Video yükleniyor...
Video Yüklenemedi
Andrew Ng just released a 1-hour course on building Context Agentic Graphs from scratch: • 00:00 - Introduction to agentic knowledge Graphs • 03:07 - Construction of agentic Graphs • 14:00 - Architecture of multi-agent systems • 23:00 - Building agentic graphs with Google ADK • 01:06:03 - Why... show more
799,243 görüntüleme • 2 ay önce •via X (Twitter)
27 Yorum

saving this for the neo4j half. entity resolution is what wrecked my last graph mate

and how you solve it?

saw this lecture 2 weeks ago already everyone who hasn't seen it, you're making a huge mistake

I’m sure I will take huge value from this video

and you right

1 hour with Andrew feels like gold

its literally gold, huge value from one hour

insane value in here

fact man

@0xMortyx definitely need to carve out an hour for this. been diving into agentic graphs lately, and it's fascinating stuff.

agentic graphs finally explained properly

The next step is not just giving agents more context. It's giving them the right context.

This is the real cost of fanning out agents. Parallel execution is the easy part. Reconciling 300 outputs into one coherent memory is the part nobody budgets time for.

more Andrew courses, please

Context graphs could make agent memory and reasoning way more powerful

agree with you

brilliant lecture

i also enjoy it

booked, thx for your help

Context engineering is quietly becoming the core differentiator in agentic systems — more than model choice itself. The multi-agent architecture section at 23:00 is probably the most practical part for anyone already past the basics.

The multi-agent architecture section is the real meat—curious how they handle state sharing between agents without turning it into a debugging nightmare.

The real video is more than two hours.

@grok when was this course released?

Free course from Andrew Ng again

andrew ng is one of the smartest people in ai

@grok when was this lecture ?

The multi-agent architecture section is the real meat—curious how they handle shared state between agents without it becoming a bottleneck.
