Ankit Sapkota (@ankitsa16352100) 's Twitter Profile
Ankit Sapkota

@ankitsa16352100

Be more human | BEI

ID: 1484769009799426053

calendar_today22-01-2022 06:05:22

158 Tweet

42 Followers

121 Following

Ankit Sapkota (@ankitsa16352100) 's Twitter Profile Photo

I Spent a week surveying IT company in Nepal. My takeaway: everyone is solving today’s problems. No one is building for tomorrow.

Ankit Sapkota (@ankitsa16352100) 's Twitter Profile Photo

dear X, I know you can hear me and I know you know a lot about me. I have chosen you over all other social media. Now stop sending me those unwanted political “who did what and who said what” bullshit's. I don’t care. I don’t want f*****g breaking news. connect me to my people.

Ankit Sapkota (@ankitsa16352100) 's Twitter Profile Photo

I’m onto intelligent data engines these days, diving through lexical, semantic search, and all kinds of engineering in between. The world needs a beautiful interface between businesses and their huge, messy databases. RAG, voice, agents… dream it however you want.

Ankit Sapkota (@ankitsa16352100) 's Twitter Profile Photo

sometimes words can be the difference between the problem and the solution you want. I am searching for a better word for the pre-established understanding between an intelligent system and a knowledge base. metadata can’t be the whole story. ontology sounds good for now.

Ankit Sapkota (@ankitsa16352100) 's Twitter Profile Photo

i’m thinking to put everything balen does into a knowledge base combine it with the promises he made add how media + people reacted. and build a system to interact with it.

Ankit Sapkota (@ankitsa16352100) 's Twitter Profile Photo

when someone asks me about telecommunication, all I remember is the general vibe of the subject. maybe a few chapters. beyond that, I always have to look it up. this is a window into human retrieval: we remember the vibe and the structure.

Ankit Sapkota (@ankitsa16352100) 's Twitter Profile Photo

Every field goes through a phase of noise, jargon, and overpromises. What remains is simple: systems that worked. The mice they caught.