Growth x Devrel x Community x Yap :) | Neuroscience × AI | also always online 😋
Bengaluru, India
Joined April 2022
- Tweets3.2K
- Following388
- Followers1.5K
- Likes2.6K
Pinned Tweet
Sending this into the universe.
I’m an engineer who loves yapping about products, creating content, and talking to people.
Built community for 4 years. 2 years in DevRel + growth.
If you’re hiring for growth, DevRel, GTM or content , let’s talk.
I’ve found every blockchain conference (in India) useless, like bro can you please wake up and talk about something genuine that one can actually use in daily life?
You’ve got me again @firefox :)
This evening, me and my brother were just playing @matiks_play when this tweet popped up!
We play it a lot and it literally hits our dopamine every time. Always cheering for you guys, all the best! Can’t wait to see what’s coming next :)
@cortisoul_ @SushantTimmapur @themohment ✨
we raised $10M to take Matiks from India to the world 🧠🌍
not long ago, it was just a few of us who believed mind sports could go mainstream. most people called it niche.
last month, 2.2M people played 48M brain games on Matiks. they could’ve kept scrolling, and they didn’t.
our longest streak has crossed 689 days. players run google meets to coach each other. one called us “the cure to my doomscrolling” and we still aren’t over it.
huge welcome to General Catalyst and Unilazer Ventures, and thank you Info Edge Ventures, Tanglin Venture Partners and Untitled Ventures for backing us again.
what’s next:
→ taking Matiks to the US
→ more games across math, memory, logic and puzzles
→ an AI engine so anyone can build a brain game millions can play
to Sudhanshu, Sushant, Mohan and the whole Matiks team: you took a simple idea and made millions fall in love with using their brains.
to our players: none of this exists without you. this one’s for you ❤️
Blr folks , it’s 2 months in (blr) HSR at Zolo, now moving to a flat. Cubbon Park area or stick with HSR , which should I choose?please help :)
Photobooth, a good matcha , and coffee in a peaceful spot surrounded by trees,You got me the perfect place 😋
📍Got Tea, HSR layout
This is 🫶🏻
For eight years, @Zetwerk, you’ve been turning the biggest of ideas into reality.
And in that spirit, we thought we’d introduce you to prototypes of the future from your toughest customers yet.
After all, most inventions start with a child asking, “What if?”
Congratulations on your IPO, Zetwerk. We can’t wait to see what you make next.
#AccelFamily • @prayanks • @acharyaamrit • @srinathramki
Sunday plan locked: chole bhature first, then Growth FAFOing at @minimi HQ with @Ozacle23
Gonna need those carbs to survive meeting @sarvagya_kul @primal_brainer & @aadharrr and actually figure out organic growth 💃😋
24,000 paid subscribers, awesome stuff
🔊 The Investment Memo, Episode 1 🎙️
Every partnership at Lightspeed starts with an investment memo, where our team writes down what it believes about a company before we partner with it. In The Investment Memo, our new series, we open that memo with the founders it was written about and read from it, so you can see how an early investment decision is made.
Our first episode is with @SuperLivingApp, which is building an AI health coach for India, starting with users in tier 2 and tier 3 towns. People already ask AI about their health, but they rarely get someone who checks in with them the next day, and SuperLiving's coaches remember each user and follow up, at a price most families can pay.
We partnered with Manavdeep Singh Grover and Gurjot Kaur early at SuperLiving's Series A. When we wrote our memo, the app had 24,000 paid subscribers after about four months. Both founders came from Pocket FM, and they have shaped the product around what their users asked for.
In the episode, Lightspeed Partner @kumarharsha2212 and Vice President @PriyalMotw89880 read our memo to Manavdeep and Gurjot section by section, from the team and market to the risks we noted, and discuss what has shifted since.
Watch the full episode here: tinyurl.com/SuperLivingPodca…
Short summary of what happened yesterday -
PHP nightmare
Today PHP got two critical bugs that let anyone run code on your server without even logging in. The details leaked before the patch was out. It affects WordPress, MediaWiki, Magento, Joomla, and Laravel. Censys says roughly 1.5 million servers are running unpatched versions.
Mass “take it offline” response
NSA, FBI, CISA, the UK, and CERT-EU all told people to take systems offline.
Japan’s Prime Minister posted at 5:55 AM telling 48.8 million people to do the same. Result: Dutch cybersecurity agency, Dutch parliament, Port of Rotterdam, government sites in Finland/France/Denmark, Wikipedia, Asana, GOG, Something Awful, University of Vienna webmail, and others all went down.
Other major incidents the same day
• GitLab: CVSS 10.0 bug that’s been there since 2014 ,
every self-hosted GitLab for 12 years.
• LiteLLM 1.107.0: Hijacked and shipped a credential stealer + remote access trojan in every wheel and Docker image (including latest). Just installing it was enough. Second time this year.
• Railway: Paused every deployment platform-wide because someone got into environment variables. Everyone has to rotate every secret stored there.
• TiDB Cloud: Someone was inside with stolen credentials for about a month. Extortion group claims they took 10.6 TB.
• Zeabur: Breached for the second time in two months and is shutting down the whole company.
• Node.js: Patched remote code execution + TLS bypass.
• Plex: CISA says it’s being actively exploited.
• Tor: Someone took over 300 exit relays to steal SSH passwords.
• SafePay ransomware: Hit over 600 critical infrastructure orgs.
• India’s biggest bank: Filed a data loss incident.
• Japan: Official state of emergency in cyberspace since October 9 (which explains the 5:55 AM post).
Certificate authority facepalm
eMudhra (a CA that has issued over 60 million certificates) was supposed to be using huge random private keys (100+ digits).
One of them was literally 4.
Others were 66 and 237.
What happens next in cybersecurity, according to the research 🧵
In 2024, researchers gave GPT-4 descriptions of 15 real vulnerabilities. It exploited 87% of them on its own. Every other model and scanner scored 0%.
arxiv.org/abs/2404.08144
In 2026, a new study gave an AI agent only Linux kernel security patches. It turned more than half of them into working exploits.
arxiv.org/abs/2602.07287
Put those together: every patch is now a blueprint for attackers.
That’s why the window keeps shrinking. In 2018, attackers needed 63 days to exploit a disclosed bug. Now it’s about 3, and almost a third are hit before the CVE is even published.
Meanwhile, AI is finding bugs faster than humans can fix them. OpenAI’s Patch the Planet has found 1,995 issues, and only 466 patches have been accepted. PHP2Shell, a 14-year-old bug in the language behind ~70% of websites, was one of them. About 1 in 3 PHP sites run versions too old to even get the fix.
And attackers are hunting AI keys. AI-related attack activity is up 89%, as platforms like Railway and OpenCode get breached for the keys they store.
What’s coming:
→ more decade-old bugs, every month
→ exploits within hours of a patch
→ stolen AI keys as criminal currency
→ AI agents as the new phishing victims
The bugs were always there. AI just turned the lights on, for everyone.
I want to meet a founder building a kill switch for API keys.
The moment a key leaks (GitHub commit, poisoned package, platform breach like Railway or OpenCode), it gets auto-revoked, a new one is issued, and every service using it is updated without anything going down.
Detection to rotation in under 60 seconds, across OpenAI, Anthropic, AWS, Stripe, everything.
This week proved we need it yesterday.
Building this or know someone who is? love to chat :)
