Every week, I’ll share a 3-minute read with you - an insight from my learnings, a useful tool that reduces your cognitive load, cool founders building cooler stuff, plus an update from our team at Shram.

🪐 The paradox of perfection

We all have heard the advice: don’t chase perfection. It’ll stall you. Ship it, iterate, move on.

But what if perfection isn’t what we think it is?

A sunflower has to pack as many seeds as possible into its flower head. It grows a seed, turns, grows another. If it picks a clean, rational angle - a third, a tenth, a half - it gets neat spokes with big empty gaps. The more orderly the pattern, the more space gets wasted.

But if it picks the most irrational number possible - the one that resists every neat fraction - it gets perfect packing. No gaps, no wasted space. That number is phi - the golden ratio.

Nature’s most perfect design comes from choosing the most irrational input. Not order. Not symmetry. The thing that refuses to fall into a pattern.

Maybe perfection was never about getting everything right. Maybe, like the sunflower, it’s about embracing what works now and iterating it with the intent to make it better.

Speaking of counterintuitive ways to absorb things - here’s a tool that changed how I consume long documents.

🤖 NotebookLM - your documents, but as a podcast

I’ve been thinking a lot about the difference between stickiness and retention lately. They sound like the same thing, but they’re not - stickiness is about how often users come back within a time period, retention is about whether they come back at all. You can have high retention with low stickiness, and that’s a slow leak most builders don’t notice.

I found a long research piece on this, uploaded it to Google’s NotebookLM, and hit “Audio Overview.” Ten minutes later, two AI hosts were breaking it down for me like a podcast episode - debating the nuances, using examples, pulling out the parts that actually mattered. I listened to the whole thing on a morning walk.

The hosts sound unsettlingly human. You can even jump in mid-episode and ask them to elaborate on something. It works with PDFs, YouTube videos, Google Docs, slides - anything you want to absorb but don’t have the time to sit and read.

Try it at notebooklm.google.

Speaking of making things unexpectedly fun - here’s a founder from our community who did exactly that with math.

🎯 Sudhanshu from Matiks

Sudhanshu is a friend from our community and one of the sharpest math minds I know. He’s also the co-founder of Matiks - a mobile gaming platform that started by turning math into a competitive sport.

The idea was simple: instead of making math feel like homework, make it feel like a game you actually want to win. Speed challenges, multiplayer duels, daily contests. You’re not studying math. You’re competing at it.

What happened next is interesting. They started with math, but the same principle - make your brain work harder and enjoy it - naturally expanded into memory games and puzzles. Turns out when you nail the core mechanic of “competition makes practice addictive,” the category almost doesn’t matter.

I’ll admit - I love Matiks! Speed math is my go-to whenever I’m bored. It’s replaced mindless scrolling for me, and 1.3 million other users seem to feel the same way.

Check them out at matiks.in.

And while Sudhanshu is making math feel like play, we spent this week making Shram feel more solid.

🚀 What we shipped this week

We spent this week making the app significantly more reliable. Less glamorous than shipping a cool update, but honestly more important.

We also added undo/redo across Shram. Made a change you didn’t mean to? Ctrl+Z your way back. Small thing, but the kind of small thing that makes you trust a tool enough to move fast in it.

If you want to try it for yourself, here's our beta access!

That’s it for now - I’ll see you next week. Until then, have a good one! :)

Every week, I’ll share a 3-minute read with you - an insight from my learnings, a useful tool that reduces your cognitive load, cool founders building cooler stuff, plus an update from our team at Shram.

📖 The shell vs the core - Actionable Gamification

A few years ago I read Yu-kai Chou’s Actionable Gamification and there’s one concept I keep coming back to.

Most companies trying to make their product more engaging do the same thing: add points, badges, and a leaderboard. Yu-kai calls this the “shell” of a game. It looks like gamification, but it doesn’t work - because slapping a progress bar on a boring experience doesn’t make it less boring.

Real engagement comes from something deeper. Yu-kai maps out 8 Core Drives - things like feeling part of something bigger than yourself, curiosity about the unknown, or the fear of losing something you already have. These are the forces that actually make people care.

As Yu-kai puts it - it would be foolish for a general to say the Greeks sent a wooden horse and won a war, so let’s build wooden horses too. The Trojan Horse worked because of the strategy behind it, not the shape of it. Same with gamification. Same with building products.

Speaking of going deeper than the surface - here’s a way to do that with anything you read.

🤖 Turn any research paper into something you'll actually read

Research papers are packed with insights but almost designed to be hard to read. Dense jargon, 40 pages of text, and by page 12 you’ve forgotten what page 3 said.

Here’s a trick: upload any research paper to Claude and ask it to condense the content while keeping the key insights. Then ask it to build an interactive version - one where you can select any paragraph and get it elaborated, explained with an example, or turned into a visual diagram.

Our teammate Ojasvika did exactly this with a paper called “Augmenting Human Intellect.” She added voice notes that got cleaned up automatically, got flowcharts generated as print-ready SVGs, and turned a paper she’d been meaning to re-read for months into something she absorbed in one sitting.

Watch her walk through it here.

While Ojasvika is rethinking how we read, here’s a founder rethinking what goes inside the products we use every day.

🎯 Shruti from Bacalt

Shruti is one of my closest friends and I’ve watched BacAlt come up from nothing over the last two years.

Here’s the problem she’s solving that most people don’t think about: 87% of the top 10 best-selling consumer products contain microplastics in their formulations. Not in the packaging - in the actual product. Your face wash, your detergent, your shampoo. Everyone talks about recycling and biodegradable packaging, but the stuff inside the bottle is the bigger problem.

Shruti is building bio-based alternatives using fermentation and agricultural waste. Her biopolymers replace the synthetic ingredients without compromising on performance or cost - which is the part that usually kills sustainable alternatives.

Check them out at bacalt.bio.

Solving the hidden problem instead of the obvious one - that’s been the theme this week. Here’s how we’re doing the same at Shram.

🚀 What we shipped this week

You know that thing where you finish a 45-minute call and realize you didn’t write down a single action item? That’s gone now.

Shram can now listen to your meetings and pull out tasks for you. Once the call ends, you get a notification on the bottom right of your screen with all the tasks - no manual note-taking, no forgetting what you promised.

It works with Google Meet, Zoom or any other audio/video app and just needs your system audio permission. That’s it.

If you want to try it for yourself, here's our beta access!

That’s it for now - I’ll see you next week. Until then, have a good one! :)

Every Friday, I’ll share a 3-minute read with you - an insight from my learnings, a useful tool that reduces your cognitive load, cool founders building cooler stuff, plus an update from our team at Shram.

📹 The eye that changed everything - Leaps in Evolution

My favourite docu-series is Leaps in Evolution - I have watched it several times and my learnings from it span beyond just biology (like most of my learnings do :P).

For 3 billion years, life on Earth was blind. Creatures floated, crawled, and fed - navigating entirely by touch, smell, and vibration. No eyes or vision. Just darkness.

Then, about 540 million years ago, a simple patch of light-sensitive skin on a tiny creature evolved into something that could form an image - the first eye.

What followed was the most explosive period of innovation in the history of life. Within just 15 million years, nearly every major animal group we know today appeared. Scientists call it the Cambrian explosion. And many believe vision was the trigger.

The logic is simple: the moment one creature could see, everything changed. Predators could hunt from a distance and prey needed new defences. A single new capability didn’t just help the species that had it - it forced the entire world to evolve around it.

This is exactly what’s happening with AI right now. It isn’t a speed upgrade - it’s a new sense. If you’re building right now, you’re in the middle of a Cambrian explosion. And the founders who “see” that clearly have a real edge.

One place this new sense is showing up? How we talk to our computers.

🤖 Wispr Flow - stop typing, start talking

Think about how much of your day is just... typing. Emails, Slack messages, prompts, docs, DMs. Your thoughts move fast but your fingers are the bottleneck.

Wispr Flow fixes that. It’s a voice dictation tool that works across every app on your computer and phone. Hit a shortcut, speak naturally, and your words appear as polished text - formatted for whatever app you’re in. It strips filler words, fixes grammar in real-time, adapts to context, and supports 100+ languages.

Wispr has raised $81M and is valued at $700M. Voice-first input is clearly becoming a real category. But here’s a question worth asking: should one voice model really work for everyone?

🎯 Mitesh from Utter

Mitesh is building Utter - a voice dictation app that starts familiar but goes somewhere very different.

On the surface, it works like you’d expect. Hold the fn key, speak, release - your text appears wherever your cursor is. Emails, messages, coding prompts, journals. Clean and simple.

But Mitesh’s core belief is that one-size-fits-all voice models are fundamentally limited. The nuances of how each of us speaks - our pauses, tone, sentence patterns, the way we trail off or emphasise certain words - are deeply personal. A general model can get you 80% of the way there. The last 20% requires something built just for you.

So after you’ve spoken about 60,000 words on Utter, it lets you download your own personal voice model, trained entirely on your speech and stored locally on your device without cloud dependency. It’s the most accurate version of your voice that software can understand - and it’s yours to keep.

I have to be honest here - despite having an active Wispr Flow subscription, I've been using Utter for the past 3 weeks and haven't looked back. It is noticeably faster, and you can see your words appearing right on your screen as you speak. Press Shift while holding fn to correct errors before sending - small things that make a big difference in flow.

I love it so much that we’re now partnering with Utter to bring the highest quality speech-to-text to Shram. So every time you talk to Shram, you’re moving one step closer to building your own voice model.

Check out Utter here.

⚡️ Access your memory from WhatsApp

Speaking of Shram getting smarter with your voice - here’s something we shipped that makes your memory smarter too.

Until now, everything Shram remembered about you lived on your device. Private, fast, secure. But also stuck on one device.

This week we added a simple toggle that syncs your on-device memory to the cloud - making it accessible from anywhere, including WhatsApp.

That means when you use Shram on WhatsApp, it already knows your context from your laptop. Your preferences, your projects, your history - all there. No re-explaining yourself every time you switch devices.

Draft a follow-up email to a lead from your phone while you’re in a cab. Get a summary of your week while you’re away from your desk. Your memory follows you.

And if you’d rather keep everything local? The toggle stays off. Your data, your choice.

If you want to try it for yourself, here's our beta access!

That’s it for now - I’ll see you next week. Until then, have a good one! :)

Every week, I’ll share a 3-minute read with you - an insight I learned from my readings, a useful tool that reduces your cognitive load, cool founders building cooler stuff, plus an update from our team at Shram.

📖 The seven emotions you can't fake

Years ago, I came across Paul Ekman’s work in “Unmasking the Face” and it changed how I read people.

Ekman spent decades studying facial expressions across cultures - from urban cities to tribes with almost no contact with the outside world. His conclusion: seven emotions are universal. Surprise, fear, disgust, contempt, happiness, sadness, and anger. Every human on the planet displays them the same way. Cultural norms might change how intensely you show them, but the expressions themselves are hardwired.

Here’s the part that stuck with me: a genuine emotion can only be displayed if it’s actually being felt. A real smile comes only from a real feeling of happiness. You literally cannot produce one while feeling contempt - not if the emotion is true.

I think about this a lot when I’m building. The same way you can't fake a smile, you can't fake craftsmanship. People feel the difference even if they can’t articulate why. Users can tell when something was built with genuine care versus when it was assembled to check boxes.

That same authenticity matters when you show up online. Which brings me to a tool that helps you do exactly that - consistently.

🤖 The best way to write and grow on LinkedIn and X

Every founder knows they should be posting on LinkedIn and X. Very few actually do it consistently. The blank page wins almost every time.

Typefully makes that blank page disappear. It gives you a clean, distraction-free editor built specifically for social content. The AI doesn’t write for you - it makes what you’ve written better. Hit a command and it’ll tighten your draft, suggest hooks, or split a long thought into a thread.

What I like is that you write once and Typefully adapts your post for X, LinkedIn, Threads, and Bluesky - each with the right tone and format for that platform. One idea, multiple audiences, one click.

Check it out at typefully.com.

While Typefully helps founders tell their story, here’s a founder building something that lets D2C brands actually understand theirs.

🎯 Yuvraj Dagur from Clevrr AI

Yuvraj is building Clevrr AI - an AI workforce for D2C brands. If you run a D2C brand, you know the pain: ten Excel sheets every month, Shopify and Meta Ads dashboards pulling you in different directions, no visibility when things go wrong. You end up being the bottleneck connecting every team and every tool.

Clevrr solves this with specialized AI agents - one each for sales, marketing, inventory, shipping, and finance. The sales agent detects revenue drop-offs and recovers abandoned carts. The marketing agent spots underperforming ads before your budget bleeds. The inventory agent predicts stockouts and auto-generates purchase orders.

46 brands already rely on Clevrr to manage over ₹150 crores in revenue. I've seen him work night after night when we used to share the same workspace. What I admire most is how he's thought about this as a team of AI colleagues rather than just another dashboard.

This kind of contextual intelligence - where AI doesn't just see data but understands what to do with it - is exactly the direction we're pushing Shram.

⚡️ Shram now listens while it looks

You’re reading an article. It’s good. You know you’ll forget about it. Normally you’d leave the tab open - joining the 30 others you’ll never get back to.

Now you can just tap, then hold the Option key and say “Remind me to read this later.” Shram captures the link, stores the context, and adds it to your to-do list. Tab closed.

Or say you just opened an email with three or four things you need to do. You know you should add them to your to-do list. You also know you won’t. Now just tap and hold Option - you don’t even need to say anything. Shram reads what’s on your screen, finds the tasks, and adds them for you.

Well, Shram now listens while it looks. Your voice adds intent to what it sees. If you say something, it acts on that. If you don't, it finds tasks by default - just like before.

If you want to try it for yourself, here's our beta access!

That’s it for now - I’ll see you next week. Until then, have a good one! :)

Every week, I’ll share a 3-minute read with you - an insight I learned from my readings, a useful tool that reduces your cognitive load, cool founders building cooler stuff, plus an update from our team at Shram.

📒 ‘The Prince’ on Mercenaries

Machiavelli wrote something in ‘The Prince‘ I find incredibly interesting: “A prince without his own army is no prince at all.”

Mercenary troops are dangerous whether you win or lose. If your mercenary commander is skilled, he will eventually turn on you. If he is unskilled, he will lose your battles. Either way, you are ruined.

If you win with mercenaries, you are at their mercy. If you lose, you are still at their mercy.

The only real power comes from having your own army - people who fight because the land is theirs, not for payment. Machiavelli watched powerful city-states crumble because they relied on hired soldiers, while states with citizen armies lasted for centuries.

I find this extremely relatable - if you have a CTO who doesn’t code, or hired contractors working on the core of your product, you are increasing your chance of failure.

You can’t build anything lasting on borrowed strength.

❤️ Lovable - Build apps without writing code

Last week, I spoke about a tool I built. Today I want to talk about my experience building it with Lovable - one of the most impressive tools I have used.

You describe what you want in plain English - “Build XYZ app” - and Lovable generates the entire thing, not just code snippets. It creates full-stack applications using React, TypeScript, and Supabase. You get a working prototype you can deploy immediately.

I wrote 380 queries to build the app I was working on, spending approximately 12 hours over a few weekends. This cost me 350 credits - approximately $60 total. I am on their Pro plan so I had to buy $20 worth of extra credits when I ran out of mine.

What makes it special is that you actually own the code. Export to GitHub, edit it yourself, or keep building with AI. Think of it as your AI intern - fast, capable, but you’ll still need to review its work. It’s best for MVPs and prototypes, not production-scale apps.

And speaking of production, here are two founders who built an entire platform to help developers with it.

🎯 Kalpesh and Vedant from Scoutflo

Two friends in our beta are solving one of the most painful problems for developers - DevOps automation.

For this, Kalpesh and Vedant built Scoutflo - Your AI SRE. When incidents happen, their AI handles alerts, finds root causes, and fixes issues in minutes. It slashes MTTR, predicts problems before they impact users, and makes engineers 10x faster.

I visited their lovely office in Mumbai last week and I was pretty impressed with what they have built! Unfortunately I am not their core ICP but if you are a developer tired of DevOps being a bottleneck, you should check them out.

This kind of automation - where you just describe what you want and the system handles it - is exactly what we shipped this week.

🚀 What we shipped this week

Shram can now read and write your Gmail, and manage your Google Calendar.

For Gmail, try saying:

• “Email John and let him know I find his quote a bit high”

• “Summarize emails I received today. Exclude promotions”

• “Send the event invite via email separately to Tanmay and Henry”

For Calendar, try saying:

• “Schedule a meeting with Noah tomorrow at 10 AM”

• “I want to meet Noah at 3pm, not 10AM - update the event”

You don't need to open Gmail or Calendar - just tell Shram either through our desktop app or WhatsApp. It reads context from your inbox threads, drafts emails that sound like you, schedules meetings without you touching your calendar.

It’s the magical experience we’ve been building toward - where Shram handles the coordination work and you focus on what actually matters.

Here's our beta access for you to try it out!

That’s it for now - I’ll see you next week. Until then, have a good one! :)

Every week, I’ll share a 3-minute read with you - an insight I learned from my readings, a useful tool that reduces your cognitive load, cool founders building cooler stuff, plus an update from our team at Shram.

⚔️ The Battle of Alesia

In 52 BC, Julius Caesar marched his army into what looked like a death trap.

The Gallic forces had retreated to Alesia, a fortified hilltop city with 80,000 warriors. Caesar had 60,000 men. He could have retreated. Instead, he chose to stay and let himself get surrounded.

A massive relief army of 250,000 Gauls was coming. So Caesar built two rings of fortifications - one facing the city, another facing outward. Twenty miles of walls and trenches.

When the relief army arrived, Caesar’s forces were sandwiched between two enemies. Outnumbered six to one. But the double walls let his troops move quickly between threats while the two Gallic armies couldn’t coordinate. Both armies collapsed.

Caesar’s superiors in Rome were furious - if he had fallen, Rome would have become extremely vulnerable. Yet he made this choice anyway. Sometimes the best options are not so obvious.

Speaking of unconventional solutions - here’s how I fixed a problem that’s bothered me for years.

📖 Learn like Neo from the Matrix

I love reading books but ever since I got into entrepreneurship, I haven’t been able to keep up. Almost every day I regret being unable to read - so I decided to fix it for myself.

I built Read by Shram that allows you to upload articles or books, set a speed you’re comfortable with, and read through without even moving your eyes.

I used to avoid books because I’m a slow reader by habit. But since using this tool, I’ve naturally pushed myself to read faster without loss in comprehension.

I read “The Cold Start Problem” by Andrew Chen in 4 hours yesterday night at a speed of 600 words per minute with brief timely pauses.

Just keep three things in mind when using this tool:

1. Comprehension > Speed - choose a speed that works for you

2. Blink often - don’t rush and give your eyes rest after each chapter

3. Pause if lost - pause and read what you couldn’t understand

(Watch this video to see how the app works)

I made this for personal use so what works for me may not be your cup of tea, but what these two founders built will definitely help you with a more universal challenge.

🎯 Abhinav and Vinayak from SuperJoin

Two founders in our beta are building something remarkable at SuperJoin.

Abhinav Das and Vinayak Jhunjhunwala started by syncing live data from Salesforce, HubSpot, and other tools into Google Sheets. No more CSV exports or manual refreshes.

But while building AI models for financial modelling, they stumbled onto something bigger. They built models that understand spreadsheets at the cell level - where tables sit, what they mean, how they relate, what drives the data.

Their models are already beating everyone on benchmarks - including Claude and Microsoft Copilot.

The last time I tried it, it genuinely felt like magic. I found myself wishing I worked with Excel more often just to use SuperJoin more.

While SuperJoin understands your spreadsheets, we heard your request and shipped something that understands what’s on your screen.

🚀 What we shipped this week

After several iterations over many months, automatic task detection is finally ready for reliable use.

We’ve personified the interaction you’d have with a real executive assistant. Your EA doesn’t constantly interrupt you while you work - their primary job is to maintain your flow. Instead, they check in periodically and tell you what needs to be done.

We’ve simulated this exact experience using the context of your screen:

• You are on Slack - Shram notices action items and suggests tasks without you having to manually create them

• You are reading an email thread - Shram extracts follow-ups and deadlines automatically

The difference is subtle but important: Shram works in the background, understanding context, only surfacing tasks when it’s confident something needs your attention.

Here's our beta access for you to try it out!

That’s it for now - I’ll see you next week. Until then, have a good one! :)

Every week, I’ll share a 3-minute read with you - an insight I learned from my readings, a useful tool that reduces your cognitive load, cool founders building cooler stuff, plus an update from our team at Shram.

🏛️ The Architecture of Happiness

Le Corbusier is probably the most famous modernist architect in the world - he even designed the city of Chandigarh, India. His clean lines, geometric forms, and “machines for living” defined 20th-century architecture.

In 1923, he built housing for factory workers in Pessac, France. Undecorated boxes with bare walls and minimalist design - exactly what modernist architecture called for.

The workers hated it.

They added shutters, painted the walls, built gardens, and put up picket fences - making the houses the complete opposite of what Corbusier envisioned.

The workers spent all day in monotonous factory work. Coming home to equally monotonous buildings felt like more of the same. They needed their homes to feel different.

This reminded me of something I learnt the hard way over time - the best experiences solve a real, deep problem people face, not the problem we think they should have.

Speaking of solving real problems - here’s a simple automation that’s been sitting on your Mac this whole time.

🤖 Automate your repetitive typing

You type the same things 50 times a day. Your email. Your LinkedIn URL. That paragraph explaining what your startup does. Calendar links, and more!

Here’s how Mac’s built-in Text Replacement handles this:

Go to System Settings > Keyboard > Text Replacements.

Then set up shortcuts like I have done for my personal use.

• jay@ → expands to my Email ID

• jay. → expands to my LinkedIn profile

• intro1 → expands to an introduction of what Shram does for new users

• cal2 → expands to my Calendar link for friends

• maps. → expands to our Shram’s office location on Google Maps

The magic: it works everywhere - Gmail, Slack, LinkedIn, iMessage, Notion. No app switching. No copy-pasting. Just muscle memory that saves you hours every week.

No more hunting for links. No more retyping your email address. No more writing the same intro from scratch. Just shortcuts that make repetitive typing disappear.

While we’re talking about automation, here’s a founder who’s automated operations at a completely different scale.

🎯 Shubham from QuicReach

One founder in our beta consistently impresses me with sheer operational excellence: Shubham Jain, the solo-founder of QuicReach.

QuicReach is a shared cab platform for airport and intercity travel. Think Uber, but you’re sharing the ride with verified co-passengers heading the same direction reducing cost of travel significantly.

What’s remarkable isn’t just the idea - it’s the execution. Shubham built QuicReach to 80,000+ users and $300k in ARR with a team of 3-4 people. Completely bootstrapped.

I use QuicReach whenever I need to get to the airport. The quality is excellent - most cabs are EVs, drivers are well-spoken and extremely timely. I remember when we met, how impressed I was with the way he had automated every single detail of his service.

If you’re in Bangalore and tired of paying ₹2000 for airport rides, download QuicReach. If you are building a company, I think its worth watching how Shubham operates - it’s a masterclass in doing more with less.

This kind of curated experience requires being everywhere your users are - which is exactly what we shipped this week.

🚀 What we shipped this week

You can now chat with Shram on WhatsApp.

Just save our number, send a message, and manage your entire to-do list through text:

• “Add call with Abhinav for next Tuesday at 3pm”

• “What do I have scheduled tomorrow?”

• “Mark the deck review as complete”

No app switching. No context switching. Just text the way you normally would.

Here's Shram’s beta access for you so that you may try it out for yourself!

That’s it for now - I’ll see you next week. Until then, have a good one! :)

Every Friday, I’ll share a 3-minute read with you - a useful tool that reduces your cognitive load, an insight I learned from a book I’m reading, cool founders building cooler stuff, plus an update from our team at Shram.

📖 Darwin’s bees - The Origin of Species

Most people know honeybees build perfect hexagonal cells - nature’s most efficient structure. But Darwin discovered bees in Latin America called Melipona that still build round cells. Not hexagons. Just circular, less efficient structures.

These bees have been around for millions of years building round cells. They’ll likely evolve to hexagons eventually - but that process takes time.

Here’s what struck me: evolution happens in steps, not leaps. There’s a messy middle where things work but aren’t optimal yet.

You see the polished version your competitor shipped and feel behind. But they went through the round cell phase too. Being in the middle of evolution is normal. You’re supposed to be imperfect right now.

Speaking of evolution and messy middles - let's talk about a real productivity mess we've all experienced.

🤖 The browser extension that organises your tab chaos

47 tabs open. Can’t find anything. Every founder I know lives like this.

We built a Chrome extension that fixes it: Organize by Shram automatically groups your tabs using on-device AI. Research tabs get grouped together. Product work gets its own space. Personal stuff stays separate.

The magic? It’s completely local - no data leaves your computer.

Download it here.

While we're solving problems with AI, here are two founders doing something equally practical with it.

🎯 Bhaskar and Rajat from Monova

Bhaskar and Rajat are building Monova - an AI styling platform that helps you optimize your existing wardrobe.

Bhaskar spent four years at Amazon on Alexa. Rajat is his co-founder and COO. They’re not trying to sell you more clothes. They’re helping you optimise what you already have - figuring out what works and how to stop buying things that sit unused.

I love how personalised they have made the onboarding and how well they have evolved the experience, starting with WhatsApp and now moving toward a completely native app experience on phone. I am sure you will find their fashion advice useful.

Check them out here.

This kind of personalisation requires context that evolves - which brings me to what we’ve been working on.

🧠 We built perfect memory

Most AI assistants use embeddings designed for search. So when you say “I’m vegetarian” in January, then order chicken in March, they see two food-related statements and think they’re compatible. But your context evolved.

We rebuilt our memory layer to preserve context as it evolves over time, not just store conversation snapshots.

We ran LongMemEval-S (the industry benchmark) and scored 100% on single-session performance, beating Supermemory, Zep, and full-context approaches. This is what lets Shram understand what you meant three weeks ago versus what you mean today.

You can read more here.

All of this comes together in what we shipped this week.

🚀 What we shipped this week

You can now chat with your to-do list. We also built a Google Calendar integration that allows you to schedule or update meetings on your calendar via Shram. Just say…

• “Delete all tasks related to the old landing page”

• “Show me everything due this week”

• “Move the investor call to next Tuesday”

…and Shram will take it from there - just like your EA would.

Before you go, here’s a short video by our design lead Ojasvika explaining how you can use Shram now :)

That’s it for this week - I’ll see you this Friday. Until then, have a good week! :)