When working with LLMs I am used to starting "New Conversation" for each request. But there is also the polar opposite approach of keeping one giant conversation going forever. The standard approach can still choose to use a Memory tool to write things down in between

I completed 2 years in VC last month. I’ve led a few deals, seen some uprounds, and navigated crises. I’ve also lost a bunch of deals. Here’s what I learned, written as notes to myself, some where I did a 180 on what I thought before I started:

20/ The only truth in VC:

Bullish on the MCP economy! V likely we will see 21st dev like case studies coming out of India soon cline.bot/blog/building-…

“LLM systems reason like people. Wait, what is reasoning exactly?" I’ve often found it hard to articulate why LLM-based AI systems are not able to fully replace a human. Why are these not “AGI”? What is missing? They seem to do well on most human exams, so haven’t we solved

The best founders operate on a different clock - and it drives most of their employees crazy..

Curious to know how JioHotstar renders the 3D images of players in IPL. While the quality is not bad by any stretch, it feels a little 'game-like' for the lack of a better word 🤔

Voice will undoubtedly be the killer app of the post-AI world and we will look back at the decades we spent sitting in front of a screen typing stuff one letter at a time. We started investing in this in 2023 first and the validation and constant improvement in all layers of

search.app.goo.gl/eBgPXtQ AI-generated 35-sec audio recap of Day 1 of India vs England Test match. Fair to assume Sundar Pichai’s love for cricket had something to do with this getting fast-tracked? 🏏🎧

The Accel Cybersecurity Summit is back on July 30, 2025. Last year’s debut brought together 400+ top minds across cybersecurity: founders, CISOs, researchers, and operators - from India, the US, Israel, UK, UAE, and Canada. What started as a summit has grown into a real

It feels to me that we are on trajectory to have keyboards & speech input go away before 2030? Any contrarians agree? And it won't be Neuralink or anything invasive or anything intrusive in public places like speech! And it might be multi-modal like thought?