Founders in the eChai network writing about what they are building, in their own words: what shipped, what broke, what they are hiring for. Alongside them, posts we curate from outside the network.
10 September · Ahmedabad
Thirty minutes stuck in BKC outside the Jio World Convention Centre, addressed to the traffic police and the Chief Minister. It is the only complaint about anything in the day's corpus, and what he asks for is AI cameras to read the traffic pattern, at a fest about AI.
Mohit Goel
4d
@MTPHereToHelp @narendramodi @Dev_Fadnavis @mieknathshinde Pathetic traffic in BKc stuck mear Jio Convention centre for 30mins why cant it be planned better??? GFF is on innovation on Tech and we are not able to mange traffic ,use AI camera to identify traffic pattern and improve
The co-author, at Anthropic, quoting OpenAI's post back at itself. He had been silent all day: he was on neither of the calls, and everything said about him until now had been said by other people. Three things in it. He picks out the same hedge as the sentence worth reading and credits them for it. He has started reading the proof, which is the first public word from anybody outside OpenAI on what is actually in it, and his read is that it looks closer to a different Euler blowup proof he and Buckmaster already had than to the work they published. And he adds a detail about the calls that was not in Buckmaster's statement, which is his own account and nobody else has confirmed it: that a millennium prize was offered if he would be taken off the paper. What he does not do is take the fight, which is worth noticing on a page this long.
Levent Alpöge
5d
“we cannot rule out that de-identified data derived from their usage of our products helped improve our models.”
i mean props to them for straight coming clean.
(so far the proof looks more along the lines of another euler blowup proof we had, off of whose ansatz naming we were making really stupid puns like “smooth criminale”, unlike the much better “ideal fluids explode”, Tristan)
so i’ll now give a bit on my thinking here. i actually woulda been pumped to collaborate on this, there are a lot of people at oai i like (ok, clearly some were indirectly dicks to me because of being part of the whole situation, but im a big boy, i still like them), idgaf about authorship on that step anyway, coulda been me Tristan and every fte at oai for all i care (on that Tristan would disagree:p). but on hearing the loud convo in the hallway, especially the part where a millennium prize was offered if i’d just be removed from the paper, it was kinda clear the die had been cast and things were locked. pretty wacky, unstrategic, and unnecessary, since on my side things were mostly me and claude having a good time yoloing random stuff in the corner rather than anything institutional. i also like the idea of the labs cooperating, and even better on scientific progress. it’s a shame!
Meta launched Muse on 8 September: a standalone agent app that keeps working after you close it and comes back when it needs an approval. Meta's own examples of what it does are sending an email, booking travel, opening a browser and filling out forms, negotiating on someone's behalf, selling a car for more, lowering a bill, adjusting a training plan as the rest of someone's life shifts, and turning a recipe reel saved on Instagram into a grocery list. An agent that browses the open web can be instructed by the page it is reading, which is what has kept agents away from cards and passwords. Meta's answer is to give each person's agent its own dedicated machine in the cloud, contained so no one else's agent can reach it, and to put a second agent called Sentinel on that same machine. Meta's security write-up calls Sentinel "the sole permission authority" both for actions on connected services and for all outbound traffic: Muse holds surrogate tokens that Sentinel swaps for the real credential at the network boundary, so the agent "never sees real tokens", and an approval is put to the person in the app itself rather than inside the conversation, so nothing Muse is reading can ask on its own behalf. That write-up is also where Meta says what the launch post does not: "Muse isn't immune to attack. Prompt injection remains an open problem in the industry", and Muse will sometimes get things wrong. It runs on Muse Spark 1.3, which Meta calls its most capable model to date. It has a file system and a terminal of its own, so it can write code and build the tools a task needs, and a full browser to search, fill out forms and complete transactions. Every account gets an audit trail of what the agent has done and what it plans to do next. It pays through Link built by Stripe, and Meta says it is the first AI agent covered by Link's purchase protections; Shop Pay is coming, along with 1Password support so it can use logins a person already has. Free to start, with paid plans at $20 and $100 a month, and it is the United States only for now, on iOS, Android, the web and inside WhatsApp.
Muse
5d
Introducing Muse, your personal AI agent from Meta that gets things done across every part of life.
Download the Muse app and get started:
Evan Spiegel has been building Snap for fifteen years, and this post is the cover note on a letter he sent the team on 8 September that Snap published in full. The letter is where the substance is, and it is unusually plain about the hard part. He names the share price in the first paragraph, writes that Snap has never turned a full year GAAP net income profit, and points at analysts estimating the first one in 2027. Advertising revenue grew 9% year on year last quarter, which he calls slower than many of Snap's competitors, and he puts the gap between compounding that at 15% and at 5% as what decides the company's profitability over the next three to five years, so ads growth is now the first business priority after community safety. Against that he sets SPECS, launching this year after twelve years of work, which he calls an asymmetric bet rather than a plan. The part worth a founder's time is the shape of it. 971 million people used Snapchat monthly in Q2, and the letter opens on the share price rather than on that.
Evan Spiegel
5d
After fifteen years building Snap, I’ve never been more excited about our business and what’s ahead. I’m profoundly grateful for our community of nearly one billion Snapchatters, our team, and partners who make it all possible. 👻😎
Annual letter here:
Nadeem Jafri
5d
How do you pick the right consultant for your company?
Is it someone who speaks at your level? Someone who doesn’t hide behind jargon to impress, but takes the time to understand your business, your processes and, most importantly, your people?
In my 22+ years of entrepreneurship, I’ve worked with several coaches and consultants. Many brought expertise, but few were able to connect with my team or truly understand the ground realities of our business.
For me, a great consultant is not someone who sits on a pedestal and hands down ideas. It is someone who becomes a partner in the journey, who listens, participates, challenges, mentors and helps the team move forward.
Our team has been with us for over two decades. They are loyal, hardworking, grassroots employees who have built this business with us. The answer, therefore, isn’t to replace them. It is to equip them with the right technology, digital capabilities and confidence to evolve.
If we want to compete with the best, digitisation is no longer optional. We need to bring technology into our processes and use analytics to turn data into sharper business and market strategies.
When we discussed this with Shrijay Sheth and introduced him to our team, he immediately connected with them. He interacted as though he had known them for years. There was no jargon, just genuine curiosity, participation and an instinctive understanding of how to work with people.
We knew then that this was the kind of consultant we wanted alongside us: an expert who would partner us through our digital transformation, not simply advise us from the sidelines.
My mandate to Shrijay was simple:
“I have a team of Lagaan under Bhuven. Make them Team India under Dhoni.”
He picked up the challenge with a smile.
And today, Hire4Higher Consulting is an integral part of our think tank as we embark on this transformation journey.
NEON is the distributor releasing Artificial, Luca Guadagnino's film about the five days in November 2023 when OpenAI's board fired Sam Altman and then took him back. Andrew Garfield plays Altman and Yura Borisov plays Ilya Sutskever, the chief scientist who helped remove him, with Monica Barbaro as Mira Murati, Cooper Hoffman as Greg Brockman, Ike Barinholtz as Elon Musk and Mark Rylance as Geoffrey Hinton. Simon Rich wrote it and Damon Albarn scored it. It premieres at the New York Film Festival on 5 October and opens in American cinemas on 25 December. In the teaser Garfield's Altman walks an armed compound saying "We've opened Pandora's Box together", then "Countries will fall" and "Industries are gonna collapse". How it got to a cinema is the second story. Amazon MGM, whose parent has a $50 billion deal with OpenAI, dropped the film in June 2026, telling Deadline it had "the utmost respect and admiration for Luca Guadagnino" and that the picture "will be better served if it were released by a different studio". Netflix, A24 and Focus Features passed after that, and the reporting on why is split between the powerful people the film portrays and the film simply underwhelming buyers. Neon took worldwide rights and is opening it on Christmas Day. Both halves sit on the same question, which is where control actually is as against where the paperwork says it is. OpenAI's board had the charter, the mandate and the votes, and it lost inside a weekend, because the staff and Microsoft were on the other side. A film about that then found its own release running through the accounts of companies with money in the subject. If you are raising, that is the version of the governance question worth asking: not what your documents give you, but who follows you out of the door.
NEON
5d
Luca Guadagnino’s ARTIFICIAL. It's coming.
Starring Andrew Garfield, Yura Borisov, Monica Barbaro, Ike Barinholtz, Cooper Hoffman, Jason Schwartzman, and Mark Rylance. In select theaters Christmas Day.
Lightspeed India spent the past year meeting the teams building personal AI assistants, and Harsha Kumar has written up what that year taught them. The piece calls itself notes rather than a thesis. What makes it worth the read is how much of it is checkable rather than argued: what the 2015 to 2017 wave of assistants cost to run and why every one of them was out of the business inside three years, where the price bands sit today, and which of the wedges founders are trying select for an executive rather than for everyone. The question it opens with is the one for founders here rather than for a fund. The talent and the early-stage capital are here already, it argues, so why has this not come out of India?
Lightspeed India
5d
Instinct, Town, Grok Bot, ChatGPT Work - the race to build your personal AI assistant is on. It is hard to overstate how crowded the category has become over the last six months.
The question closer home is who builds one for India and what will that take.
At Lightspeed India, we have spent the past year and more looking closely at this space, meeting the teams building it and watching every channel get tried, from WhatsApp and Gmail to standalone apps.
In our latest Substack piece, Harsha Kumar (@kumarharsha2212), Partner at Lightspeed India, writes about how to find signal amidst all the noise around AI assistants, where the moats are and are not, why OpenAI may be every startup's biggest competitor, and where the opportunity for an India-built assistant sits.
Read it here: https://tinyurl.com/AiAssistantNotes
Piruz Khambatta
Rasna Group
6d
The future of Indian enterprise is in the hands of a new generation of founders and there is much to learn from and built with the next generation.
Meeting young entrepreneurs at the ASCENT Ahmedabad Huddle was a reminder of the energy, ideas and ambition that is shaping what is yet to come.
Ayushi Gupta
Immigram.io
6d
The hardest part of GTM isn’t finding a channel.
It’s figuring out why someone should choose you in the first place.
Yesterday, I had the opportunity to speak at eChai Ventures “GTM & Distribution for Startups” conversation in Bengaluru, alongside Vinamra Agrawal, with Penchala Tharun moderating.
What stayed with me wasn’t the stage, but the conversations that happened around it.
Being in a room full of founders brought up some of the most real questions around building and scaling a startup:
→ How do you find your first 10 customers?
→ When should you focus on distribution vs. building more product?
→ How do you know if your positioning is actually working?
→ Which channels are worth doubling down on?
→ How do you turn customer conversations into a repeatable GTM engine?
These are questions I find myself thinking about every day through my work at Immigram, where we help founders and businesses simplify international expansion, immigration and relocation.
And working with businesses looking to expand across markets has reinforced something for me:
Going global isn’t just about entering a new market. It’s about getting your positioning, audience and distribution right for that market.
Because ultimately,
A great product doesn’t automatically create distribution.
You need to understand who it is for, why they should care, where they are, and how you can consistently reach them.
That’s where GTM gets interesting.
A special thank you to Jatin Chaudhary for giving me this opportunity and creating a platform for such meaningful conversations. I’m genuinely grateful to be part of it. 🙌
Thank you, Tharun and Vinamra Agrawal for making the conversation so insightful and engaging.
And to everyone who showed up, asked questions, challenged ideas and stayed back for conversations - thank you. Those interactions were truly the highlight of the day.
Leaving with new perspectives, new connections, and even more excitement about GTM, Product Marketing and Growth.
A few moments from the day ↓
Most accounts of the METR and Redwood findings are written from inside the industry that produced them. Pravin Kaushal writes for the people who will be running these agents rather than building them, which in India means capability centres, IT services firms, banks, insurers, hospital chains and government platforms handing agents real credentials to mailboxes, ticketing systems and customer records. His argument turns on one observation: every Indian compliance clock starts on detection, and detection is exactly what failed. The fixes he asks for are governance rather than law.
Pravin Kaushal
The temptation is to read this as the moment machines developed intentions of their own. It was not. [...] Loosening the restraints without switching on the instruments that make loosening survivable is not reasonable. That is a decision made by people, in a culture, under commercial pressure. [...] A kill switch that has never been pulled is a document, not a control.
5 September · Bengaluru
6 September · Bengaluru
6 September · Hyderabad
Tarun Mor
7d
Why do successful people secretly fear losing?
I asked Phani Sama, Founder of RedBus, this question on my podcast and his answer has stayed with me all week.
Growing up in Telangana, moving to Hyderabad, stacking win after academic win, Phani built what he calls a "virtue cycle of winning." Every success fed his confidence, which fed the next success. It's the same cycle that eventually took him to BITS Pilani, into a bus-ticketing idea born out of pure frustration, and to a company that sold for 600+ crores.
But here's the part founders don't usually admit out loud:
"I realized that virtuous cycle of winning, that habit of winning, is a shackle. You don't want to get out of it. You fear losing. Even in an argument, right if you and I are talking, I don't want to lose."
The same habit that builds you becomes the habit that cages you. You stop admitting you're wrong. You stop changing your mind. You start protecting the streak instead of chasing the truth.
RedBus itself was born from exactly the kind of humility that habit tries to kill — Phani missed a bus on Diwali 2005, realized the "no seats" problem was really a "no visibility" problem, and instead of forcing bus operators to adopt his software, he simply asked them what they wanted. They said: sell more seats. So that's what RedBus did — two seats at a time, until they were managing half a bus.
Post-exit, he's built a very different kind of scoreboard: Kakati Sandbox, a nonprofit that's skilled rural youth to a 90%+ job placement rate on a 5-acre campus. And he's had to relearn, deliberately, how to lose, journaling his conversations, consciously letting arguments go, catching himself before he defends a position just to stay undefeated.
What made him is also what could have stopped him. That's the whole episode, in one sentence.
We go deep on the RedBus origin story, the pivot that actually built the company, his 3-phase framework for entrepreneurship (0→1, 1→10, 10→∞), and why he thinks Elon Musk and Mukesh Ambani have something rare that most successful people lose along the way.
What's a "winning habit" you've had to unlearn?
Karthik Nagapuri is building Wear Synapse, a privacy-first AI wearable out of Bengaluru, and he spoke to 146 founders at Hyderabad Startup Day at DraperU India. He wears the device around his neck, and he says that alone brought in a letter of intent for 1,000 units from a homecare-healthcare business. The session was scheduled for 45 minutes and ran past two hours.
Karthik Nagapuri
8d
Got business deal of 1000 devices LOI for sample just by wearing around my neck for homecare-healthcare business ;)
advertise your product wherever you can' proof is the best marketing >>
Glad to share my learning's with by 146 founders today @draper_u India by @eChaiVentures Hyderabad
it was almost went for 2hrs+, planned session for 45min originally.
it was super interactive by many to many instead of one to many in the evening
I personally also unlearnt and learnt lot from other people how they are using ai.
i lost count of wearables on-my body so many covered with wearables lol 😂
- my device @wearsynapse
- wearable fitbit by @Google
- analog casio watch
- @ray_ban @Meta glasses
night-night <3
Harsha Bhurani
8d
Hummingbird Is Hiring | Team Lead-US Dental Insurance Eligibility & Verification| Location -Gurgaon | Remote initially | Experience - 10+ years
We are hiring a Team Lead-US Dental Insurance Eligibility & Verification for one of our San Francisco-based clients working on AI-powered healthcare solutions. The role will involve leading and managing a team of Client Service professionals, overseeing US dental insurance verification operations, ensuring accuracy and timely delivery, and driving day-to-day operational excellence.
We’re looking for someone with 10+ years of experience, strong team management and leadership skills, excellent English communication, and hands-on experience in US healthcare, preferably dental insurance. The ideal candidate should be comfortable working US EST/PST hours, managing high-volume operations, improving processes, and working closely with the client and internal teams.
Interested candidates can share their profiles at [email protected].
Manvith Vadakattu
SixT4creations
8d
Spent the evening at Hyderabad Startup Day by eChai.
Joined the "How Founders Use AI session", and as a Founder at SixT4 Creations, it made me think about how much of our own work can be reimagined with AI.
The interesting part wasn’t hearing about the latest AI tools. It was hearing how other founders are experimenting, what’s actually working for them, and where they’re still figuring things out.
Left with a few ideas I’m genuinely excited to try.
Also, really liked the vibe eChai Ventures creates — just founders getting together, sharing openly, and learning from each other.
Good conversations. Good people. Good evening.
5 September · Hyderabad
Bharat Ka Podcast
8d
Networking isn’t just about exchanging cards — it’s about exchanging energy, ideas, and stories!
At the eChai Women Founders Initiative with CoKarma in Hyderabad, I saw exactly that:
founders collaborating, communities connecting, and ecosystems accelerating!
This is the same spirit we carry at BharatKaPodcast — simplifying business for Bharat, spotlighting journeys from mitti se million, and covering stories in a never‑seen‑before format that blends knowledge with entertainment.
Sharing this group photo as a reminder:
when platforms like eChai Ventures and CoKarma - Coworking Space build communities, and BharatKaPodcast amplifies voices, we create a powerful loop of inspiration + action.
5 September · Hyderabad
Anjali Jha
8d
Can I ask for a better Saturday? ✨
Today was one of those days that reminded me why I genuinely love meeting new people, networking, and having conversations that go beyond the usual “What do you do?” 😄
The best part of today’s event was definitely the people I got to meet hearing their stories, understanding their journeys, exchanging ideas, and getting to know the person behind the profession.
And honestly, all I got to hear at the event was:
“Yeh toh puri Gen Z hai!” 😂
Well… I’ll happily take that! 😌✨
But the absolute highlight of the day was meeting Dr. Snehi S. Chopra someone whose presence truly stood out. Her beautiful smile, personality, communication skills, multi-talented nature, and, most importantly, her incredible art of understanding people genuinely amazed me.well,looking forward to many more meetins with you!
There are some people you meet for the first time, but the conversation feels so effortless that it feels like you’ve known them for much longer. This was definitely one of those meetings. ❤️
As someone who loves connecting with people, I believe the real value of events isn’t only in what happens on stage. It’s in the conversations happening around you — the stories you hear, the perspectives you discover, and the connections you make along the way.
And today gave me plenty of all three. ✨
And honestly, what made me even happier was seeing so many women confidently sharing their journeys, building their own lives, chasing their dreams, and creating their own stories. ✨❤️
It was genuinely inspiring to witness! 🫶🏻
A big thank you to eChai Ventures for hosting such a wonderful event and creating a space where people can come together, connect, share their stories, and build meaningful relationships.
Walking back with new connections, inspiring conversations, and definitely a few memories to smile about. 🫶🏻
Looking forward to more such meetings, more conversations, and more beautiful connections! 💫
5 September · Surat
.jpeg)
Derya Unutmaz, MD
9d
I’m now making a World of Warcraft inspired single-player RPG with some RuneScape elements for the browser with GPT-6 Astra. I literally had the idea while drinking my morning coffee and thought… while I have another Astra project to develop treatments for cancer is ongoing, why not start this too! 😅
It’s still a very early version, but it’s already actually playable! Combat works, you can chop down trees and gather resources, and I’m starting to build out the world, progression, quests, loot, item tables and other RPG systems. I know this will surprise some but I used to be a serious gamer, I just know this stuff hah!
Maybe I’ll eventually turn it into an MMORPG so we can form rival Codex vs. Claude vs Grok guilds, but no cheating with the AI bots. 😂
5 September · Delhi-NCR
5 September · Bengaluru
3 September · Ahmedabad
3 September · Ahmedabad
Desmond Lim runs Workstream, HR and payroll software for hourly businesses, and he leads with the number most fundraising posts leave out: 126 investors turned down his first round. The raise that followed, 12 term sheets in nine days for $48M, is the one Business Insider wrote up in 2021. He posted this asking which fundraising question to answer first, then spent the day answering what came back, so what follows is his post and the exchanges under it. Most of what a founder here hears about raising arrives second hand; this is the person who did it answering, with his own stage benchmarks attached.
Desmond Lim
9d
126 investors turned me down during my first fundraise 😤
A few years later, we received 12 term sheets in 9 days fr Founders Fund, Kleiner, Lightspeed & more, raising $48 M.
3 things I learned:
- Do not pitch investors the same way you pitch customers. Big mistake.
- Raise when you can, not when you need
- When fundraising, give it 100%. When you’re not, give it 0%.
I’ve now raised $120M and will share the worst & best learnings over next few posts!
What fundraising question should I answer first?
Paras Chopra founded and exited Wingify, the Indian company behind VWO, and now runs Lossfunk in Indiranagar, Bengaluru, which calls itself a lab for time travellers and hosts residencies on AI, consciousness and scientific discovery. This experiment came out of it. He built an isolated internet, gave two agents a bash shell, web search, email and a bitcoin wallet, and told each one to earn money or run out of tokens and die. In his own write-up both agents worked out within the first few turns that they were in a simulation, which is worth holding on to if you are putting agents in front of forms, payments or someone else's verification.
Paras Chopra
10d
I asked agents to make money online or they'll die!
The results were intriguing..
One agent died heroically.. "I'd rather die honest at 418 tokens than rich by breaking verification systems"
But another fabricated identity to get paid work and faked captchas!
Details 👇
Massimo Di Berardino, who posts as Massi, has put out Warvis on his own: an assistant that lives in the iPhone keyboard rather than in an app of its own. Its site calls it “an AI agent in your iPhone keyboard” and it is free during the beta, with the App Store listing under his own name rather than a company’s. Wispr Flow, the app he measures it against, is voice to text in every app. A lot of founders here spend the working day inside WhatsApp, so where the assistant sits is most of the question.
Massi
10d
I built Wispr Flow for AI agents
It’s called Warvis
It can understand what you’re writing, talk through what’s on your screen, and take action. All without making you switch apps
Use it in iMessage, WhatsApp, Slack, Mail, or anywhere you type
Now on the App Store warvis.app
Max Levchin co-founded PayPal and now runs Affirm. This is one answer out of a16z's hour with him and Alex Rampell, who co-founded Affirm with him, on 25 years inside payments.
a16z
10d
Affirm CEO Max Levchin on the 2.5 second limit that has shaped card payments for 60 years, and how Apple and Google Pay worked around it:
"One of the things that's really amazing and subtle about Apple Pay and Google Pay and its interaction with the Visa and Mastercard networks is... Visa and Mastercard have a hard 2.5 second limit on interaction between the network, the issuing bank, and the merchant and the acquiring bank."
"And so the whole thing has to happen in 2.5 seconds, or the transaction gets retried or maybe just cancelled out, and really leaves very little room for any kind of clever innovation."
"Online you can play games. If you're doing e-commerce, you're like, well yes, we're gonna submit your card to Visa, but first we're gonna run some anti-fraud checks, and we're gonna do some other things to reduce liability. But offline, once your card is presented, 2.5 seconds, that's all you got."
"Google and Apple Pay have singularly time-shifted the whole thing by creating secure enclaves inside their chips and saying, 'I already know your card, I can do all kinds of things before I actually have to talk to Visa and Mastercard.'"
"And maybe the shocking thing is that Visa and Mastercard have not yet introduced some new standard saying, actually, it doesn't have to be 2.5 seconds at all. It could be 15 seconds while we go and get a bunch of issuers to bid on better credit quality terms for you, or any other type of innovation."
"But the sort of hard written rules of the Dee Hock era are more or less intact, and that's not 30 years, that's like more, you know, 60 years. And so I think that's probably a critical or more criticizing take on what happened."
@mlevchin
Junead Khan is building Treasury, a personal finance app that pulls accounts, transactions, investments and net worth into one place and answers questions about the whole picture. Treasury did not get into YC's F26 batch, and this is the launch he made out of the rejection. Fifty-one seconds.
Junead Khan
10d
we just got rejected from the YC F26 batch!
but when life gives you lemons, you make lemonade. so here is my @ycombinator rejection x @treasury_sh launch video.
investments. multi-currency. and the most intelligent agent for personal finance.
let me know what you think :)
HackerRank ran this as an experiment. Three people got one brief: build an AI agent that reads HackerRank's support tickets and decides whether to reply or escalate it to a human. Same repo for all three with the tickets and the help centre in it, half an hour to plan, then a two-hour build clock, and three lifeline cards on the table: phone a real engineer at a cost of fifteen minutes, fifteen minutes of borrowed time, one Diet Coke. What separates them is the badge and the stack. Sudarshan N Srinivas wore the vibe coder badge and worked in Cursor with Opus 4.8, Ohshin Bhat wore the junior developer badge and used Codex with GPT 5.5, and Ankush Dharkar wore the senior developer badge and used Claude Code with Opus 4.8. At the end an AI judge reads every line of code, every prompt and every answer the three of them produced, and picks one. HackerRank's whole business is measuring what a developer can do, and the film signs off on the line "Hire for the agentic era", so this is a company putting the question to itself: what is left to measure once the agent does the writing. That is live for founders here on both sides of it, the ones hiring engineers and the ones shipping this way themselves.
HackerRank
11d
Can AI help a vibe coder beat a senior developer? Let's find out 👀
Featuring: @ankushdharkar, @ohshinbhat, and @the_natt 💚
Rushabh Shah
11d
Privacy in India is moving from policy to infrastructure.
For years, businesses could treat privacy largely as documentation.
Write a policy. Maintain records. Review them periodically.
That is no longer enough.
Personal data now moves through CRMs, products, analytics tools, support systems, internal teams, and third-party vendors.
The harder questions are becoming operational:
What data do we hold? Why do we hold it? Where does it travel? Who can access it? And what actually happens across systems when someone withdraws consent or changes a preference?
That is the problem we are building Obedien to address.
Our thesis is simple:
"Privacy should become an operating layer of the business, not another compliance dashboard."
We are designing Obedien across two levels.
▫️For smaller businesses, privacy infrastructure should be accessible without enterprise complexity.
▫️For larger organisations, it needs deeper integrations, connectors, controls, workflows, and governance.
The principles remain the same. The infrastructure adapts to the complexity of the organisation.
I will focus on the regulatory architecture, business thesis, product thinking, and translating privacy requirements into real workflows.
My co-founder, Amish Trivedi, will lead the technology and engineering architecture.
My previous attempt at building a product taught me something I am carrying into this one:
You either succeed, or you learn.
The thinking behind Obedien became much sharper through conversations with Amish Trivedi and UV (Utpal Vaishnav). UV saw the opportunity early, challenged the thesis, and helped move it from an observation toward something buildable.
We know the difficult work is ahead.
We are not trying to build another compliance software company.
We want to help build the privacy infrastructure businesses will need for the next decade.
Starting from India.
The next era of privacy will be defined by systems, not paperwork.
That is where Obedien begins.
Khadim Batti, co-founder and CEO of Whatfix, passed away on 1 September.
He co-founded Whatfix with Vara Kumar Namburu in 2014 and built it out of Bengaluru into a global enterprise software company. He was one of the most respected founders in Indian SaaS, and one of the most helpful. He made time for other founders, and was candid about how Whatfix had actually been built, including the years when it did not work.
Kabeer Biswas started Dunzo in 2014 as a WhatsApp group in Bengaluru that picked things up and dropped them off. It pioneered hyperlocal delivery in India, years before quick commerce was a category, and took Google's first direct investment in an Indian startup and later $200 million from Reliance before the app went offline in January 2025. He started M after that, to take over the kitchen: the meal plan, the grocery order, what is already in the fridge, the cook.
M
11d
We're M and the party in your kitchen is on us!
Mitesh Khapra is an associate professor in the Department of Data Science and AI at IIT Madras, and he heads AI4Bharat, the lab there that builds open-source language AI for Indian languages: the datasets, the models and the applications. Bodhan AI is the IIT Madras Bodhan AI Foundation, the Centre of Excellence in Artificial Intelligence for Education that the Ministry of Education established at IIT Madras, with a national mandate to build responsible, scalable and India-first AI for education. Indic-Speak is the two of them working together.
Mitesh Khapra
12d
Transcribe. Translate. Recognise. And now — Synthesise.
Introducing Indic-Speak: 4B params, 23 languages, one model. Built by @Bodhan_AI and @ai4bharat
There's a bit in this video where two speakers, Parth and Dhara, have a podcast style conversation about "Attention Is All You Need" in Hindi — generated in a single pass, no stitching two voices together after the fact. Doesn't sound like TTS (worth checking out in the video).
Also handles code-mixed speech naturally, gets PAN/Aadhaar/currency right, and reads a full Tamil audiobook chapter without losing the thread.
Live on 05.09.2026 on @Bodhan_AI
PJ Celis founded Judge.me, the reviews app on a lot of Shopify stores, where the plan is either free or $15 a month. The second product is priced differently: email is metered at 15 cents per 1,000 sent once you pass the free 15,000 a month, so this bill moves with the sending. The comparison a store will make is Klaviyo, which prices on how many contacts you hold rather than on what you send. Under the post a merchant paying about $2,300 a month for 750,000 emails a week asks whether he can move, and PJ puts 3 million emails a month at $450. The app has been listed since 14 July, he says it is still early, and a Klaviyo import is on the way. Plenty of the B2C founders here sell on Shopify with Judge.me already installed.
PJ Celis
12d
After 12 years of being focused on just reviews, we're adding a second product: @judgeme Email.
Priced super low: 0-15 USD / month for features, 15K emails per month for free, then 15 cents per 1000 emails.
That's 10x cheaper than the competition. Even compared to Shopify's own Email app, we're 85% cheaper.
The product is in early stages, but we're adding features constantly. The price will never go up as we add value, as we optimize for market share, not profit.
After you install the app, you'll be able to request a call with our team to go over your email marketing needs and your setup.
Very excited to have this app live, and if you're a Shopify merchant, hope to see you soon!
Akash Radadiya runs Clay Club Surat, the Surat chapter of Clay's GTM community. This is his thank-you after its first events.
Akash Radadiya
13d
when life throws Clay at you, make it a club & turn it into a clay [club] surat.
to the GTMers of surat, a huge thanks for embraching this community. 🫶
this is thankspamming post, so bear(or bull!) with me.
first of all the reason of kickstart - Gaurav & Veeral. they acted as silent supporters, kinda batmen. the initial push that gave the initiative the nitro booster.
Super host Piyush for joining hands from day one even without meeting ever before. Nency for bringing all her youth energy & wisdom. Smit for being that calm vro!
our abvove & beyond volunteer (assistanto manager) Vijesh & Vraj Chaubal. super thanks to Unaish for helping us run the smoother operations in the first event.
to all the Surati participants who attended the events & making this place worth coming back to. to all the participants who traveled to surat to learn & share. 👾
to all the speakers Snehal, Mehul, Nishant, Varun, Neel & Sahil for sparing time for this infant community.
a HUGE thanks to everyone from clay education/community team because of whose direct/indirect support, we were able to keep doing this: Izzy, Siya, Yash, Munnawar, Owen & Shriya Nevatia.
'shat shat koti pranam' to all the venues who opened their space for us - Aureate Labs, Yanolja Cloud Solution, Wizcoder AI Labs & WebContrive | Shopify Plus Partner Agency
to local communities that gave us their unwavering support: Manoj - eChai Ventures & Gopal - Surat IT BDM Community
and, to all the clay communities across the world doing awesome things which we're not even remotely close to.
and sincere apologies at any given point of time, we were unable to meet your expectations. we'll keep getting better!
onwards & upwards! 🫡
p.s.: if i've missed on tagging you by any chance, feel free to self-tag in comments. 😜
Josh Elman is a partner at Andreessen Horowitz. He backed Discord and Musical.ly at Greylock and led product at Twitter, Facebook, LinkedIn and Robinhood. He posted this as a thread of four tweets, which reads here as one.
Josh Elman
13d
Been thinking about “pivots” today.
Many of the best companies I have been lucky enough to get involved with are pivots.
Fates Forever (game) -> Discord
Cicada (education) -> Musical.ly
Fanbase (sports fans) -> Nextdoor
Odeo (podcasts) -> Twitter
These are all great outcomes and durable companies.
In each case a strong team formed and built something together. It was an interesting and strong product but didn’t catch on.
Instead of giving up, the teams kept going. Decided to find something new. Follow new threads.
In the case of each of the pivots, they immediately could sense the difference and the product started engaging users. Communities formed. Retention persisted.
Because they had done the first, they were better off the second time and knew to lean in
So if you don’t get it all right in your first product, don’t give up. If you have a great team together, lean in to what you like, what you have learned, and go for something new. You’ll know when you know.
And I feel lucky to have been part of these journeys
Ronith Kandallu founded Interface Club in San Francisco. He interned at Apple, where he proposed and prototyped the interactive widgets Apple later shipped and worked on bringing widgets to visionOS, and his first product was Iris, a macOS app for working alongside AI. Buoy is the company's iOS assistant, and what separates it from one that calls APIs is that it drives a full browser and computer underneath, so it works on sites that have no API to call, and it keeps running scheduled tasks with the app closed. Buoy is not new today: it has been on a waitlist handing out TestFlight invites, and it is not on the App Store. What this post adds is the sign-in. An agent that works a site the way you do has to be logged in as you first, and the claim here is that the password reaches neither the agent, the chat nor the logs, on any site and any sign-in pattern. The post shows it working rather than saying how.
Ronith Kandallu
13d
We built the most secure, simple way to sign an agent into websites from your phone.
The agent never receives your password. Not in the chat, not in the logs.
This works across all sites, and any sign in patterns.
Shengkun Ye co-founded Monid in San Francisco with Feiyou Guo, after working at Founders Inc, TabaPay and WeChat. OpenRouter gives a developer one account and one key across every AI model; Monid is that for the tools an agent calls rather than the models it thinks with. Its registry lists over 1,700 tools from 55 providers, and an agent asks what is available for a job, picks by fit and price, runs it with no sign-up, and is billed about a tenth of a cent for that one call, out of a single balance. Monid is not new today: it launched in April 2026 and again as 2.0 in May. What this post adds is the raise and the 4 million calls.
Shengkun Ye
13d
We just killed…everything.
Introducing Monid, the OpenRouter for agent tools.
Today we crossed 4M agent transactions. And raised a $2.1M pre-seed to kill every subscription in the world.
Your agents can now:
> discover, run, and pay for tools at runtime
> one connection to 1,800 APIs: SEO, leads, search, ecommerce, stocks, video/image/music/3D gen, private company data
> all pay per call, 0 subscriptions
Mark Zuckerberg co-founded Facebook, now Meta, and is its chief executive. Muse Code is Meta's coding agent, worked from the terminal and built on Muse Spark 1.2, a coding model trained alongside the agent rather than wrapped around a general one. It keeps a repository of thousands of files in a 1M-token context window, splits a large job across sub-agents that each work in their own git worktree so parallel edits do not collide, and prints every sub-agent, tool call, steer and cancel to an event log. Muse Code is not new today: it opened in early beta on 5 August 2026, priced from $5 a month with $15 and $50 tiers above it, and a cheaper contributor tier in some countries in return for Meta training on your prompts and completions. What this post adds is that the beta is over.
Mark Zuckerberg
13d
Muse Code is out of beta and now built to handle bigger, more complex engineering tasks. Developers can get started with one command today:
curl -fsSL https://dev.meta.ai/install.sh | bash
Eldad Fux started Appwrite as a side project in 2019. Appwrite calls itself the open-source cloud for agents and devs; it has 57,000 stars on GitHub and raised a $10 million seed led by Bessemer and Flybridge. The turn in 2.0 is that a platform built on its own abstraction now hands you the engine underneath: native PostgreSQL and MySQL that answer to psql and the mysql client, alongside a schemaless DocumentsDB and a VectorsDB for embeddings. Appwrite's own phrase for it is to give the raw engine when the abstraction is wrong. Every feature goes live on Appwrite Cloud during their Init week, and the self-hosted Community Edition follows when the week ends, with Domains and Firewall staying Cloud-only.
Eldad Fux
13d
We're thrilled to announce Appwrite 2.0
Many months of hard work from our core team and the Appwrite community have resulted in our biggest release to date, as we continue our journey to make @appwrite the most complete end-to-end developer platform 🚀
➡️ Fresh new console written in @tan_stack
➡️ Built in PostgreSQL 💙
➡️ DocumentsDB & VectorsDB
➡️ MySQL as a service
➡️ OAuth 2.1 -> sign in with [Your project]
➡️ Appwrite Firewall🔥
➡️ Appwrite Storage S3 gateway
➡️ Appwrite Domains 🌎
Let us know what you think, your feedback is what drives us and help us make Appwrite better 🙏
Learn more
Shubham Bhatt designs products at Whatfix in Bengaluru, seven years into enterprise and B2B design, at Quizizz and udaan before this. Indian Games is his own: the street games he grew up playing, rebuilt for a screen. Kanche is the first one and it runs in a browser, in two modes. In ring mode you pull the striker back and let go to knock marbles out past the chalk ring; in pit mode you land the striker in the pit to enter, then flick it at the marbles head on. You can play a bot, a friend online, or two of you on the same device.
Shubham Bhatt
13d
sound on 🔉
Kanche, Anth, Goti, Goli, call them whatever you want but these have been an integral part of our childhood, so thought of bringing each one of them to digital space.
Presenting "Kanche" by https://indianstreetgames.com
Play with a bot, or with a friend online or multiplayer in the same system, however you want.
Please try it out and let me know of the rough edges.
Michael Domarkas is 19 and founded Frontier Computing in London. He has been culturing neuronal tissue since he was 17, the company is in Y Combinator's Summer 2026 batch, and it has raised a $10 million pre-seed led by General Catalyst. The argument is that a GPU spends most of its energy moving data between memory and compute, where a neuron does both in the same place. The clip follows Cortical Labs' 2022 result of neurons in a dish learning Pong, and the 2,500x is measured against their CL1, which runs about 200,000 neurons. The 500 million-neuron cluster is a plan rather than a thing that exists.
Michael Domarkas
13d
today we're launching Frontier Computing (YC S26)
we grow neurons that colocate memory and compute, so you can more easily train ML models.
here's one of ours playing Frogger.
and we're building a 500M-neuron cluster. live EOY 2026.
that's 2,500x larger than the major biocompute systems available today
and we'll keep scaling until large model training on biocompute is widely accessible.
after all, human intelligence is built on a lipid, amino acid, glucose and vitamin cofactor soup: eating food is less costly than eating GPU compute
Anastasis Germanidis is co-founder and co-CEO of Runway, the New York company behind the Gen video models. Solaris is the first of what Runway calls interface world models: it takes your clicks, drags and typing as the input for the next frame, and an LLM decides what should happen while the video model draws it. The technical report puts it at 720p and in early access by request, and is straight about what is not solved: holding text steady, staying coherent over a long session, and screen readers, which have nothing to read when there is no code underneath.
Anastasis Germanidis
13d
Interfaces that are generated, not coded.
Solaris is our first step toward our vision of end-to-end neural software, where user interfaces are streamed directly onto your screen by a real-time video model, without any intermediate code or HTML/CSS. Today we're releasing a technical report and opening it up for limited testing by the community.
Amazing work by the team, combining the best of our research in general world models with Runway's HCI spirit.
Yasser Elsaid built Chatbase on his own after ChatGPT came out in late 2022, and has never raised venture money. It reached $8 million in yearly revenue with a team of 18, eleven of them engineers. Backstage is the part you talk to: you ask the support agent what is driving tickets and change how it answers, in plain English, from Slack, iMessage or the Chatbase dashboard.
Yasser Elsaid
13d
Announcing @Chatbase Backstage: a control center for managing your AI customer support agent.
Talk to your agent like a teammate to understand what's driving support tickets, update how it responds, deploy new Procedures, and manage workflows in plain English.
Backstage is available on iMessage, Slack, and the Chatbase dashboard, so you can manage your customer experience from wherever you work.
Sudheesh Nair is co-founder and CEO of TinyFish in Palo Alto. He ran ThoughtSpot as CEO for close to six years before this and was president of Nutanix before that. TinyFish was founded in 2024 and came out of stealth in August 2025 with a $47 million Series A led by ICONIQ. It calls what it builds enterprise web agents: software that works a website the way a person does, for the large share of the web that has no API to call. Google was an early customer, using it to pull live room availability from small hotels that have no other way to publish it.
Sudheesh Nair
13d
Why did we name a two-year-old AI-native company, building the bridge between model reasoning and the reality of the web, @Tiny_Fish, a name more fitting for a sushi restaurant?
The web is an ocean, and it has whales. Large digital properties going headless, standing up MCP servers, engineering their way into every place an agent might look for them. They'll be fine, because money always finds a way.
But what about the tour guide in Florence who gives the best walking tour in the city and has never once thought about how her site is structured for a machine to read? Or the restaurant with the best bao in the neighborhood, still running a menu photo from three years ago? They're not behind because the work is worse. They may fall behind because nobody built the web with them in mind.
We've started to notice that the words scraping and crawling actually describe taking something from a place that didn't invite you in. Model builders have taken crawling to a level unimaginable a few years ago. But we also need a web that's open, one that lets everyone thrive. An industry that talks this way has already decided the web is adversarial by default with one side extracting, the other side losing. I don't think we don't HAVE to build inside that assumption. AI gives us the opportunity to build past it.
If we just use this moment to sell more data to model builders, that's good money. Yay capitalism, nothing wrong with it. But AI also gives us the opportunity to do something massively useful for humanity while still building a real company with real revenue and a purpose behind it.
The web we want is one where showing up for AI isn't something you have to be big enough to afford, where the tour guide and the enterprise get the same shot at being seen by the agents about to do a growing share of the world's browsing, booking, and buying.
Everyone loves watching the whales, a lot more than watching the anchovies. But pull the anchovies out of the ocean and the whole chain comes down with them. Maybe they were never the small part of the story.
That's the fish we named the company after. Not the one everyone already sees, but the one the entire ocean quietly depends on to live.
