Sana Gupta retweeted
the kalaari fellowship has been a j curve for my learning over the 2 years. This would be a great opportunity for anyone that resonates with the following -
> deep interest in the early stage startup ecosystem
> curious about frontier technologies and their broader implications on society
> social! you like talking to founders, operators, building communities, hosting events et al
> creative and original - you like building povs *and* presenting them in digestible formats like writing, decks, interactive websites etc etc etc
the role is a mixture of a lot of things but the ambiguity is a feature; which you can exploit to make the most out of this opportunity!
above all this fellowship is a court-side seat for arguably what is the best and most exciting quarters/halves/innings of our indian startup ecosystem.
if this excites you, do apply to our latest fellowship batch :)
This quoted post is unavailable.
Sana Gupta retweeted
Kalaari Fellowship is a two-year, paid, immersive VC apprenticeship program with opportunities to work alongside investors and founders, explore emerging technologies, evaluate opportunities, contribute to investment thinking, and gain hands-on experience across the venture ecosystem.
If you're inspired by the progress being made in frontier tech sectors, want to build a future in venture investing, and contribute to the next chapter of India's innovation story — apply for Kalaari Fellowship Batch 5.0.
Applications are now live. Link in comments.
Sana Gupta retweeted
Quick commerce shelves have become the new D2C launchpad in India: 70%+ respondents to a BofA survey willing to try unfamiliar brands featured on QC apps
Sana Gupta retweeted
Fascinating heatmap mapping the shared users across e-commerce, social and delivery platforms in India.
Sana Gupta retweeted
Localmaxxing : pushing more inference to local models.
Over five weeks, I tested how much of my daily work can run on a local 35B model instead of cloud frontier models. The answer : half.
Many reasons to use local models : privacy, cost, asset depreciation.
But the only one that really matters is latency.
I ran a head-to-head benchmark. Qwen 3.6 35B-A3B-4bit on my MacBook Pro M5 vs Claude Opus 4.5 via API.
Result : 2.1x faster locally. Mean 2.8s vs 5.8s.
The local model isn't smarter. Opus scores ~20% higher on reasoning benchmarks. Local models lag frontier by 3-4 months, and for complex tasks, that gap matters.
But for routine agent tasks, it rarely does.
If half the work runs 2x faster on my laptop, I'll take that trade every time.
My little computer is about to earn its keep.
tomtunguz.com/localmaxxing/
