From organising startup events to making DeepTech bets — Arpit’s journey has been anything but linear. In the first episode of Unscripted, Snigdha Sengupta from The Runway sits down with Arpit Agarwal (Arpit), Partner at Blume Ventures, to trace his evolution — from volunteering at

WhatsApp is now the Mecca of spammers. Can’t block them. Meta WhatsApp Are you actively promoting this? How else is the same business flouting the “norms” repeatedly? TRAI can you help?

Many many such stories. Sadly all true

At a Macro level, when India was growing due to economic reforms, Bihar was regressing. Per Capita Income rose only 0.12% for Bihar vs 4.08% for India. Given the high population of Bihar, this damaged not only the state but the country. en.wikipedia.org/wiki/Economy_o…

Well said. My two cents: a lot of work remains before we can say AI is the best teacher. It already works for autodidacts but that is not even 1% of the student base. Making it work for the 99% is the big challenge; best done in small segments - one at a time.

Incredible work Carolynn Levy, the creator of the SAFE 🫡

One of the core life hacks - pay attention to what people do and the outcomes they generate. NOT to the titles and statements - they usually mean very little. If the professors or influencers were so good, they would be investing instead of talking. amit

Gen Z, believe me, these things will not change. Accept them now else be forced to later: Money is very important Being rich beats being poor Being very rich usually feels good Your own house is your own house A car is better than a bike Family always comes first Families

If you are wondering what were the CXOs thinking, read between the lines: av CXO tenure is less than 5 years in Fortune 500 (linkedin.com/posts/diversit…) and they're measured QoQ for performance. None had incentives for a long enough horizon Noah Smith 🐇🇺🇸🇺🇦🇹🇼

A live case study of Tragedy of commons. India is full of them

Why do LLMs do well on evals but fail on real world task - Ilya says that its because the RL training is focussed on making the models do well on the evals. Makes a lot of sense - incentives drive performance youtu.be/aR20FWCCjAs?si…