tight feedback loops in agents is all we need.

enterprises always like to think they are ahead in tech by adopting them, innovating, normal users always behind (not the geeks) but historically normal users always dictated what the world used including the enterprises. behaviour is the king. that’s for another day why some

If you don't understand this, you will not understand why LLM-based agents are irreparably failing for a general-purpose problem solving. An agent (by the way it was the topic of my PhD 20 years ago) to be useful, must be rational. Being rational means to always prefer an

AI wrapper - lame, undifferentiated, commoditized AI harness - mysterious, technical, strategic

It was never easy and with each passing day you only start doubting your decision as I do mine at times. but that’s the point nation building isn’t easy Sacrifices are never easy If India had all the things sorted we wouldn’t need amazing people to be here And I truly still

this is cute

AI interviewing humans was already pretty novel. AI interviewing AI? That's the frontier. Synthetic user feedback is one of the most bleeding-edge developments in AI-powered customer discovery right now. The concept: train AI to simulate what specific customer segments would

hahaha no truer words were ever spoken. I literally joke about this with my husband that so many smart folks having real jobs don’t have time to do AI because its so exploratory in nature you cannot learn that with a project or tutorial, its consistent learning everyday, everyday

I think they meant to say that anyone can think of an idea, but not everyone can afford to work towards them or implement them. The privileged or one with authority might think they are cheap 😝

Indian behaviour has been leading the software digital and mobile interfaces from many years now, but still what Indian investors wish to see is there something in US comparable to your product? How sad it is to hear this, after so many years I thought we have improved and

Internet was expensive when it started So were cars And electricity We need it to be used to become usable.

If we consider AI hallucinations as dreams - solving for this can have two steps 1. wake up (awareness) 2. restart from when you were awake

I think most of the people who expected too heightened output from AI are genuinely pissed and don’t see things going in the direction they had envisioned or trusted. this creates a very unstable thinking, one side is continuing on the same pattern but with too cautious outlook

I have a question about cold dms. So the suggestions are always, have a line or two and that should excite the other person to respond. I don’t know why though I do understand the gimmicky nature of this and a bait kind of thing, also that the people you are trying to reach

The cost of avoiding structure is that every day has to be renegotiated from scratch.

to build something exceptional using AI, consider problems above the feasibility in AI, AI will follow up soon, but don’t dwell a lot on the possibilities of tech right now. you might build something because you don’t know it was impossible.