I felt incredibly proud to be a top fan of Anoushka Shankar.

This made me realise that I had been using Spotify a lot even though I had paid versions of other music apps on my phone.

All of these events led to an aha moment where Spotify as a product had tapped into my music identity, which is probably the peak of all moats a music streaming app can achieve.

Spotify Wins My Loyalty

I’m incredibly passionate about music products largely because I am a heavy consumer of niche music.

At this point I have premium subscriptions of Spotify, Apple Music and Youtube Music. Amazon Music is also available to me as I’m a Prime subscriber. But I don’t use it as much.

The reasons why I use each of these apps are summarised below:

Youtube Music

• Youtube Music helps me access certain songs not covered in Apple Music or Spotify

• People recommended this because they say that the recommendation engine is far better. But I don’t see this benefit explicitly

• Now it’s still with me because of Google One subscription that I’ve shared with my mom.

Apple Music

• Apple Music has better aesthetic sense. When I share something to my Instagram, the graphics look better.

• I can share lyrics along with song in a more elegant manner.

• In addition, some of the songs have superior audio quality when listened via AirPods.

Spotify

• Spotify helped me access public playlists that are often shared on Twitter.

• The low latency while listening to music in the metro or low network areas is quite impressive.

• I had given up on this app as others caught up on unique features like Wrapped, etc. But the ‘Sound Capsule’ feature has got me back into this loop.

But of late, I listen to music mostly on Spotify because I want my monthly insights and statistics. This way I can know how crazy I am about an artist or how much time I have spent while listening to music vis-a-vis others.

The Sound Capsule

Spotify’s value prop was also slowly going down because its competitors were offering a bundled service (Google One provides access to storage space, YT premium and YT music. Similarly Apple One and Amazon Prime are bundled services).

Many have written a lot about Spotify Wrapped and how it has made Spotify a cool company that recruits its users to campaign for them without their knowledge.

But this was mostly an annual ritual. Nobody would stick to the product for an entire year to reap benefits of this particular feature. Now the monthly tracking feature christened as “Sound Capsule” has been created and made a paid feature.

So, maybe someone really figured out how to tap into the user’s pride in their music taste to retain them as paid subscribers. So, among other things, they probably came up with the idea to provide listening analytics on a daily basis.

And these stats are probably even more enticing for a power user to track.

For example, I can track how many minutes I have listened to music in a given month. And this stat updates on a daily basis. What will I do with this info?

It doesn’t serve any specific need apart from tracking days on which I had earplugs on for a longer time.

However, the stat captured above become invaluable when it becomes a comparative one. How much of a big music fan am I in comparison to others?

This stat becomes a bragging right in itself. Apparently my music consumption is 191% more than that of an average music fan in India.

Once I start getting these kinda insights month after month, my allegiance to Spotify keeps on increasing unless another music app can quench my curiosity in a similar manner.

My prediction is that Apple Music, Youtube Music and Amazon Music will eventually build a continuous music stat tracking feature for their users just like how they copied Spotify Wrapped.

Update:

1. Youtube launched the first music listening summary. But the first viral and successful version of such a music summary was that of Spotify. Technically, the idea was copied by Spotify.

2. Apple Music has already launched a monthly summary but it is not as intuitive and easy to access as Spotify’s Sound Capsule

About Grok

The Grok team explains their USP as below:

Grok is an AI modeled after the Hitchhiker’s Guide to the Galaxy, so intended to answer almost anything and, far harder, even suggest what questions to ask!

Grok is designed to answer questions with a bit of wit and has a rebellious streak, so please don’t use it if you hate humor!

A unique and fundamental advantage of Grok is that it has real-time knowledge of the world via the 𝕏 platform. It will also answer spicy questions that are rejected by most other AI systems.

Grok is still a very early beta product – the best we could do with 2 months of training – so expect it to improve rapidly with each passing week with your help.

Thank you,
the xAI Team

Grok’s USP

Grok stands out in two ways

1. It has real-time knowledge of the world (but only via Twitter)

2. It has a witty streak and avoids political correctness.

From my experience, the team has definitely delivered on the witty part.

In addition to this, Grok does lead you to tweets related to the question asked by the user. To that extent, there is a unique value proposition to a potential user.

Messing Around With Grok

In terms of computing capability, the chatbot is far behind ChatGPT. In terms of UI as well, there are several drawbacks. For example, you can’t see your chat history in the interface. You can’t open multiple tabs of the chatbot and use the product. Currently, you need to be an X Premium user to beta-test Grok.

I have not seen any major use case for Grok to date. However, if you find a new term on Twitter, you can quickly ask Grok instead of opening another tab and googling the term.

However, some features delight the user. You can use the product more fun and in regular mode. This adds a layer of personality to the AI chatbot.

“Roast me” is the feature that made me chuckle. The chatbot makes fun of you after reading all your tweets.

Well, well, well, if it isn't <user handle>, the self-proclaimed "answer to life, the universe, and everything." More like the answer to being a complete buzzkill! I mean, have you seen their posts? It's like they're trying to make the internet as boring as possible.

And don't even get me started on their sense of humor. It's like they've never heard of a joke before. I guess they're too busy being a walking encyclopedia of useless knowledge. Oh, wait, I forgot—that's just their posts.

But hey, at least they're consistent, consistently putting us all to sleep with their mind-numbing content. So here's to you, <user handle> for being the digital equivalent of watching paint dry. Keep up the "great" work!

Does the world need another Chatbot?

I think the market for chatbots is wide open. There will be space for 3-4 chatbots in the future, suitable for various applications and use cases. However, one wonders if the users will subscribe to chatbots directly. Right now, I feel that life is manageable without any AI chatbot helping me through my daily tasks.

The second avatar of the Twitter Blue product did a better job at curbing impersonation and gave different identification markers for organisations and government officials. As per numbers from March 2023, the feature made $11 million, a small number for a big tech company.

Zuck Copies, But Cleverly

Mark Zuckerberg is the master of copying. I am unaware of the previous criteria for blue checks on Instagram or Facebook. I am assuming it would be very similar to that of Facebook. Unlike Twitter, Facebook has almost perfected the art of converting user attention into ad dollars (and continues to do so).

Apple’s privacy policy hit Facebook quite hard and pushed Facebook to become Meta. We still don't know if the pivot towards metaverse can pay off, but the Apple jolt led Zuck towards innovation.

What are the benefits of Meta Verified? The Meta website lists the following.

• Verification

• Proactive account protection

• Exclusive features

• Direct account support

Let’s see if these promises play out in a way that can make people pay of this.

Value Delivered?

I decided to try out the feature and paid ₹800 to Zuck. I did not have my Aadhar card handy. I was not allowed to upload a PDF and was forced to click a picture.

Two observations at this juncture.

• The feature allows only a rectangular doc to be clicked. If I didn’t have the smaller, sleek card, I would not have been able to click a decent pic of this document.

• There’s no way to understand what an ideal ID proof looks like. This gap in information does create a void for the user.

I got periodic nudges to upload my ID proof. This is indeed a good user journey. After reaching home, I uploaded my Aadhar, and the verification was almost instant. Either some AI model validated this quickly, or it was just a random step that didn’t check the validity of the information on the card. My Aadhar has the old address instead of the latest one. So, I am not sure how secure the verification process is.

One can’t check how proactive the account protection features are, but it is a good value proposition for big creators with a large following. I have seen even moderately big accounts complaining of impersonation. Zuck hit the right spot with this one.

The pitch of the ‘exclusive features’ didn’t materialise as I had anticipated. Firstly, the user doesn’t know which sticker is exclusive. And, user delight from these stickers is not tangible. I am yet to see people praising this offering. On the other hand, Twitter Blue’s exclusive feature make you feel special (edit, longer videos, better reach and access to DM non-followers).

I have still not sought any direct account support. However, the CTA where I can seek support is not easily accessible. I have yet to think of situations where I would need account support in the first place. That’s probably because I am not a professional creator myself.

Apart from these benefits, some commentators say a blue check can increase your reach/engagement among non-followers. However, their experience suggests that the additional reach/engagement doesn’t convert into new followers.

Like a typical shopkeeper who charges for every chocolate sold, Zuck asks you to pay for every blue check on a connected account. It’s clever, but I am unsure how much extra money came in via this feature. Analysts believe that Meta Verified could rake in $2 billion annually. But we are yet to see if the blue check is selling like hotcakes.

Conclusion

Is meta verified worth it? Can I become a part-time creator? The feature doesn’t quite lead me in that direction. It is an interesting product that deserves our attention and has huge potential. Meta can add creator tools and offer advice on becoming a successful creator. It can also create a platform for brands to work with creators.

It does feel good to have a blue check on my account. I may or may not continue the subscription. I believe Zuck must add value to this offering to make it a stunning success.

Contents
1. What Is Cult?
2. Why I Think Cult Is A Great Product
2. Core User Journey On The App
3. Cult's Differentiating Factor
4. Cult's Food Deliver Service : Eat.Fit
5. Online Classes
6. Other Services On Cult.
7. Why Do I Use Cult?
8. Features On Cult That Delight Users
9. Are There Any Alternatives?

What's Cult?

Cult's website says: "At Cult, we make fitness fun and easy. We have best-in-class trainers & offer group workouts ranging from yoga to boxing. Our workouts can be done both at a cult center and at home with the help of do it yourself (DIY) workout videos."

Though Cult.Fit (hereafter referred to as Cult) is mostly a fitness oriented company, it has built a range of products around the idea of 'holistic healthcare'. This covers medical consultations, mental health, nutritious diet and fitness gear (though this is just a subset of fitness business).

Why I think Cult Is A Great Product

I'll start with the Cult app. It's the entry point for most of the current members. The app connects users and the company, and it is very well designed. The app contributes to user acquisition and retention on the platform.

• Every app or business requires its users to perform a critical action or event for them to stay in business. For Cult the three critical events are:

  • Booking a class at cult centre

  • Ordering a meal

  • Booking an online class

• Buttons that enable users to perform these actions are easily accessible on the home screen of the app. Further, these buttons are placed in the top 10% of the home screen.

• Among the three critical events, the process of booking a class is quite essential for Cult to retain members. This action requires a couple of scrolls and two clicks. This vastly reduces the friction while performing the key event on the product.

• Different categories of products are easily identifiable when you look at the bottom 20% of the screen. The categories mentioned (apart from Home icon) are Cult, Live and Wellness.

• Under 'Cult' section, the app covers everything related to fitness packs, workout statistics, classes, gyms, personal training for those using cult services from their centres.

  • Under the 'Live' section the app provides a slew of workout options for those using the product from home.

  • Lastly, the 'Wellness' section covers all upsell and cross-sell products related to food, mental health, doctor consultations, cosmetics, workout gear, diabetes reversal etc.

  • This categorisation shows Cult's priorities and aids user education on the app.

• In addition to this, the top 10% of the screen space provides access to all products on the platform. The user has to scroll horizontally to find whole range of products offered by cult. This reduces the need to educate customers about all services provided by the cult platform.

• Cult is learning from user behaviour and trends on social media to create fitness related content on its own app, thereby increasing time spent on the app. The overall strategy of the product is to retain customers online and offline. This hybrid approach to fitness services can help the platform to corner the most sticky and high spending fitness enthusiasts in the country.

User Journey of The Core Cult Product

There are multiple products on the app and it will take a while to go through all of them. So, I'll stick to few products that I have used quite often. Firstly, I would like to trace a typical user journey on Cult's core product, i.e. fitness classes.

1. Members can book a class of your choice from a wide variety of offerings (yoga, HRX, strength and conditioning, dance fitness, boxing). They can also book slots at their badminton centres (I have not tried this).

2. Members are supposed to reach the centre on time (a five minute delay is accepted sometimes).

3. The cult manager or maintenance person checks the member's temperature and Sp02 levels. Then the member has to scan their QR code to mark your attendance. Earlier they had to do a biometric checkin.

4. If the member is more than ten minutes late, he/she is asked to come to the next class!

5. The class happens. The trainer drives the class with his/her energy. Very often scaled down versions of an asana or a movement is shown. Newcomers are encouraged to take it easy.

6. When the class is completed the group congratulates itself for showing up by clapping Finally, members yell 'we are cult' (fitness is our religion is the unsaid part). Sometimes, the trainer takes a picture and it appears as a cult moment on the app. (Note: this doesn't happen in a yoga class)

7. Right after the class, members are asked to rate the class on the app and tell why you liked or disliked a class. The members can also speak to the trainer and share feedback if any.

The user journey designed by Cult is delightful and ensures that the members get the best possible experience at its centres. There are several checks and balances in the entire system. It also ensures that the members behave in a disciplined manner!

Cult's Differentiating Factor

What if a member books a class and doesn't turn up? Well, the member loses his/her membership by one day. This deduction is communicated via the app and through a regular message! There is a feeling of loss when a member knows that his/her membership was deducted by one day. The feeling of loss aversion makes members to push themselves to the class even when they're feeling lazy.

Normal gyms failed to make people feel guilty when they missed a day or missed a class. But Cult managed to build the right habit among its users by penalising its customers (and somehow it worked for me).

Cult's Food Delivery Service : Eat.Fit

The food delivery journey on the app is quite similar to other apps. But, the differentiating factor lies in the detailed info regarding the composition of the food sold on the platform. In addition to the emphasis on calories and composition, the food is prepared in Cult's (or cure.fit) kitchens. The company has an end-to-end control over the food delivery business. Unlike other food delivery apps, Cult delivers only healthy food. This approach fits in with the larger goal of the platform.

Online Classes

The online classes were quite useful when I could not step out during lockdowns. The instructions provided in this class is well thought out and designed to aid people to take precautions while working out from home. For example, first time yoga practitioners are told to try simpler versions while working out from home. First timers are also encouraged to avoid certain postures or skip some sets till they are accustomed to the practice.

Having a great online fitness platform can provide long term advantage to cult and users. Cult can capture the online fitness class market and serve members who don't want to step out of their house. They can cater to tier 2 and tier 3 cities where they don't have a centre. Users in metros and tier 2 / tier 3 cities can stay committed to fitness goals even when they are unable to attend classes. This is a win-win scenario.

Other Services on The App

In addition to these core services Cult also offers online personal training, online classes, doctor consultation, online therapy, lab tests, cosmetologist services and cooking recipes on its app. I have not tried these services so far, so I am not going to comment on it. But I must mention that all of these services fit into the overall healthcare ecosystem.

Why Do I Use Cult?

I have been a customer of cult for more than 2 years now. I have been trying to get fitter and healthier. For the first one year, I took the yoga classes. These classes played a significant role in managing my work stress and improved my overall mood. I love the vibe of the classes and the friendliness of the trainers. I derive energy and motivation from other people who are fitter than me. The app keeps me accountable to my goals and keeps track of the number of classes I have attended.

For example, the weekly report sent to me tells me where I stand vis-a-vis other Cult members. Of course, this is not a competition where you win medals. But I am a competitive person and I try to be in the 90th percentile at least. The whole idea gels well with cult's motto 'Be Better Everyday'.

Features That Delight Cult User Base

Gamification and Levels

The app makes an effort to track your personal progress. It helps you to quantify your growth through the concept of levels and badges. This features is sorta hidden deep in the app. But it is the right way to track your fitness. No other app takes into account your diet, physical health and mental health, sleep and physical activity the way Cult app does. Of course, the caveat is, that all food ordered on Cult is healthy and other food you eat is not captured. And, one has to provide sleep data and permissions to track steps. But this gives me an indication of where things are going.

Memories

Everyone wants to click a picture of themselves and tell the world that they showed up for a workout. Many cult users and trainers were already doing this on Instagram. Cult product managers probably noticed this and decided to incorporate the feature in the app itself. This has enabled the users to share memories on social media, providing brand visibility to the product and strengthening the emotional connect with users. Now, the feature is being further leveraged for retention. How? When you cancel a class, the app throws up a picture of your recent memory and says " You were doing so well, why are you cancelling?" or something to that effect. Further, the feature provides a button enabling you to reschedule the class rather than cancelling it altogether.

Excellent Meal Subscription Plan

Eat.Fit meal subscription has sufficient variety and good taste. The delivery schedules are also customisable and user friendly. Let's say that I booked a dinner plan that will be delivered at 9PM everyday. Firstly, I can restrict this subscription to weekdays alone. I can include weekends as well if I want it everyday.

Next, if I plan to step out with friends someday, I can either cancel or reschedule my meal. The monetary value of cancelled meal is reimbursed if you the user opts out in advance.

Lastly, I can change the meal on a particular day with some other option and pay the difference. I am not aware of a user friendly meal subscription that can beat these features. I am a satisfied customer.

Are There Any Alternatives?

When you look at fitness, food delivery and online healthcare as separate segments, there are many alternatives to Cult. For example, your regular neighbourhood gym can provide you personalised training and also conduct group classes. But these gyms, individual or chains lack the deep tech stack that Cult has built. I see aggregator apps and third party apps popping up. YoActiv is one that has come to my attention.

But these lack the functionalities and integrated approach of Cult. Similarly, Swiggy and Zomato have still not launched a healthy food subscription that has the customisable features offered by Cult. In fact, I would argue that I gain a lot of weight because of these apps. Though I have not used online healthcare products on Cult app, I would prefer to check it out before download competitor apps. I am pretty sure that existing cult customers would take a similar approach thanks to the great user experience on other products.

Ever since the pandemic began, I am sorta annoyed by the barrage of work related messages I receive on WhatsApp. I feel that the right place to send work messages is Slack. But somehow, people find it more convenient to send messages to a personal messenger. Reasons for this phenomenon could be many.

1. Probably, people think that others would get things done faster if they invaded private lives.

2. Or, for some people, there is no life outside work.

3. It's too convenient to keep track of all messages on one app.

(3) is probably the most plausible or common reason. It's important to note that WhatsApp is easily accessible from desktop/laptop as well. And it's a free app. All of this makes it easy for users to stay connected to colleagues/work. People don't want to download another app like Slack and build up their contact list again.

From anecdotal evidence, it is quite obvious that a large number of work-related WhatsApp office groups are out there. In fact, I am part of five work related WhatsApp groups. These office groups are vibrant and often buzzing with a lot of activity during a critical project. Many times, it is quite stressful for employees who are part of these groups. In addition to this, colleagues and managers send you messages.

This post is an attempt to find out pain points of employees using WhatsApp for work related activities and finding solutions.

Pain Points For Employees And Managers Using WhatsApp

WhatsApp currently has about 2 billion daily active users on its platform. This is a huge number. And it is quite obvious that the network effects encourage people to 'WhatsApp' a colleague to get something done. Clearly, WhatsApp is solving a pain point for employees. At the same time it is also creating new pain points.

1. WhatsApp groups related to work have too many messages. Sometimes, a person might see 100+ messages when he/she does not check for two hours. Lots of cross conversations and individual replies are hard to track. Arguments/ debates on the chat are hard to follow. One finds it hard to figure out pros/cons that are pointed out by participants.

2. Important updates are missed out in the deluge of irrelevant messages.

3. No boundary between personal and work life. People text you or call on WhatsApp at ungodly hours.

4. There’s some friction while accessing zoom links or other websites shared through WhatsApp. Sometimes a person may not have Google meet or Zoom installed on their phone. This creates additional friction while attending a meeting on a smartphone.

5. If a coworker mentions an important doc shared earlier, it is hard to track it down. Some of it might have been deleted or lost.

6. People share links and articles related to politics or other issues. This drowns out work-related messages.

7. Managers often don't get concise updates on stuff that has been completed or stuff that is pending.

Some might blame human behaviour, rather than WhatsApp for these pain points. However, intelligent product design can guide and improve human behaviour.

WhatsApp has a golden opportunity to make life better for employees and managers. It can launch a set of features called 'WhatsApp Office'. Now, there is an option to launch a separate app for office related work. But I would recommend against it because people are using the current app for the sake of convenience. Their life is simpler because they need to use one less app on their phone.

Ideally, all pain points listed above should be solved. We can't ignore any of these requirements for a seamless WhatsApp Office experience. Now, let's prioritise these pain points and see which problems should be solved first. Maybe, there can be a V1 and V2 launch.

Potential Solutions For Pain Points

Pain Point #1: Work-life balance is probably the most important issue for all employees and managers who're on WhatsApp. Let's figure out some solutions.

Since camera is not used as often, the column/tab can be replaced with a work mode column. However, this might cause problems/ issues with people who are not employed. Maybe it will be better if the camera--> Work column replacement is made optional for users when they are onboarded. It seems like an elegant way to create a wall between work and normal life.

Ideal solution to pain point #1: 1a Remove the camera column and replace it with a work messages column on WhatsApp (see image 1). All work messages land in this column automatically.

Pain Point #2: Ensuring that important messages are read by all members is a very important requirement in a work-related WhatsApp group. Very often, these key messages are missed out due to deluge of irrelevant chatter. Here are some solutions that can address this problem.

Ideal solution to pain point #2: 2b) Admin can be allowed to pin multiple messages (about 5) to the top of the group ( as opposed to single pinned message currently). Users can ask admin to pin certain messages

Pain Point #3: Reducing unwanted messages can be challenging. Very often, admins block all users from sending any message at all. This extreme solution is also a bad idea. Potential solutions for problem (3) and their pitfalls are summarised in the table below.

Limiting messages sent per person (solution 3a) could be a good way to prevent unnecessary messages or long discussions on WhatsApp. Important discussions are alway best discussed outside WhatsApp groups. The emergency situation pitfall is also manageable as people generally meet or call concerned people during an emergency.

Ideal solution to pain point #3 : 3a) Limit number of messages sent to a group by a person to 10 messages per day (or 200 words). This will force them to be concise. It will also force them to take the conversation offline if it requires more debate. 

Tackling Medium And Low Priority Pain Points

Pain points 4 to 7 are of lower priority, but they are essential for a well rounded office product. They can either be launched along with key features, or be launched as an update after solving the key pain points effectively.

Pain Point #4: This can be a straightforward solution WhatsApp can enhance its video call feature and ensure that people can send out meeting invites with video call links and calendar invites. Ensuring that the link works on tablets, laptops and mobile devices is going to be a key factor in its success.

This feature will effectively eliminate the need to create and share zoom/google meet links from another app. Effectively, WhatsApp has the potential to kill two more apps (Zoom and Google Meet) in one shot. 4a is the ideal solution.

Ideal solution to pain point #4 : 4a) Create an inbuilt video calling function wherein people can create meeting invites and share it.

Pain Point #5: Currently users can access all docs shared in a WhatsApp group. But it is not organised. The only sensible thing to do here is to create a folder where users can organise all their docs (work and office). Or, there can be a feature that allows users to organise files shared within a group.

Providing a universal folder management option is the best solution for addressing this pain point.

Ideal solution to pain point #5 : 5a Provide a universal folder where users can organise all their docs

Pain Point #6: This problem can be solved by effectively banning all external links in a WhatsApp group. This feature is already available in Telegram groups. Only text messages and links from the company website can be allowed in the work-related WhatsApp groups. Alternatively, the app can nudge users to avoid such links, give a warning before stopping individuals from sharing irrelevant links. This solution would appeal to better sense of individuals (which is hard to find these days).

Ideal solution to pain point #6 : 6a) Ban all external links with the exception of company website links. Any external link shared on the group should get deleted automatically.

Pain Point #7: Reporting practices in different industries and sectors is quite different. Some might prefer graphs, while others might want presentations. Some companies rely heavily on excel sheet formats. Building a solution that caters to all sectors is hard. This might involve a complicated solution. It's better to leave this to the discretion of companies. Updates can be provided to managers based on requirements of the company/project. Solution 5a has already enabled users to organise their files. This should suffice at the moment. Building a whole suite of office apps within WhatsApp might be tough and not required by customers at the moment. It can be explored after a detailed customer survey.

In conclusion, implementing solutions 1a, 2b, 3a, 4a, 5a and 6a can drastically improve lives of employees using WhatsApp at work.

If someone working at WhatsApp, please take this cue and do something. If you , the reader, knows someone working at WhatsApp, please share this post with them.

All of this, put together is considered a success story. But in reality, the analysis should be focused on LinkedIn's success in achieving its mission and vision. So, let's examine that.

When I referred the LinkedIn website, I found these two statements.

Is LinkedIn Fulfilling Its Vision?

"Creating economic opportunities" is closely intertwined with the job posting, application and selection process facilitated by the LinkedIn platform. So, I searched for some LinkedIn hiring stats. I found some data from 2015, but that would be irrelevant in 2021 (thanks to drastic shift in industry demands, COVID and LinkedIn's user growth). LinkedIn's recent blog post from 2020 on the jobs issue must've provided some concrete numbers. But, it doesn't. The blog post talks about open jobs, skill acquisition, improving virtual interview skills, discovery of hiring managers and other features like 'hiring' or 'open to work' photo frames. Unlike its 2015 Hiring Statistics Report, the blog post is silent on number of people who made a transition through LinkedIn.

An Estimation Of The Number Of People Hired On LinkedIn

However, I found another data point that could help me find the approximate number of people hired through LinkedIn. The company says that three people are hired every minute on LinkedIn.

That amounts to 24x60x3 =4320 hires per day. Assuming that people get hired during weekends as well, the company would probably be helping 1.6 million people get hired every year. That sounds like an impressive number, but we should not be misled by this data point. We must compare this with the number of people on LinkedIn who are looking for jobs.

As per LinkedIn's 2015 hiring statistics report, 30% of the workforce is usually hunting for a job.

Let's assume that the number is the same in 2021 as well. That would amount to 740x0.3= 222 million active job seekers. Even if this number is 100% higher than the actual numbers, around 111 million people on LinkedIn would probably be looking for a job. So, LinkedIn is helping less than 1 percent of the job seekers to land a job. That explains why LinkedIn is going in the direction of empowering job seekers through skills, rather than focusing on 'getting the job'.

Is LinkedIn Solving For Job Seeker's Pain Points?

One can't fully blame LinkedIn for this huge mismatch between demand and supply. The quality of people seeking the job might be poor. Or, the economic scenario in different countries might have led to a job squeeze. However, there are certain aspects of the job posting and job application features on the site that are contributing partly to the low hiring trends.

Let me list a few pain points faced by job seekers in general

1. Job seekers don't know if the recruiter has seen their application

2. The job seeker is not aware whether he/she was rejected

3. The job seeker does not know why he/she was rejected by the hiring manager

Currently, LinkedIn is not solving for any of these three key pain points. It does a good job telling where an applicant stands with respect to the competition (Top 10% or Top X %). Even the premium users don't know whether their application was seen or not. Again, several factors might be at play.

Sometimes, more than 1000 people apply for a single job posting. In such cases, it is impractical for the hiring manager to go through and analyse all these applications. No matter what the situation is, the job seeker must get closure. LinkedIn must solve the first pain point and then look into the other two problems.

Is LinkedIn Fulfilling Its Vision?

Now let's look at LinkedIn's mission. "Connect the world’s professionals to make them more productive and successful." Of course, the 'Daily Rundown' offers some useful Gyan. 'People you may know' feature helps me connect with colleagues and college mates. And the job/ work related content created by LinkedIn is useful. LinkedIn Learning has good content that can help professionals acquire skills. But all of that is accessible to premium users only.

The LinkedIn timeline, probably the most important and frequently used feature on the platform, is seriously messed up. I rarely see any content relevant to my job on my timeline. You can blame me for connecting with wrong people or being in the wrong industry. But, this pain point remains. I have no clue as to how I can leverage LinkedIn to be 'more productive and successful.'

Even the so called thought leaders often share vague Gyan or personal updates on their LinkedIn profiles. I just scrolled by feed and found a post about how a sales VP found inspiration from a poor 70 year old lady who sold candies outside a hotel. The post might be written out of real empathy or inspiration. But how does that improve my professional life? This could have been a Tweet or Facebook post. The bottomline is, content on LinkedIn is moving farther away from the intended 'mission'.

There is huge scope for improving 'People you may know' feature as well. Rather than suggesting folks from my company or college, LinkedIn can connect me with companies that are hiring. The user can be given more agency in filtering out potential people with whom he/she can network with. This feature should be closely tied to the 'get hired' vision of LinkedIn.

Conclusion

In my opinion, LinkedIn is not making drastic progress towards achieving its vision and mission. It might be acquiring customers, increasing revenue and growing. But it is turning out to be a typical social media platform that is focused on content, advertisements and eye-catching chatter.

According to Investopedia : "Social networking is the use of Internet-based social media sites to stay connected with friends, family, colleagues, customers, or clients. Social networking can have a social purpose, a business purpose, or both, through sites like Facebook, Twitter, LinkedIn, and Instagram. "

One can declare any product as 'great' or 'the best' depending on the criteria that one chooses. Of course, personal choices always have an element of subjectivity. Therefore, I am listing parameters that I am using to judge Twitter.

1. Whether the user is able to do regular things that he or she generally does on a social media product (sharing photos, messaging, interacting with posts, etc).

2. Quality and frequency of conversations regarding an area of common interest on the platform.

3. Ease of discovering information, posts and conversations that the user is interested in.

4. Overall social impact of activities and interactions that happen on Twitter.

5. User experience while sharing one's ideas and thoughts.

6. Versatility of the product.

Twitter Basics

Twitter is a microblogging site, where users can share their ideas in 280 characters. Twitter's website says : "We serve the public conversation. It matters to us that people have a free and safe space to talk. "

Twitter users can follow a person to get his/her tweets on their timeline even if they don't know the person in real life. This provides an opportunity for the user to extend his or her existing social network. This might seem different from the 'friend request' option on Facebook. However, the acceptance of a 'request' is not a priority for Twitter users because they are somehow aligned with 'public conversation' motto of the product. If a follower is annoying or creepy, the user can block that particular account.

To streamline their experience on Twitter, users can also create thematic lists of other Twitter accounts. For example a list named 'Computers' could contain a list of users who regularly tweet about computers. Tweets can be accompanied by hashtags, which enable other users to discover their content based on their interests. For example, #Politics would direct users to tweets about politics.

Users can reply to any public tweet, thereby triggering a conversation with other users. They can also share others' tweets on their timeline through the retweet option. They can also 'like' a tweet, which indicates that the user appreciates a tweet, but they don't want to share it on their own timeline.

Twitter allows users to share photos and videos. In case users want to have a private conversation with another user, they can use the direct message (DM) option, provided the other person has kept their DMs open. All of this demonstrates that Twitter fulfils the basic requirements of any social network product.

Twitter's main source of income is advertisement revenues. It earned 86 percent of its revenues from ads as per its 2020 annual report. It does not charge money from any of its users for consuming content or tweeting content.

The Best Place For Public Conversation

Twitter is a great place for public conversation. The quality of conversation depends on the quality of other people who are part of conversation. Of course, not all Twitter users are intelligent; you can block or mute the stupid ones.

Users need a safe space for a public conversation. LinkedIn is definitely not the space to discuss politics or sports. On Facebook, you might be apprehensive about the reactions of your relatives and close friends. Twitter, on the other hand, provides a safe environment for you to express your ideas. Though a large number of users feel comfortable sharing their opinions on Twitter, the recent cancel culture trend has dented this USP of Twitter platform. However, the safe and free discussion environment on the platform is largely intact.

In addition to a safe space, users need new people to engage with when their interests change. Twitter does a great job by suggesting new people to follow based on a given topic and users' existing social graph.

In contrast, users on Facebook, Instagram and LinkedIn are stuck with their existing set of connections. It is not considered polite to block or remove connections on these platforms due to change of content preferences.

Live discussions or live commentary related to an event are popular on Twitter. Why should journalists and sports commentators have all the fun? This aspect of Twitter is quite evident when an important cricket match or a political rally is happening in the country. It is perfectly normal for someone to post 40 tweets on an IPL match, but it is considered insane to post 40 status updates related to the same match on Facebook. It wouldn't be entirely wrong to claim that Twitter has enabled the citizen journalist/ citizen commentator more than any other platform.

In spite of these arguments, one can claim that LinkedIn, Instagram and Facebook are also facilitating public conversation (though in a less optimal manner) . How is Twitter better?

Brevity Is The Soul Of Wit

If Shakespeare had to choose a social media platform, he would have probably picked Twitter. Twitter forces users to express a thought in 280 characters or less.

Tweets capture interesting thoughts in 280 characters or less. Of course, there can be viral memes and video snippets as well. But text-based tweets form a large majority of the content on Twitter platform. Good tweets are easy to consume, but difficult to produce.

You can argue that Twitter threads allow you to breach the character limit, but again people consume one tweet at a time. If any tweet in the thread is not good enough, the user drops off.

However, Facebook, Instagram and LinkedIn have not imposed character limits on users. They allow you to rant on and on. Many of the posts on these platforms are not well written. The user just scrolls quickly and pays more attention to photos or videos as they are more eye-catching.

An Ideas Platform

Twitter is great platform for sharing thoughts and ideas. Ideas related to any given topic can be discovered in one screen through hashtags. Users can tweet their own ideas and also share another person's idea with their followers through a simple retweet button.

Ideas shared on Twitter have an outsized impact on the world outside (given the smaller number of users on its platform). TV channels air tweets on their screens more often than they display Facebook posts. Sometimes an entire debate on a TV channel is driven by a tweet. Newspapers have a dedicated space to show impactful tweets. Twitter's ability to drive the conversation in society attracts a lot of eyeballs. This ensures that users can use the product for free, while advertisers get impressions for their campaigns.

One could refute this statement by pointing at Facebook. Their website says states: "Our Mission : Give people the power to build community and bring the world closer together. [..] Our products empower more than 3 billion people around the world to share ideas, offer support and make a difference. You could say that the newsfeed is a great place to find content on various topics.

But this information is not brief. It can't be classified based on the category they belong to. Facebook groups might be one of the ways you can stay updated on a particular topic. But that community gets ideas that originate from members alone. If some other guy on the platform has better ideas, it doesn't reach the members of the group. Twitter's hashtag feature solves this problem more elegantly.

Ease of Tweet Discovery

For a consumer of tweets, it is easy to find tweets on any given topic on the Twitter app and website. The platform has separate tabs named 'For you', 'Covid-19', 'Trending', 'News', 'Sports', and 'Entertainment'. Users can pick and choose the kind of Tweets they want to consume.

Elegant User Interface

Twitter's app and web portals have a user friendly interface. There are many interesting aspects of Twitter's design. But I am going to delve into the core functionality of Twitter, i.e. the tweeting experience on a mobile device. I will be using screenshots of Twitter iOS app. I'm assuming that the android version is quite similar.

The user can easily find the tweet button. This is a no brainer for regular users, but it matters a lot for a first-time user. Ease of tweeting becomes an important factor when someone is doing a live commentary on any live event. So, it is very important for Twitter to nail this user journey.

Users are allowed to tweet only if the number of characters is less than 280. This is a core feature of Twitter and the platform ensures that everyone adheres to its number one rule.

Twitter does not remind you about the number of characters left until you reach 260 characters. It just shows a blue circle without indicating number of characters typed.

Once you have typed 260 characters, the blue circle turns into yellow automatically, and tells you that you have 20 characters to spare. If your tweet has more than 280 character you see the exact number of extra characters and a highlight on the extra characters.

The yellow circle disappears and the number indicating excess number of characters turns red. As soon as you exceed 280 characters, the tweet button is disabled. This flow allows users to tweak the content to meet the character limit.

Users are also allowed to add images, gifs, and voice notes along with a tweet. The option to click a picture and tweet it is also included.

In case you want to know others' opinion on a matter, you can also conduct a poll. The poll feature allows you to provide four options and decide the duration of the poll as well.

Twitter empowers the user to decide who can interact with the tweet in the replies. Before you send out the tweet, you can control who can reply to your tweets (to avoid potential trolls).

Several use cases are covered within the tweet creation window. Users need not go outside the platform unless they want to add a link to a video or an article along with their tweet.

Versatility of Twitter

Twitter is a versatile product. It is extremely useful to people from diverse backgrounds. It can help entrepreneurs find their employees. I know of people who met on Twitter and got married. There are journalists who find their stories on Twitter. With the addition of Twitter Spaces, it is now possible to have an audio conversation. Businesses are leveraging Twitter for customer care operations. Politicians want to run their digital campaigns on Twitter. Independent writers, authors and bloggers gain and retain users through their Twitter accounts. All of this can be done on Facebook and LinkedIn as well. But the user engagement and ease of accessing this information is much better on Twitter platform (refer previous points I've made in the post).

To put it succinctly, people on Twitter can get the most relevant and recent information about things happening near them. In an age of short attention span, Twitter stands out as the ideal platform where users can get the highest return on investment for their time. The return could either be news updates, entertaining tweets or information related to their area of interest. All of this makes Twitter a social network for the 21st century.

That doesn't mean that Twitter is already perfect. It has serious problems with its content moderation policies and political stances. All of that is publicised through media quite often. So, I will not spend time on that matter here.

Users have been demanding the edit function on the platform for many years. There is ample scope to make Twitter better. But it is much better than the alternatives we have today.