Pinned Tweet
If you need an investor that can put on their big boy pants and lead your round, call me.
58
4
282
been building chrome extensions to make software i use less annoying –
@Superhuman - removes the sidebar when i narrow the window so i can see just the emails
@vimcal - event previews no longer popup when moving the cursor across my calendar; it needs an actual click
@Superhuman - removes the sidebar when i narrow the window so i can see just the emails
@vimcal - event previews no longer popup when moving the cursor across my calendar; it needs an actual click
2
0
8
993
Highlight for me is the behind-the-scenes video at the end
what an insane pitch man, 10/10 @tankots
wisprflow.tanyaarora.in
if tanya doesn't get hired after this, we'll be collectively disappointed 😔
@WisprFlow
0
0
21
6,854
Wut
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: fronely read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weight-sp…Community noteLLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math.
rikiyat.github.io/blog/weight-sp…
github.com/RikiyaT/lm_und…
huggingface.co/Qwen/Qwen3.5-0…
huggingface.co/RikiyaT/Qwen3.…
0
0
0
375
If most Indian SMBs already run their business inside WhatsApp, how much does Meta's Muse have to do to win them?
In the US, Muse has to connect Shopify, QuickBooks and Canva just to become the first screen. But here, only 21% of SMBs even use business software & 97% of retail SMBsfirst screen is already Meta's.
I think the SMB story for Muse could be bigger in India than in the US, just because the path to them is so much shorter here.
In the US, Muse has to connect Shopify, QuickBooks and Canva just to become the first screen. But here, only 21% of SMBs even use business software & 97% of retail SMBsfirst screen is already Meta's.
I think the SMB story for Muse could be bigger in India than in the US, just because the path to them is so much shorter here.
7
1
59
5,032
great post x.com/omooretweets/s…
Muse for SMBs might be the most underrated story in consumer agents right now
They have near-infinite consumer distribution at very low cost, with a base of 3.64 billion daily active users across Meta apps
But even more importantly, more than 500M SMBs use Meta for marketing. coverability (preferred or not) to that SMB base, it could be a huge boost, especially for local businesses who can hyper-target users with urgent purchase needs
Meta said in their Q3 earnings calls (and in blog posts prior) that they plan to build an ecosystem of connectors for SMBs in Muse - analytics across your Meta accounts plus QuickBooks, Canva, Stripe, etc.
Many of these businesses are already trusting Meta with their growth / ad dollars - I could see Muse becoming almost a nervous system for SMBs that allows Meta to capture more and more of their activity over time
Shopify has already leaned into this (@tobi) - Muse can search across the Shopify catalog and complete purchases for consumers, and businesses can manage their Shopify store in Muse via a connector
And, the early Muse user base has shown a willingness to purchase via agents. 25% have completed a transaction via Muse, with average purchase size over $200. And, the top activity on Muse is shopping
Very curious to see what comes next here, especially with holiday gifting on the horizon 👀
0
0
2
270
Really cool
Weekend hack: HackerTube. I was annoyed with YouTube's recommendations so I had Claude filter down 43M comments over 20 years of HN to 6,454 videos that keep being recommended. *Significantly* better viewing experience, ~2 hours with Opus 5.5:
hackertube.org
- Excellent (state is kept in local browser session)
- Thumbnail is a representative frame from middle of the video instead of the super duper A/B test hyper-optimized user uploads
- Underlying data from the comment analysis as one zip file, point your Claude at it and ask for your own version
- HackerTube Longs ✊
A bit more on the process:
I use Claude Code (desktop app for Mac), Opus 5.5 on Medium effort as my default. I only talk to Claude and have it do everything, the only "manual work" I helped with in tandem was in the browser (reminiscent of MenuGen): 1) I had to create some kind of API key from Google Cloud UI to query YouTube for video metadata, 2) I booted up a VPS box for serving and 3) I bought the domain name. Once the box was up I gave Claude the IP and told it update, secure, harden, deploy and optimize the code, and it did the whole thing.
Many awesome projects that would have normally taken weeks/months even can now be few-shotted for breakfast. What else are we not thinking of.
0
0
1
622
Our portco Orange Health has just crossed the milestone of opening up their 100th collection centre! I think there's a bunch to learn from Dhruv, Tarun & how the team thinks about expansion + scale:
1/ They know how sticky their customer is (net revenue retention sits at 70-75% overt's been depth > breadth for them. More orders in the same neighbourhood mean the hubs, the pickup fleet and the labs all get used harder. In BLR they reckon each incremental rupee returns about 3x any other city.
2/ When they do expand to new cities, it has to be earned. Pune opened only once the first few markets were cash positive together, and the plan is for established cities to fund new ones. It's a pretty rare kind of discipline at this stage.
3/ The same discipline applies to new categories too. Radiology was meant to wait until BLR turned profitable, and it's now next, with CT, MRI and ultrasound for customers who already come to them for blood tests.
4/ Physical centres are a priority despite building an online-first business. The centres are there to build trust. They spend 3-4x what competitors do on designing them, and about 40% of orders placed at a centre turn into home visits.
Great piece by @thearc_hq on all of this if you want the full picture.
1/ They know how sticky their customer is (net revenue retention sits at 70-75% overt's been depth > breadth for them. More orders in the same neighbourhood mean the hubs, the pickup fleet and the labs all get used harder. In BLR they reckon each incremental rupee returns about 3x any other city.
2/ When they do expand to new cities, it has to be earned. Pune opened only once the first few markets were cash positive together, and the plan is for established cities to fund new ones. It's a pretty rare kind of discipline at this stage.
3/ The same discipline applies to new categories too. Radiology was meant to wait until BLR turned profitable, and it's now next, with CT, MRI and ultrasound for customers who already come to them for blood tests.
4/ Physical centres are a priority despite building an online-first business. The centres are there to build trust. They spend 3-4x what competitors do on designing them, and about 40% of orders placed at a centre turn into home visits.
Great piece by @thearc_hq on all of this if you want the full picture.
2
1
17
1,322
for us conviction is less a first cheque thing and more something we try to practice over the second & third. we’ve followed on in 8-10 companies now, which for a concentrated portfolio like ours is most of it.
we're neither spark nor have we invested in anthropic but @BadCapitalVC believes strongly in concentration so we've done 2-4 follow ons in fair few of our portcos
0
0
7
600
Arjun Malhotra retweeted
we're neither spark nor have we invested in anthropic but @BadCapitalVC believes strongly in concentration so we've done 2-4 follow ons in fair few of our portcos
Only venture firm named in Anthropic's S-1: Spark Capital @sparkcapital
It led the $450M round in 2023, then invested in five more rounds, nearly $800M in all. It's worth $11B at $1.5T valuation for @AnthropicAI
The most underrated skill in venture is writing the second, third and fourth cheque into a company you already know.
2
1
15
2,647
I think long-form is the future for reasons that run deep beyond just a product decision -
1/ @mosseri shares that paying creators for performance has been ROI-negative. The funds were a fixed pool split by views, and as the creator base grew the same pot got spread thinner each quarter, so payouts kept falling even for the same work. I think long form could thus give IG a format where the ad load could finally cover the payout.
2/ YouTube just reached 15.8% of all US TV watch time, more than double Netflix, with connected TV at 44% of its US viewing. The living room is where the next decade of video ads gets sold and IG has almost no footing there. The 90 min horizontal format is what gets it onto the TV at all.
I can see why that’s tempting even if the odds aren’t clear.
1/ @mosseri shares that paying creators for performance has been ROI-negative. The funds were a fixed pool split by views, and as the creator base grew the same pot got spread thinner each quarter, so payouts kept falling even for the same work. I think long form could thus give IG a format where the ad load could finally cover the payout.
2/ YouTube just reached 15.8% of all US TV watch time, more than double Netflix, with connected TV at 44% of its US viewing. The living room is where the next decade of video ads gets sold and IG has almost no footing there. The 90 min horizontal format is what gets it onto the TV at all.
I can see why that’s tempting even if the odds aren’t clear.
Instagram launches long-form video called "Features"
Creators can link up to 8 clips that show up in the feed and drive to the full video per @mosseri
1
0
4
792
what's the best solution to give openclaws the ability to make phone calls in India?
finding that latency via twilio for india calls is too high.
finding that latency via twilio for india calls is too high.
10
0
12
3,813
Every attempt to fix Indian manufacturing starts with the same instinct: consolidate the small players into something big. That playbook keeps failing, and it fails on the economics of the base itself.
India's manufacturing strength lives in thousands of clusters, with most units employing < 10 people. Each shop is specialised & retools fast because it is small, hence the firm size IS a feature.
When you roll those units into one large plant to chase scale, you kill the flexibility + specialisation that made them competitive in the first place. You end up trading away your actual structural advantage to play a scale game that China has won.
I think coordinating these small players works better than consolidating them for India. Zetwerk built its business on this premise. It orchestrates multiple small units across renewables, capital goods, aerospace & electronics - all while staying asset-light.
Building that layer however is expensive. Zetwerk does roughly ₹16k cr in revenue against a ₹1.6k cr loss. Those losses represent the upfront cost of wiring the network together. In this model, you have to fund the orchestration years ahead of the margin.
All that said, I'm still bullish on the model. Zetwerk may or may not be the one that cracks it but the coordination approach is worth backing regardless of who gets there.
India's manufacturing strength lives in thousands of clusters, with most units employing < 10 people. Each shop is specialised & retools fast because it is small, hence the firm size IS a feature.
When you roll those units into one large plant to chase scale, you kill the flexibility + specialisation that made them competitive in the first place. You end up trading away your actual structural advantage to play a scale game that China has won.
I think coordinating these small players works better than consolidating them for India. Zetwerk built its business on this premise. It orchestrates multiple small units across renewables, capital goods, aerospace & electronics - all while staying asset-light.
Building that layer however is expensive. Zetwerk does roughly ₹16k cr in revenue against a ₹1.6k cr loss. Those losses represent the upfront cost of wiring the network together. In this model, you have to fund the orchestration years ahead of the margin.
All that said, I'm still bullish on the model. Zetwerk may or may not be the one that cracks it but the coordination approach is worth backing regardless of who gets there.
Zetwerk is now listed on the NSE and BSE.
To everyone who helped build Zetwerk, thank you. We hope you feel proud today.
To our new shareholders, welcome. We have a lot more to build.
Here’s to building an enduring manufacturing company from India, for India and the world.
1
2
26
3,260
I now have a Granola agent that sits in on my WhatsApp calls.
I no longer have need for any material attachments.
I am untethered
I am free
I no longer have need for any material attachments.
I am untethered
I am free
7
1
113
8,444
Arjun Malhotra retweeted
friends, we have a new website - goodcapital.vc - pls report any bugs/breakages & would love to hear feedback - what's good, bad & what can be improved!
8
3
48
2,807
Arjun Malhotra retweeted
fruits of our labour, aka the new good cap website, is now live :) goodcapital.vc
if you come across any bugs/have any feedback - pls let us know!
if you come across any bugs/have any feedback - pls let us know!
2
1
34
1,942
Yes, SMBs haven't adopted AI at any real scale, but an interesting distinction here is that this is not a capability issue so much as a coordination one. Every new city, site or lab adds its own layer of scheduling, supervision, quality control and planning. As the physical footprint grows, management overhead balloons and margins erode under the weight of it.
While AI can make these ops-heavy businesses easier to scale from the inside, it's natural for the adoption curve to be slower (and slower in India) where many firms haven't embedded software to begin with
While AI can make these ops-heavy businesses easier to scale from the inside, it's natural for the adoption curve to be slower (and slower in India) where many firms haven't embedded software to begin with
SMBs are so far from adopting AI.
I thought the explosion of consumer agent calls and texts would pull SMBs into adoption over the next few years. But many don't even have basic online booking.
(I'm attempting to pay someone to train our dog)
2
1
7
839