Skip to content
HeyJared

Sean‘s AI Stories

I make videos about building AI products and startups. I enjoy working on 0-1 problems and take 120% responsibility in what I believe in building. I also share interesting conversations I had with friends and great connections in my network.

YouTube · US · Official site

Indexed videos, last 90 days
9
Latest publication
Sep 22, 2026
Audience
~37.3K subscribers
Earliest in this view
Jul 5, 2026

Latest videos

  1. Video · Sep 22, 2026

    You Can Learn Jev VS LLM In 22 Min | System 1 vs System 2, AGI, Eval (opens the original)

    Excerpt · 23 min · 27,830 views by Sep 30, 2026

    Read excerpt

    Try https://waku.one, Me https://seanchen.io, Open Source https://github.com/shenseanchen/waku-agent | Jev came out of TypeSafe AI a few days ago and the launch post is past 38 million views, because it does something no chatbot does: it never talks to you. JSON in, JSON out, a probability on every answer. I draw System 1 and System 2 out on the whiteboard, step by step, so you can follow it whether or not you come from a technical background, then I race Jev against Claude Opus, Haiku 4.5 and G

  2. Video · Sep 18, 2026

    You Can Learn Personal Agent Harness (GrokBot vs Meta Muse AI) System Design | Waku Memory, Compute (opens the original)

    Excerpt · Neutral tone · 27 min · 19,690 views by Sep 30, 2026

    Read excerpt

    Try https://waku.one, Me https://seanchen.io | If you've been waiting to see what Musk's and Zuckerberg's personal AI agents actually do, this is the side by side. GrokBot came out of SpaceXAI, Meta Muse came out of Alexandr Wang's team, and I ran the same tests on both: install them, look at the architecture, check where they keep memory, hand them my calendar, and let them try to buy something. I draw both system designs out on the whiteboard, step by step, so you can follow it whether or not

  3. Video · Aug 16, 2026

    You Can Learn AI Agent Memory Layers | Graph RAG, Vector DB, SQLite, Hermes, Waku (opens the original)

    Excerpt · 30 min · 36,874 views by Sep 30, 2026

    Read excerpt

    Try https://waku.one, Me https://seanchen.io | Every LLM call starts with amnesia. The reason ChatGPT and Claude seem to remember you is not the model, it is a memory system somebody built around the model. In this video I whiteboard the three questions every agent memory has to answer, then I race five real memory layers against each other on the exact same set of facts and watch them fail in different ways. Step by step, no jargon, whether or not you come from a technical background. Here's wh

  4. Video · Aug 12, 2026

    What being a female startup founder in UK is like | London vs SF, Fundraising, Imposter Syndrome (opens the original)

    Excerpt · 66 min · 1,715 views by Sep 30, 2026

    Read excerpt

    I spent an hour in a London cafe with two female founders who are both in their early 20s and both building things I couldn't have built. Myah is building rebeauty.app, a resale marketplace for unopened beauty products. Victoria is building Ventureship.co.uk, which replaces CV screening with real project work. We ended up talking less about their products and more about the part nobody films: raising in the UK when investors want conviction and proof at the same time, getting told no, finding yo

  5. Video · Aug 1, 2026

    You Can Learn AI Agent Graph Engineering In 21 Min | Loop, Harness, System Design, Waku Agent (opens the original)

    Excerpt · Neutral tone · 22 min · 51,124 views by Sep 30, 2026

    Read excerpt

    Try https://waku.one, Me https://seanchen.io | On July 18th, Peter Steinberger asked one question on X: are we still talking about loops, or did we shift to graphs yet? Two weeks later the whole AI agent space is arguing about whether loop engineering is dead. It isn't. A loop and a graph do two different jobs, and the systems worth building use both. In this video I draw the entire AI agent engineering ladder on the whiteboard, from prompt to context to skills to loop to graph, so you can see w

Publishing over time

Last 90 days. Choose a month to open its work.

Recurring subjects

Named in the text we hold. One piece can cover several.

Audience

~37.3K subscribers

Measured Sep 19, 2026

Source's subscribers, not the number who saw an individual piece.

How this was measured

About this data

Counts cover the work we have indexed. Tone needs enough text and a confident classification. Excerpts and episode notes are not full articles or transcripts.

Identity or attribution wrong? Suggest a correction.

See coverage about Sean‘s AI Stories