Why this channel exists

Most AI coverage lands in one of two places: a press release that says a model is revolutionary, or a paper that assumes you already read the last forty. This channel sits in the gap.

Every explainer takes one piece of work and walks it through in plain language — what problem it set out to solve, what it actually did, what the numbers show, and where the claim outruns the evidence. No hype, and no pretending the hard parts are simple.

The podcast

AI with Kishor is the longer-form companion. Where the videos are tight and edited, the podcast has room for the argument to develop — and for a guest to disagree.

Who makes it

I am an AI researcher — Member of Technical Staff at Anthropic and Founder & CEO of ELITE Research Lab. My work centres on AI safety, AI for science, and low-resource languages, alongside AI agents, LLMs, explainable AI, and NLP, across 73+ publications and 880+ citations. I explain the papers I read for work the way I would to a colleague — what a piece of work claims, what it actually shows, and where it stops.

Get in touch

Spotted a mistake, want a paper explained, or have an idea for the podcast? I read everything. You can email me at kishorexplainsai@gmail.com.