Mrigank Bhargava

Still stuck answering “khane mein kya banana hai?” every morning? most apps just help you plan your diet. cookmate actually gets it done. It lives on WhatsApp and coordinates with your cook - voice notes, recipes, alternatives, ordering groceries. one less thing to think about every day. link: cookmate.food

Brilliant read in how AI is enhancing music production 🙌🏼

Folks, I have a 2.5 BHK on rent at Prestige Ferns Residency, Harlur. I’ve lived here by myself for the last 4 years, but I no longer stay in Bangalore full-time and ended up spending only about 3 months here last year. This year is likely to be similar, so I’m looking for someone to share the apartment. The house is fully furnished by me, has a green, minimal vibe, and is part of one of the best gym and sports club communities in Bangalore. One room is available for long-term sublet. Details: Total rent: ₹63K (room-wise split can be discussed). Fully furnished with a stocked kitchen. You just need to bring your bags and move in. PS: The only condition is maintaining the house in its original state. @BangaloreRoomi please do your magic.

A BIG ONE: Today, we're excited to announce that @fibr_ai has raised $7.5M in Seed funding, led by @Accel (@AccelIndia), with participation from WillowTree Ventures, MVP Capital, and an amazing group of operator angels who know how hard real systems are to build. I don’t think about funding as a milestone. I think about it as time. Time to keep chasing a belief without cutting corners, time to keep sitting with uncomfortable product questions, time to build the thing the way it actually needs to be built. The belief we’re chasing is simple, but deeply structural: the website, the URL itself, is broken. URLs were built for static pages. But today’s traffic arrives with context, from ads, search, and now AI systems, and the website still resets everything to the same page. That breaks down fast. Human visitors arrive mid-journey. AI agents increasingly evaluate sites before a person ever clicks. The static web wasn’t built for either. At Fibr AI, we’re rebuilding the website as a system, not a set of pages: an AI-native experience layer where each URL observes intent, retains context, and improves over time. This funding lets us keep going deep on the hard infrastructure problems and build web systems that actually learn as they run. Grateful today to the team that obsesses over the details no one sees, and our customers who push us with real problems, not hypotheticals. Back to building. #AgenticWeb #AI #Technology #startup

Video editing just became a lot easier 🚀 Congrats on the launch @sxmawl and @ishandeveloper

@ycombinator @usecardboard Congrats on the launch @sxmawl and @ishandeveloper 🚀🚀

.@usecardboard is an agentic video editor where you upload raw footage and edit it with AI. Unicorns, media agencies, and creators with millions of followers are already using Cardboard. It lives in your browser and helps you go from a sea of raw clips into a finished video in minutes. Congrats on the launch @sxmawl and @ishandeveloper! ycombinator.com/launches/PM3-c…

ive probably “written” more code in the last 2 weeks than during the first 7yrs of my career writing code full time combined the era of “the idea guy” has arrived

@sxmawl Congrats 🚀

Tl;dr: we got into YC W26!!!! We’re building Cardboard (agentic video editor) A few months ago, I shut down my last startup and booked a flight to SF (didn't go into debt). I shared my story publicly for the first time, and it ended up getting 85k+ impressions on LinkedIn. I was vulnerable, but I knew I wanted to do the -1 to 0 journey again and just follow my curiosity tbh. In SF, I met founders, researchers, and builders I’ve looked up to for many years. They are earnest, truth-seeking, and obsessed with their craft. It was like my X feed is out of this tiny rectangle. The thing that stood out most was simple: bias to action. On Nov 1, I started building with my school best friend, @ishandeveloper. We picked a problem we both felt personally: video editing is still painfully slow and manual today. We realised that there isn’t a VS Code-like open, extensible editor we can build agents on top of so we had to build everything from scratch. The momentum was great, we were able to get to an agent that single-shot Cardboard’s product demo from random screen recordings. I asked him - do you want to apply to YC? And he said “yes”. And in the process, we realised @garrytan was right, filling the application definitely levels up your idea. It forced us to rethink everything from first principles. We were pretty stoic about the application. So we just kept our heads down and started shipping features every single day. We submitted an honest update to YC, and after that everything moved really fast. We got an interview the next day and acceptance the day after that. Huge thanks to @maitreya_wagh, @srinisarkar, @JayantPanwar17 and @_shubhankar for helping us with the application. The YC news was a breath of fresh air after a really long time. We doubled down, Ishan quit, we moved out of Bangalore, and in the middle of all that, we shipped like a 10-person team with just 2 people. Since then, we’ve shipped: - Keyframes, transcription + captions, canvas editing - Advanced background options, playback speed - Agent overhaul and AI video understanding - Music library, crop and text tools. - Blazing fast exports and a new landing page We’re launching in January 2026 - so keep an eye out. Meanwhile, we're inviting design partners: if you’re leading or working in growth or marketing teams that use video as a distribution channel to ship launch videos, ads, recaps, testimonials, or weekly content, and you care about speed, I’d love to work closely with you. Cardboard will be the fastest way for a team to go from raw assets to a polished story. Not “a better timeline.” An editor that behaves like a teammate. It understands your footage, takes first passes, iterates with feedback, and helps you ship more videos with the same team. DM me if you want early access to a design partner slot.

We launched @arcxtrade last week. Traders showed up. Feedback poured in. Now we’re hiring a Full-Stack Engineer to help us build what users are already asking for. Frontend-heavy initially but you will definitely be fiddling with backend too. Bonus points if you’ve built in web3 before. 📍 In-person role in Gurgaon, India ⏱️ Looking for someone who can start immediately (at least part-time to begin with). DM us if this sounds like you.

@TheIshanGoswami @Nihalgoyal3 @aman_jaiinn @ShivayMadan Amazing, lfg

Introducing Health Wrapped! Get your 2025 Apple health insights. Your steps, workouts, and movements beautifully visualised. Built by my friends @Nihalgoyal3 @aman_jaiinn @shivaymadan Try Now: healthwrapped.com

Optimize for Momentum

Wide range LPs >>> Tight range LPs for most. Been digging through a ton of WETH/USDC concentrated liquidity positions using @revertfinance and the pattern is pretty clear: - Over a long enough horizon, wide range LPs tend to outperform tight range LPs. - People get baited by screenshots + FOMO from 400%+ APYs, but: Wide ranges make more sense for most: 1. Less brutal IL when price moves, and over time you beat holding 2. More passive: No need to rebalance every few days or every week 3. You stay in range longer: I’d rather earn a sustainable 30% APY for 12 months than a flashy “120% APY” for a week then sit out of range Stop chasing unsustainable APYs Be wide, be boring, be sustainable.

Been gathering my thoughts on Uniswap's UNIfication proposal. It surely strengthens value capture for $UNI by directing protocol fees to burns, but it also means LPs receive a smaller share of fees. For the 0.01% and 0.05% fee tier pools roughly a 25% reduction in fee revenue, and for the 0.30% and 1% tiers it’s closer to 16.7%. That definitely will change the yield math for LPs. There’s also the MEV internalisation angle, which could offset some of that by reducing LVR but I feel it still wouldn't recover a large chunk of the value back. The bigger thing on my mind is simply that liquidity is not sticky and relatively easy to rotate. LPs usually go where the net yield is best, be it direct fees or incentives. With this change, I do think liquidity positioning will gradually shift. • Some LPs will likely move from lower fee tiers to higher fee tiers trying to rebalance returns especially for wide-range LPers. • Some capital slowly shifts to protocols like Aerodrome where emissions boost yields. • Some liquidity moves from V2 (yes there is still a large chunk there) and select V3 pools → V4 / L2s as a phased rollout plays out. This will be in short term Personally I’ve preferred wide range LPing on Uniswap because it usually outperformed the constant “chase the range”/"narrow" strategies in the long run, and it avoided some of the automated liquidity rotation you see on Aerodrome making returns more stable which is something I prefer. btw amazing content by @richardnhsu who explains why wide range LPs > Narrow ranges in the long run But with this fee shift, it might be worth revisiting and slowly comparing outcomes across options. Keen to learn from others who are actively optimizing concentrated LP strategies and just see how the incentives rebalance in real market conditions.

@sxmawl CFBR

Subletting 1 Room | Koramangala 1st Block - Available Nov–Dec (2 months) Spacious 3BHK, just 5min walk from Coolulu Turf / Loco Bear / D-Mart. Well-ventilated, plenty of sunlight, clean, aesthetic, and fully setup - kitchen stocked, Wi-Fi, water, electricity, washing, all sorted. 0 setup charges - just move in with your bags. Rent: ₹25k (incl. maintenance) House help: ₹1.3k Flatmates: both BITS Pilani CS grads - one founder (Accel/Antler-backed), another founding eng at an a16z-funded startup. DM if interested / want more details.

What if you could have your own Perplexity like framework that can be run with any LLM? > Introducing OpenProbe - an open-source deep research agent, our creation from the @BerkeleyRDI AgentX hackathon! > Inspired by @SentientAGI's OpenDeepSearch (ODS) > Outperforms GPT-4o-Search on DeepMind's FRAMES Benchmark (+1.5% accuracy) Try it out! github.com/Open-Probe/Ope… (1/8)

🤖 Earlier this week the Agents Payment Protocol (AP2) was announced by @googlecloud, with EigenCloud as a launch partner to power the verification layer of the agentic economy. Today, we show a glimpse of how EigenCloud provides the trust rails for the coordination and transactions of the incoming trillions of agents. Learn more 🧵⬇️

@Padamn__ happy birthday 🫡

Turned 26. Ran the same for good measure

Founder mood swings in one chart: ↓ No growth → depression ↑ Growth → euphoria